Commit
•
7866ae8
1
Parent(s):
6431ed8
Training in progress, step 32000
Browse files- README.md +357 -0
- added_tokens.json +5 -1
- all_results.json +14 -0
- config.json +1 -1
- eval_results.json +9 -0
- merges.txt +0 -0
- pytorch_model.bin +2 -2
- tokenizer.json +0 -0
- tokenizer_config.json +44 -0
- train_results.json +8 -0
- trainer_state.json +3328 -0
- training_args.bin +2 -2
- vocab.json +0 -0
README.md
ADDED
|
@@ -0,0 +1,357 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
base_model: roberta-base
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
model-index:
|
| 7 |
+
- name: roberta-2020-Q1-filtered
|
| 8 |
+
results: []
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 12 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 13 |
+
|
| 14 |
+
# roberta-2020-Q1-filtered
|
| 15 |
+
|
| 16 |
+
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an unknown dataset.
|
| 17 |
+
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: 2.7087
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 1e-05
|
| 38 |
+
- train_batch_size: 16
|
| 39 |
+
- eval_batch_size: 16
|
| 40 |
+
- seed: 42
|
| 41 |
+
- optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-08
|
| 42 |
+
- lr_scheduler_type: linear
|
| 43 |
+
- lr_scheduler_warmup_steps: 1400
|
| 44 |
+
- training_steps: 2400000
|
| 45 |
+
|
| 46 |
+
### Training results
|
| 47 |
+
|
| 48 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 49 |
+
|:-------------:|:-----:|:-------:|:---------------:|
|
| 50 |
+
| No log | 0.07 | 8000 | 2.9780 |
|
| 51 |
+
| 3.1296 | 0.13 | 16000 | 2.8948 |
|
| 52 |
+
| 3.1296 | 0.2 | 24000 | 2.8590 |
|
| 53 |
+
| 2.9018 | 0.26 | 32000 | 2.8033 |
|
| 54 |
+
| 2.9018 | 0.33 | 40000 | 2.7938 |
|
| 55 |
+
| 2.8331 | 0.39 | 48000 | 2.7695 |
|
| 56 |
+
| 2.8331 | 0.46 | 56000 | 2.7614 |
|
| 57 |
+
| 2.7723 | 0.52 | 64000 | 2.7417 |
|
| 58 |
+
| 2.7723 | 0.59 | 72000 | 2.7249 |
|
| 59 |
+
| 2.75 | 0.65 | 80000 | 2.7202 |
|
| 60 |
+
| 2.75 | 0.72 | 88000 | 2.7112 |
|
| 61 |
+
| 2.735 | 0.78 | 96000 | 2.7229 |
|
| 62 |
+
| 2.735 | 0.85 | 104000 | 2.7371 |
|
| 63 |
+
| 2.7137 | 0.91 | 112000 | 2.7059 |
|
| 64 |
+
| 2.7137 | 0.98 | 120000 | 2.7121 |
|
| 65 |
+
| 2.7155 | 1.04 | 128000 | 2.7249 |
|
| 66 |
+
| 2.7155 | 1.11 | 136000 | 2.7131 |
|
| 67 |
+
| 2.7152 | 1.17 | 144000 | 2.7000 |
|
| 68 |
+
| 2.7152 | 1.24 | 152000 | 2.7030 |
|
| 69 |
+
| 2.7151 | 1.3 | 160000 | 2.7214 |
|
| 70 |
+
| 2.7151 | 1.37 | 168000 | 2.7076 |
|
| 71 |
+
| 2.7166 | 1.44 | 176000 | 2.7106 |
|
| 72 |
+
| 2.7166 | 1.5 | 184000 | 2.7197 |
|
| 73 |
+
| 2.7144 | 1.57 | 192000 | 2.7101 |
|
| 74 |
+
| 2.7144 | 1.63 | 200000 | 2.7235 |
|
| 75 |
+
| 2.7179 | 1.7 | 208000 | 2.7066 |
|
| 76 |
+
| 2.7179 | 1.76 | 216000 | 2.7283 |
|
| 77 |
+
| 2.7231 | 1.83 | 224000 | 2.7203 |
|
| 78 |
+
| 2.7231 | 1.89 | 232000 | 2.7111 |
|
| 79 |
+
| 2.7284 | 1.96 | 240000 | 2.7217 |
|
| 80 |
+
| 2.7284 | 2.02 | 248000 | 2.7251 |
|
| 81 |
+
| 2.7242 | 2.09 | 256000 | 2.7181 |
|
| 82 |
+
| 2.7242 | 2.15 | 264000 | 2.7238 |
|
| 83 |
+
| 2.7171 | 2.22 | 272000 | 2.7488 |
|
| 84 |
+
| 2.7171 | 2.28 | 280000 | 2.7315 |
|
| 85 |
+
| 2.7312 | 2.35 | 288000 | 2.7469 |
|
| 86 |
+
| 2.7312 | 2.41 | 296000 | 2.7363 |
|
| 87 |
+
| 2.7386 | 2.48 | 304000 | 2.7398 |
|
| 88 |
+
| 2.7386 | 2.54 | 312000 | 2.7477 |
|
| 89 |
+
| 2.7457 | 2.61 | 320000 | 2.7536 |
|
| 90 |
+
| 2.7457 | 2.67 | 328000 | 2.7483 |
|
| 91 |
+
| 2.7496 | 2.74 | 336000 | 2.7529 |
|
| 92 |
+
| 2.7496 | 2.8 | 344000 | 2.7492 |
|
| 93 |
+
| 2.7521 | 2.87 | 352000 | 2.7612 |
|
| 94 |
+
| 2.7521 | 2.94 | 360000 | 2.7701 |
|
| 95 |
+
| 2.7649 | 3.0 | 368000 | 2.7705 |
|
| 96 |
+
| 2.7649 | 3.07 | 376000 | 2.7828 |
|
| 97 |
+
| 2.7516 | 3.13 | 384000 | 2.7680 |
|
| 98 |
+
| 2.7516 | 3.2 | 392000 | 2.7843 |
|
| 99 |
+
| 2.762 | 3.26 | 400000 | 2.7916 |
|
| 100 |
+
| 2.762 | 3.33 | 408000 | 2.7692 |
|
| 101 |
+
| 2.7789 | 3.39 | 416000 | 2.7834 |
|
| 102 |
+
| 2.7789 | 3.46 | 424000 | 2.7788 |
|
| 103 |
+
| 2.7879 | 3.52 | 432000 | 2.8037 |
|
| 104 |
+
| 2.7879 | 3.59 | 440000 | 2.7919 |
|
| 105 |
+
| 2.7853 | 3.65 | 448000 | 2.8077 |
|
| 106 |
+
| 2.7853 | 3.72 | 456000 | 2.7903 |
|
| 107 |
+
| 2.7976 | 3.78 | 464000 | 2.8109 |
|
| 108 |
+
| 2.7976 | 3.85 | 472000 | 2.7957 |
|
| 109 |
+
| 2.789 | 3.91 | 480000 | 2.8023 |
|
| 110 |
+
| 2.789 | 3.98 | 488000 | 2.8126 |
|
| 111 |
+
| 2.8089 | 4.04 | 496000 | 2.8154 |
|
| 112 |
+
| 2.8089 | 4.11 | 504000 | 2.8123 |
|
| 113 |
+
| 2.7915 | 4.17 | 512000 | 2.8146 |
|
| 114 |
+
| 2.7915 | 4.24 | 520000 | 2.8250 |
|
| 115 |
+
| 2.8094 | 4.31 | 528000 | 2.8206 |
|
| 116 |
+
| 2.8094 | 4.37 | 536000 | 2.8182 |
|
| 117 |
+
| 2.8196 | 4.44 | 544000 | 2.8351 |
|
| 118 |
+
| 2.8196 | 4.5 | 552000 | 2.8394 |
|
| 119 |
+
| 2.8316 | 4.57 | 560000 | 2.8397 |
|
| 120 |
+
| 2.8316 | 4.63 | 568000 | 2.8403 |
|
| 121 |
+
| 2.8444 | 4.7 | 576000 | 2.8351 |
|
| 122 |
+
| 2.8444 | 4.76 | 584000 | 2.8574 |
|
| 123 |
+
| 2.833 | 4.83 | 592000 | 2.8617 |
|
| 124 |
+
| 2.833 | 4.89 | 600000 | 2.8578 |
|
| 125 |
+
| 2.839 | 4.96 | 608000 | 2.8577 |
|
| 126 |
+
| 2.839 | 5.02 | 616000 | 2.8727 |
|
| 127 |
+
| 2.8427 | 5.09 | 624000 | 2.8586 |
|
| 128 |
+
| 2.8427 | 5.15 | 632000 | 2.8808 |
|
| 129 |
+
| 2.8599 | 5.22 | 640000 | 2.8960 |
|
| 130 |
+
| 2.8599 | 5.28 | 648000 | 2.8883 |
|
| 131 |
+
| 2.8694 | 5.35 | 656000 | 2.8885 |
|
| 132 |
+
| 2.8694 | 5.41 | 664000 | 2.8873 |
|
| 133 |
+
| 2.8626 | 5.48 | 672000 | 2.8930 |
|
| 134 |
+
| 2.8626 | 5.54 | 680000 | 2.8988 |
|
| 135 |
+
| 2.8921 | 5.61 | 688000 | 2.9117 |
|
| 136 |
+
| 2.8921 | 5.68 | 696000 | 2.9122 |
|
| 137 |
+
| 2.8884 | 5.74 | 704000 | 2.9001 |
|
| 138 |
+
| 2.8884 | 5.81 | 712000 | 2.9094 |
|
| 139 |
+
| 2.8974 | 5.87 | 720000 | 2.9110 |
|
| 140 |
+
| 2.8974 | 5.94 | 728000 | 2.9045 |
|
| 141 |
+
| 2.903 | 6.0 | 736000 | 2.9337 |
|
| 142 |
+
| 2.903 | 6.07 | 744000 | 2.9316 |
|
| 143 |
+
| 2.9057 | 6.13 | 752000 | 2.9447 |
|
| 144 |
+
| 2.9057 | 6.2 | 760000 | 2.9363 |
|
| 145 |
+
| 2.9146 | 6.26 | 768000 | 2.9438 |
|
| 146 |
+
| 2.9146 | 6.33 | 776000 | 2.9475 |
|
| 147 |
+
| 2.9221 | 6.39 | 784000 | 2.9394 |
|
| 148 |
+
| 2.9221 | 6.46 | 792000 | 2.9371 |
|
| 149 |
+
| 2.9316 | 6.52 | 800000 | 2.9494 |
|
| 150 |
+
| 2.9316 | 6.59 | 808000 | 2.9727 |
|
| 151 |
+
| 2.9421 | 6.65 | 816000 | 2.9759 |
|
| 152 |
+
| 2.9421 | 6.72 | 824000 | 2.9665 |
|
| 153 |
+
| 2.9538 | 6.78 | 832000 | 2.9650 |
|
| 154 |
+
| 2.9538 | 6.85 | 840000 | 2.9761 |
|
| 155 |
+
| 2.9594 | 6.91 | 848000 | 2.9901 |
|
| 156 |
+
| 2.9594 | 6.98 | 856000 | 2.9732 |
|
| 157 |
+
| 2.9564 | 7.05 | 864000 | 2.9897 |
|
| 158 |
+
| 2.9564 | 7.11 | 872000 | 2.9801 |
|
| 159 |
+
| 2.9561 | 7.18 | 880000 | 2.9839 |
|
| 160 |
+
| 2.9561 | 7.24 | 888000 | 2.9888 |
|
| 161 |
+
| 2.9669 | 7.31 | 896000 | 3.0000 |
|
| 162 |
+
| 2.9669 | 7.37 | 904000 | 2.9786 |
|
| 163 |
+
| 2.9649 | 7.44 | 912000 | 2.9946 |
|
| 164 |
+
| 2.9649 | 7.5 | 920000 | 3.0002 |
|
| 165 |
+
| 2.9665 | 7.57 | 928000 | 2.9960 |
|
| 166 |
+
| 2.9665 | 7.63 | 936000 | 3.0068 |
|
| 167 |
+
| 2.9708 | 7.7 | 944000 | 2.9938 |
|
| 168 |
+
| 2.9708 | 7.76 | 952000 | 3.0126 |
|
| 169 |
+
| 2.981 | 7.83 | 960000 | 2.9959 |
|
| 170 |
+
| 2.981 | 7.89 | 968000 | 2.9960 |
|
| 171 |
+
| 2.9805 | 7.96 | 976000 | 2.9919 |
|
| 172 |
+
| 2.9805 | 8.02 | 984000 | 3.0058 |
|
| 173 |
+
| 2.9705 | 8.09 | 992000 | 3.0232 |
|
| 174 |
+
| 2.9705 | 8.15 | 1000000 | 3.0047 |
|
| 175 |
+
| 2.9715 | 8.22 | 1008000 | 3.0069 |
|
| 176 |
+
| 2.9715 | 8.28 | 1016000 | 3.0019 |
|
| 177 |
+
| 2.9695 | 8.35 | 1024000 | 3.0216 |
|
| 178 |
+
| 2.9695 | 8.41 | 1032000 | 3.0219 |
|
| 179 |
+
| 2.9762 | 8.48 | 1040000 | 3.0182 |
|
| 180 |
+
| 2.9762 | 8.55 | 1048000 | 3.0332 |
|
| 181 |
+
| 2.9786 | 8.61 | 1056000 | 3.0017 |
|
| 182 |
+
| 2.9786 | 8.68 | 1064000 | 3.0236 |
|
| 183 |
+
| 2.9889 | 8.74 | 1072000 | 3.0273 |
|
| 184 |
+
| 2.9889 | 8.81 | 1080000 | 3.0197 |
|
| 185 |
+
| 2.9842 | 8.87 | 1088000 | 3.0376 |
|
| 186 |
+
| 2.9842 | 8.94 | 1096000 | 3.0323 |
|
| 187 |
+
| 2.9912 | 9.0 | 1104000 | 3.0317 |
|
| 188 |
+
| 2.9912 | 9.07 | 1112000 | 3.0225 |
|
| 189 |
+
| 2.9919 | 9.13 | 1120000 | 3.0361 |
|
| 190 |
+
| 2.9919 | 9.2 | 1128000 | 3.0432 |
|
| 191 |
+
| 2.9872 | 9.26 | 1136000 | 3.0307 |
|
| 192 |
+
| 2.9872 | 9.33 | 1144000 | 3.0482 |
|
| 193 |
+
| 2.9823 | 9.39 | 1152000 | 3.0354 |
|
| 194 |
+
| 2.9823 | 9.46 | 1160000 | 3.0419 |
|
| 195 |
+
| 2.9882 | 9.52 | 1168000 | 3.0567 |
|
| 196 |
+
| 2.9882 | 9.59 | 1176000 | 3.0395 |
|
| 197 |
+
| 3.0079 | 9.65 | 1184000 | 3.0572 |
|
| 198 |
+
| 3.0079 | 9.72 | 1192000 | 3.0403 |
|
| 199 |
+
| 3.0243 | 9.78 | 1200000 | 3.0472 |
|
| 200 |
+
| 3.0243 | 9.85 | 1208000 | 3.0523 |
|
| 201 |
+
| 3.0127 | 9.92 | 1216000 | 3.0534 |
|
| 202 |
+
| 3.0127 | 9.98 | 1224000 | 3.0434 |
|
| 203 |
+
| 3.0106 | 10.05 | 1232000 | 3.0687 |
|
| 204 |
+
| 3.0106 | 10.11 | 1240000 | 3.0678 |
|
| 205 |
+
| 3.0063 | 10.18 | 1248000 | 3.0652 |
|
| 206 |
+
| 3.0063 | 10.24 | 1256000 | 3.0768 |
|
| 207 |
+
| 3.0187 | 10.31 | 1264000 | 3.0692 |
|
| 208 |
+
| 3.0187 | 10.37 | 1272000 | 3.0621 |
|
| 209 |
+
| 3.0202 | 10.44 | 1280000 | 3.0663 |
|
| 210 |
+
| 3.0202 | 10.5 | 1288000 | 3.0537 |
|
| 211 |
+
| 3.0219 | 10.57 | 1296000 | 3.0725 |
|
| 212 |
+
| 3.0219 | 10.63 | 1304000 | 3.0664 |
|
| 213 |
+
| 3.0232 | 10.7 | 1312000 | 3.0724 |
|
| 214 |
+
| 3.0232 | 10.76 | 1320000 | 3.0476 |
|
| 215 |
+
| 3.0247 | 10.83 | 1328000 | 3.0729 |
|
| 216 |
+
| 3.0247 | 10.89 | 1336000 | 3.0646 |
|
| 217 |
+
| 3.0335 | 10.96 | 1344000 | 3.0604 |
|
| 218 |
+
| 3.0335 | 11.02 | 1352000 | 3.0631 |
|
| 219 |
+
| 3.0182 | 11.09 | 1360000 | 3.0669 |
|
| 220 |
+
| 3.0182 | 11.15 | 1368000 | 3.0626 |
|
| 221 |
+
| 3.0124 | 11.22 | 1376000 | 3.0535 |
|
| 222 |
+
| 3.0124 | 11.29 | 1384000 | 3.0768 |
|
| 223 |
+
| 3.016 | 11.35 | 1392000 | 3.0615 |
|
| 224 |
+
| 3.016 | 11.42 | 1400000 | 3.0689 |
|
| 225 |
+
| 3.0133 | 11.48 | 1408000 | 3.0699 |
|
| 226 |
+
| 3.0133 | 11.55 | 1416000 | 3.0647 |
|
| 227 |
+
| 3.0227 | 11.61 | 1424000 | 3.0705 |
|
| 228 |
+
| 3.0227 | 11.68 | 1432000 | 3.0706 |
|
| 229 |
+
| 3.0267 | 11.74 | 1440000 | 3.0694 |
|
| 230 |
+
| 3.0267 | 11.81 | 1448000 | 3.0721 |
|
| 231 |
+
| 3.021 | 11.87 | 1456000 | 3.0690 |
|
| 232 |
+
| 3.021 | 11.94 | 1464000 | 3.0603 |
|
| 233 |
+
| 3.0144 | 12.0 | 1472000 | 3.0658 |
|
| 234 |
+
| 3.0144 | 12.07 | 1480000 | 3.0720 |
|
| 235 |
+
| 3.0204 | 12.13 | 1488000 | 3.0668 |
|
| 236 |
+
| 3.0204 | 12.2 | 1496000 | 3.0773 |
|
| 237 |
+
| 3.0085 | 12.26 | 1504000 | 3.0848 |
|
| 238 |
+
| 3.0085 | 12.33 | 1512000 | 3.0568 |
|
| 239 |
+
| 3.0146 | 12.39 | 1520000 | 3.0783 |
|
| 240 |
+
| 3.0146 | 12.46 | 1528000 | 3.0736 |
|
| 241 |
+
| 3.02 | 12.52 | 1536000 | 3.0534 |
|
| 242 |
+
| 3.02 | 12.59 | 1544000 | 3.0684 |
|
| 243 |
+
| 3.0229 | 12.65 | 1552000 | 3.0767 |
|
| 244 |
+
| 3.0229 | 12.72 | 1560000 | 3.0569 |
|
| 245 |
+
| 3.0152 | 12.79 | 1568000 | 3.0788 |
|
| 246 |
+
| 3.0152 | 12.85 | 1576000 | 3.0663 |
|
| 247 |
+
| 3.02 | 12.92 | 1584000 | 3.0670 |
|
| 248 |
+
| 3.02 | 12.98 | 1592000 | 3.0683 |
|
| 249 |
+
| 3.0128 | 13.05 | 1600000 | 3.0718 |
|
| 250 |
+
| 3.0128 | 13.11 | 1608000 | 3.0847 |
|
| 251 |
+
| 3.016 | 13.18 | 1616000 | 3.0664 |
|
| 252 |
+
| 3.016 | 13.24 | 1624000 | 3.0688 |
|
| 253 |
+
| 3.0007 | 13.31 | 1632000 | 3.0741 |
|
| 254 |
+
| 3.0007 | 13.37 | 1640000 | 3.0663 |
|
| 255 |
+
| 3.0241 | 13.44 | 1648000 | 3.0607 |
|
| 256 |
+
| 3.0241 | 13.5 | 1656000 | 3.0635 |
|
| 257 |
+
| 3.0103 | 13.57 | 1664000 | 3.0731 |
|
| 258 |
+
| 3.0103 | 13.63 | 1672000 | 3.0649 |
|
| 259 |
+
| 3.0188 | 13.7 | 1680000 | 3.0587 |
|
| 260 |
+
| 3.0188 | 13.76 | 1688000 | 3.0704 |
|
| 261 |
+
| 3.0217 | 13.83 | 1696000 | 3.0664 |
|
| 262 |
+
| 3.0217 | 13.89 | 1704000 | 3.0627 |
|
| 263 |
+
| 3.0282 | 13.96 | 1712000 | 3.0714 |
|
| 264 |
+
| 3.0282 | 14.02 | 1720000 | 3.0688 |
|
| 265 |
+
| 3.0166 | 14.09 | 1728000 | 3.0521 |
|
| 266 |
+
| 3.0166 | 14.16 | 1736000 | 3.0538 |
|
| 267 |
+
| 3.0134 | 14.22 | 1744000 | 3.0641 |
|
| 268 |
+
| 3.0134 | 14.29 | 1752000 | 3.0639 |
|
| 269 |
+
| 3.0032 | 14.35 | 1760000 | 3.0588 |
|
| 270 |
+
| 3.0032 | 14.42 | 1768000 | 3.0646 |
|
| 271 |
+
| 3.0136 | 14.48 | 1776000 | 3.0629 |
|
| 272 |
+
| 3.0136 | 14.55 | 1784000 | 3.0578 |
|
| 273 |
+
| 3.0086 | 14.61 | 1792000 | 3.0529 |
|
| 274 |
+
| 3.0086 | 14.68 | 1800000 | 3.0615 |
|
| 275 |
+
| 3.019 | 14.74 | 1808000 | 3.0566 |
|
| 276 |
+
| 3.019 | 14.81 | 1816000 | 3.0659 |
|
| 277 |
+
| 3.024 | 14.87 | 1824000 | 3.0615 |
|
| 278 |
+
| 3.024 | 14.94 | 1832000 | 3.0530 |
|
| 279 |
+
| 3.0089 | 15.0 | 1840000 | 3.0797 |
|
| 280 |
+
| 3.0089 | 15.07 | 1848000 | 3.0700 |
|
| 281 |
+
| 3.0174 | 15.13 | 1856000 | 3.0748 |
|
| 282 |
+
| 3.0174 | 15.2 | 1864000 | 3.0643 |
|
| 283 |
+
| 3.0176 | 15.26 | 1872000 | 3.0628 |
|
| 284 |
+
| 3.0176 | 15.33 | 1880000 | 3.0630 |
|
| 285 |
+
| 3.0164 | 15.39 | 1888000 | 3.0722 |
|
| 286 |
+
| 3.0164 | 15.46 | 1896000 | 3.0744 |
|
| 287 |
+
| 3.0302 | 15.53 | 1904000 | 3.0739 |
|
| 288 |
+
| 3.0302 | 15.59 | 1912000 | 3.0700 |
|
| 289 |
+
| 3.0204 | 15.66 | 1920000 | 3.0751 |
|
| 290 |
+
| 3.0204 | 15.72 | 1928000 | 3.0598 |
|
| 291 |
+
| 3.0147 | 15.79 | 1936000 | 3.0522 |
|
| 292 |
+
| 3.0147 | 15.85 | 1944000 | 3.0655 |
|
| 293 |
+
| 3.0245 | 15.92 | 1952000 | 3.0569 |
|
| 294 |
+
| 3.0245 | 15.98 | 1960000 | 3.0623 |
|
| 295 |
+
| 3.0069 | 16.05 | 1968000 | 3.0600 |
|
| 296 |
+
| 3.0069 | 16.11 | 1976000 | 3.0639 |
|
| 297 |
+
| 3.0068 | 16.18 | 1984000 | 3.0775 |
|
| 298 |
+
| 3.0068 | 16.24 | 1992000 | 3.0669 |
|
| 299 |
+
| 3.0275 | 16.31 | 2000000 | 3.0627 |
|
| 300 |
+
| 3.0275 | 16.37 | 2008000 | 3.0645 |
|
| 301 |
+
| 3.0164 | 16.44 | 2016000 | 3.0667 |
|
| 302 |
+
| 3.0164 | 16.5 | 2024000 | 3.0490 |
|
| 303 |
+
| 3.0148 | 16.57 | 2032000 | 3.0618 |
|
| 304 |
+
| 3.0148 | 16.63 | 2040000 | 3.0545 |
|
| 305 |
+
| 3.022 | 16.7 | 2048000 | 3.0651 |
|
| 306 |
+
| 3.022 | 16.76 | 2056000 | 3.0687 |
|
| 307 |
+
| 3.0235 | 16.83 | 2064000 | 3.0516 |
|
| 308 |
+
| 3.0235 | 16.89 | 2072000 | 3.0761 |
|
| 309 |
+
| 3.0194 | 16.96 | 2080000 | 3.0807 |
|
| 310 |
+
| 3.0194 | 17.03 | 2088000 | 3.0601 |
|
| 311 |
+
| 3.0142 | 17.09 | 2096000 | 3.0721 |
|
| 312 |
+
| 3.0142 | 17.16 | 2104000 | 3.0653 |
|
| 313 |
+
| 3.0183 | 17.22 | 2112000 | 3.0617 |
|
| 314 |
+
| 3.0183 | 17.29 | 2120000 | 3.0622 |
|
| 315 |
+
| 3.0092 | 17.35 | 2128000 | 3.0682 |
|
| 316 |
+
| 3.0092 | 17.42 | 2136000 | 3.0732 |
|
| 317 |
+
| 3.0071 | 17.48 | 2144000 | 3.0763 |
|
| 318 |
+
| 3.0071 | 17.55 | 2152000 | 3.0675 |
|
| 319 |
+
| 3.0272 | 17.61 | 2160000 | 3.0671 |
|
| 320 |
+
| 3.0272 | 17.68 | 2168000 | 3.0622 |
|
| 321 |
+
| 3.0235 | 17.74 | 2176000 | 3.0789 |
|
| 322 |
+
| 3.0235 | 17.81 | 2184000 | 3.0623 |
|
| 323 |
+
| 3.0179 | 17.87 | 2192000 | 3.0784 |
|
| 324 |
+
| 3.0179 | 17.94 | 2200000 | 3.0629 |
|
| 325 |
+
| 3.0209 | 18.0 | 2208000 | 3.0731 |
|
| 326 |
+
| 3.0209 | 18.07 | 2216000 | 3.0946 |
|
| 327 |
+
| 3.0237 | 18.13 | 2224000 | 3.0653 |
|
| 328 |
+
| 3.0237 | 18.2 | 2232000 | 3.0590 |
|
| 329 |
+
| 3.0164 | 18.26 | 2240000 | 3.0707 |
|
| 330 |
+
| 3.0164 | 18.33 | 2248000 | 3.0546 |
|
| 331 |
+
| 3.0206 | 18.4 | 2256000 | 3.0742 |
|
| 332 |
+
| 3.0206 | 18.46 | 2264000 | 3.0793 |
|
| 333 |
+
| 3.0138 | 18.53 | 2272000 | 3.0560 |
|
| 334 |
+
| 3.0138 | 18.59 | 2280000 | 3.0870 |
|
| 335 |
+
| 3.0377 | 18.66 | 2288000 | 3.0742 |
|
| 336 |
+
| 3.0377 | 18.72 | 2296000 | 3.0676 |
|
| 337 |
+
| 3.0227 | 18.79 | 2304000 | 3.0625 |
|
| 338 |
+
| 3.0227 | 18.85 | 2312000 | 3.0736 |
|
| 339 |
+
| 3.0359 | 18.92 | 2320000 | 3.0801 |
|
| 340 |
+
| 3.0359 | 18.98 | 2328000 | 3.0710 |
|
| 341 |
+
| 3.0248 | 19.05 | 2336000 | 3.0692 |
|
| 342 |
+
| 3.0248 | 19.11 | 2344000 | 3.0677 |
|
| 343 |
+
| 3.0235 | 19.18 | 2352000 | 3.0896 |
|
| 344 |
+
| 3.0235 | 19.24 | 2360000 | 3.0778 |
|
| 345 |
+
| 3.0187 | 19.31 | 2368000 | 3.0700 |
|
| 346 |
+
| 3.0187 | 19.37 | 2376000 | 3.0743 |
|
| 347 |
+
| 3.0189 | 19.44 | 2384000 | 3.0780 |
|
| 348 |
+
| 3.0189 | 19.5 | 2392000 | 3.0867 |
|
| 349 |
+
| 3.0184 | 19.57 | 2400000 | 3.0793 |
|
| 350 |
+
|
| 351 |
+
|
| 352 |
+
### Framework versions
|
| 353 |
+
|
| 354 |
+
- Transformers 4.35.0.dev0
|
| 355 |
+
- Pytorch 2.0.1+cu117
|
| 356 |
+
- Datasets 2.14.5
|
| 357 |
+
- Tokenizers 0.14.0
|
added_tokens.json
CHANGED
|
@@ -1,3 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"</s>": 2,
|
| 3 |
+
"<mask>": 64000,
|
| 4 |
+
"<pad>": 1,
|
| 5 |
+
"<s>": 0,
|
| 6 |
+
"<unk>": 3
|
| 7 |
}
|
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 19.57,
|
| 3 |
+
"eval_loss": 2.708733320236206,
|
| 4 |
+
"eval_runtime": 133.4182,
|
| 5 |
+
"eval_samples": 103275,
|
| 6 |
+
"eval_samples_per_second": 774.07,
|
| 7 |
+
"eval_steps_per_second": 48.382,
|
| 8 |
+
"perplexity": 15.010250287695158,
|
| 9 |
+
"train_loss": 2.9400340771484377,
|
| 10 |
+
"train_runtime": 198144.865,
|
| 11 |
+
"train_samples": 1962232,
|
| 12 |
+
"train_samples_per_second": 193.798,
|
| 13 |
+
"train_steps_per_second": 12.112
|
| 14 |
+
}
|
config.json
CHANGED
|
@@ -22,7 +22,7 @@
|
|
| 22 |
"position_embedding_type": "absolute",
|
| 23 |
"tokenizer_class": "BertweetTokenizer",
|
| 24 |
"torch_dtype": "float32",
|
| 25 |
-
"transformers_version": "4.
|
| 26 |
"type_vocab_size": 1,
|
| 27 |
"use_cache": true,
|
| 28 |
"vocab_size": 64001
|
|
|
|
| 22 |
"position_embedding_type": "absolute",
|
| 23 |
"tokenizer_class": "BertweetTokenizer",
|
| 24 |
"torch_dtype": "float32",
|
| 25 |
+
"transformers_version": "4.35.0.dev0",
|
| 26 |
"type_vocab_size": 1,
|
| 27 |
"use_cache": true,
|
| 28 |
"vocab_size": 64001
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 19.57,
|
| 3 |
+
"eval_loss": 2.708733320236206,
|
| 4 |
+
"eval_runtime": 133.4182,
|
| 5 |
+
"eval_samples": 103275,
|
| 6 |
+
"eval_samples_per_second": 774.07,
|
| 7 |
+
"eval_steps_per_second": 48.382,
|
| 8 |
+
"perplexity": 15.010250287695158
|
| 9 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f4a0b30f688781c6fdd50a63198ba658b059d997fb6be1d00b2fef843a978b03
|
| 3 |
+
size 539931509
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
CHANGED
|
@@ -1,4 +1,47 @@
|
|
| 1 |
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
"bos_token": "<s>",
|
| 3 |
"clean_up_tokenization_spaces": true,
|
| 4 |
"cls_token": "<s>",
|
|
@@ -9,5 +52,6 @@
|
|
| 9 |
"pad_token": "<pad>",
|
| 10 |
"sep_token": "</s>",
|
| 11 |
"tokenizer_class": "BertweetTokenizer",
|
|
|
|
| 12 |
"unk_token": "<unk>"
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"additional_special_tokens": [],
|
| 45 |
"bos_token": "<s>",
|
| 46 |
"clean_up_tokenization_spaces": true,
|
| 47 |
"cls_token": "<s>",
|
|
|
|
| 52 |
"pad_token": "<pad>",
|
| 53 |
"sep_token": "</s>",
|
| 54 |
"tokenizer_class": "BertweetTokenizer",
|
| 55 |
+
"tokenizer_file": "/home/CIN/jdps/.cache/huggingface/hub/models--vinai--bertweet-base/snapshots/118ab1d567653bec16bbb081eafb6f8942f72108/tokenizer.json",
|
| 56 |
"unk_token": "<unk>"
|
| 57 |
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 19.57,
|
| 3 |
+
"train_loss": 2.9400340771484377,
|
| 4 |
+
"train_runtime": 198144.865,
|
| 5 |
+
"train_samples": 1962232,
|
| 6 |
+
"train_samples_per_second": 193.798,
|
| 7 |
+
"train_steps_per_second": 12.112
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3328 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 2.7100777626037598,
|
| 3 |
+
"best_model_checkpoint": "./model_tweets_2020_Q2/checkpoint-192000",
|
| 4 |
+
"epoch": 19.569471624266146,
|
| 5 |
+
"eval_steps": 8000,
|
| 6 |
+
"global_step": 2400000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.07,
|
| 13 |
+
"eval_loss": 2.977957248687744,
|
| 14 |
+
"eval_runtime": 125.5213,
|
| 15 |
+
"eval_samples_per_second": 822.769,
|
| 16 |
+
"eval_steps_per_second": 51.426,
|
| 17 |
+
"step": 8000
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.13,
|
| 21 |
+
"learning_rate": 9.939131159843243e-06,
|
| 22 |
+
"loss": 3.1296,
|
| 23 |
+
"step": 16000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.13,
|
| 27 |
+
"eval_loss": 2.894831418991089,
|
| 28 |
+
"eval_runtime": 126.0129,
|
| 29 |
+
"eval_samples_per_second": 819.559,
|
| 30 |
+
"eval_steps_per_second": 51.225,
|
| 31 |
+
"step": 16000
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.2,
|
| 35 |
+
"eval_loss": 2.8589611053466797,
|
| 36 |
+
"eval_runtime": 125.8909,
|
| 37 |
+
"eval_samples_per_second": 820.353,
|
| 38 |
+
"eval_steps_per_second": 51.275,
|
| 39 |
+
"step": 24000
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.26,
|
| 43 |
+
"learning_rate": 9.872425581589261e-06,
|
| 44 |
+
"loss": 2.9018,
|
| 45 |
+
"step": 32000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.26,
|
| 49 |
+
"eval_loss": 2.8033480644226074,
|
| 50 |
+
"eval_runtime": 125.7264,
|
| 51 |
+
"eval_samples_per_second": 821.427,
|
| 52 |
+
"eval_steps_per_second": 51.342,
|
| 53 |
+
"step": 32000
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.33,
|
| 57 |
+
"eval_loss": 2.7938032150268555,
|
| 58 |
+
"eval_runtime": 125.7192,
|
| 59 |
+
"eval_samples_per_second": 821.474,
|
| 60 |
+
"eval_steps_per_second": 51.345,
|
| 61 |
+
"step": 40000
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.39,
|
| 65 |
+
"learning_rate": 9.80572000333528e-06,
|
| 66 |
+
"loss": 2.8331,
|
| 67 |
+
"step": 48000
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 0.39,
|
| 71 |
+
"eval_loss": 2.7694976329803467,
|
| 72 |
+
"eval_runtime": 127.0405,
|
| 73 |
+
"eval_samples_per_second": 812.93,
|
| 74 |
+
"eval_steps_per_second": 50.811,
|
| 75 |
+
"step": 48000
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.46,
|
| 79 |
+
"eval_loss": 2.7614457607269287,
|
| 80 |
+
"eval_runtime": 125.7185,
|
| 81 |
+
"eval_samples_per_second": 821.478,
|
| 82 |
+
"eval_steps_per_second": 51.345,
|
| 83 |
+
"step": 56000
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.52,
|
| 87 |
+
"learning_rate": 9.739014425081299e-06,
|
| 88 |
+
"loss": 2.7723,
|
| 89 |
+
"step": 64000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.52,
|
| 93 |
+
"eval_loss": 2.7416625022888184,
|
| 94 |
+
"eval_runtime": 126.1624,
|
| 95 |
+
"eval_samples_per_second": 818.588,
|
| 96 |
+
"eval_steps_per_second": 51.164,
|
| 97 |
+
"step": 64000
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 0.59,
|
| 101 |
+
"eval_loss": 2.7248806953430176,
|
| 102 |
+
"eval_runtime": 126.0454,
|
| 103 |
+
"eval_samples_per_second": 819.348,
|
| 104 |
+
"eval_steps_per_second": 51.212,
|
| 105 |
+
"step": 72000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.65,
|
| 109 |
+
"learning_rate": 9.672308846827316e-06,
|
| 110 |
+
"loss": 2.75,
|
| 111 |
+
"step": 80000
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.65,
|
| 115 |
+
"eval_loss": 2.7202229499816895,
|
| 116 |
+
"eval_runtime": 126.948,
|
| 117 |
+
"eval_samples_per_second": 813.522,
|
| 118 |
+
"eval_steps_per_second": 50.848,
|
| 119 |
+
"step": 80000
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.72,
|
| 123 |
+
"eval_loss": 2.7112038135528564,
|
| 124 |
+
"eval_runtime": 126.8524,
|
| 125 |
+
"eval_samples_per_second": 814.135,
|
| 126 |
+
"eval_steps_per_second": 50.886,
|
| 127 |
+
"step": 88000
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.78,
|
| 131 |
+
"learning_rate": 9.605603268573334e-06,
|
| 132 |
+
"loss": 2.735,
|
| 133 |
+
"step": 96000
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 0.78,
|
| 137 |
+
"eval_loss": 2.7228710651397705,
|
| 138 |
+
"eval_runtime": 126.981,
|
| 139 |
+
"eval_samples_per_second": 813.311,
|
| 140 |
+
"eval_steps_per_second": 50.834,
|
| 141 |
+
"step": 96000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.85,
|
| 145 |
+
"eval_loss": 2.7370951175689697,
|
| 146 |
+
"eval_runtime": 126.6893,
|
| 147 |
+
"eval_samples_per_second": 815.183,
|
| 148 |
+
"eval_steps_per_second": 50.951,
|
| 149 |
+
"step": 104000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.91,
|
| 153 |
+
"learning_rate": 9.538897690319354e-06,
|
| 154 |
+
"loss": 2.7137,
|
| 155 |
+
"step": 112000
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 0.91,
|
| 159 |
+
"eval_loss": 2.7059037685394287,
|
| 160 |
+
"eval_runtime": 126.3306,
|
| 161 |
+
"eval_samples_per_second": 817.498,
|
| 162 |
+
"eval_steps_per_second": 51.096,
|
| 163 |
+
"step": 112000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.98,
|
| 167 |
+
"eval_loss": 2.7120730876922607,
|
| 168 |
+
"eval_runtime": 126.2744,
|
| 169 |
+
"eval_samples_per_second": 817.862,
|
| 170 |
+
"eval_steps_per_second": 51.119,
|
| 171 |
+
"step": 120000
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 1.04,
|
| 175 |
+
"learning_rate": 9.472192112065373e-06,
|
| 176 |
+
"loss": 2.7155,
|
| 177 |
+
"step": 128000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 1.04,
|
| 181 |
+
"eval_loss": 2.7248668670654297,
|
| 182 |
+
"eval_runtime": 126.8126,
|
| 183 |
+
"eval_samples_per_second": 814.391,
|
| 184 |
+
"eval_steps_per_second": 50.902,
|
| 185 |
+
"step": 128000
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.11,
|
| 189 |
+
"eval_loss": 2.7130985260009766,
|
| 190 |
+
"eval_runtime": 126.6262,
|
| 191 |
+
"eval_samples_per_second": 815.589,
|
| 192 |
+
"eval_steps_per_second": 50.977,
|
| 193 |
+
"step": 136000
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 1.17,
|
| 197 |
+
"learning_rate": 9.405486533811392e-06,
|
| 198 |
+
"loss": 2.7152,
|
| 199 |
+
"step": 144000
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 1.17,
|
| 203 |
+
"eval_loss": 2.6999881267547607,
|
| 204 |
+
"eval_runtime": 126.4279,
|
| 205 |
+
"eval_samples_per_second": 816.868,
|
| 206 |
+
"eval_steps_per_second": 51.057,
|
| 207 |
+
"step": 144000
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 1.24,
|
| 211 |
+
"eval_loss": 2.703012704849243,
|
| 212 |
+
"eval_runtime": 126.2932,
|
| 213 |
+
"eval_samples_per_second": 817.74,
|
| 214 |
+
"eval_steps_per_second": 51.111,
|
| 215 |
+
"step": 152000
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 1.3,
|
| 219 |
+
"learning_rate": 9.338780955557409e-06,
|
| 220 |
+
"loss": 2.7151,
|
| 221 |
+
"step": 160000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 1.3,
|
| 225 |
+
"eval_loss": 2.721385955810547,
|
| 226 |
+
"eval_runtime": 126.566,
|
| 227 |
+
"eval_samples_per_second": 815.977,
|
| 228 |
+
"eval_steps_per_second": 51.001,
|
| 229 |
+
"step": 160000
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 1.37,
|
| 233 |
+
"eval_loss": 2.707641839981079,
|
| 234 |
+
"eval_runtime": 126.5896,
|
| 235 |
+
"eval_samples_per_second": 815.826,
|
| 236 |
+
"eval_steps_per_second": 50.992,
|
| 237 |
+
"step": 168000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 1.44,
|
| 241 |
+
"learning_rate": 9.272075377303427e-06,
|
| 242 |
+
"loss": 2.7166,
|
| 243 |
+
"step": 176000
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 1.44,
|
| 247 |
+
"eval_loss": 2.7106387615203857,
|
| 248 |
+
"eval_runtime": 126.9356,
|
| 249 |
+
"eval_samples_per_second": 813.602,
|
| 250 |
+
"eval_steps_per_second": 50.853,
|
| 251 |
+
"step": 176000
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 1.5,
|
| 255 |
+
"eval_loss": 2.719717025756836,
|
| 256 |
+
"eval_runtime": 127.5317,
|
| 257 |
+
"eval_samples_per_second": 809.798,
|
| 258 |
+
"eval_steps_per_second": 50.615,
|
| 259 |
+
"step": 184000
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 1.57,
|
| 263 |
+
"learning_rate": 9.205369799049446e-06,
|
| 264 |
+
"loss": 2.7144,
|
| 265 |
+
"step": 192000
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 1.57,
|
| 269 |
+
"eval_loss": 2.7100777626037598,
|
| 270 |
+
"eval_runtime": 126.3318,
|
| 271 |
+
"eval_samples_per_second": 817.49,
|
| 272 |
+
"eval_steps_per_second": 51.096,
|
| 273 |
+
"step": 192000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 1.63,
|
| 277 |
+
"eval_loss": 2.723472833633423,
|
| 278 |
+
"eval_runtime": 127.1568,
|
| 279 |
+
"eval_samples_per_second": 812.186,
|
| 280 |
+
"eval_steps_per_second": 50.764,
|
| 281 |
+
"step": 200000
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.7,
|
| 285 |
+
"learning_rate": 9.138664220795464e-06,
|
| 286 |
+
"loss": 2.7179,
|
| 287 |
+
"step": 208000
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.7,
|
| 291 |
+
"eval_loss": 2.706564426422119,
|
| 292 |
+
"eval_runtime": 127.089,
|
| 293 |
+
"eval_samples_per_second": 812.62,
|
| 294 |
+
"eval_steps_per_second": 50.791,
|
| 295 |
+
"step": 208000
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 1.76,
|
| 299 |
+
"eval_loss": 2.7282984256744385,
|
| 300 |
+
"eval_runtime": 127.4927,
|
| 301 |
+
"eval_samples_per_second": 810.047,
|
| 302 |
+
"eval_steps_per_second": 50.63,
|
| 303 |
+
"step": 216000
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 1.83,
|
| 307 |
+
"learning_rate": 9.071958642541483e-06,
|
| 308 |
+
"loss": 2.7231,
|
| 309 |
+
"step": 224000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 1.83,
|
| 313 |
+
"eval_loss": 2.7203216552734375,
|
| 314 |
+
"eval_runtime": 127.4298,
|
| 315 |
+
"eval_samples_per_second": 810.446,
|
| 316 |
+
"eval_steps_per_second": 50.655,
|
| 317 |
+
"step": 224000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.89,
|
| 321 |
+
"eval_loss": 2.711085319519043,
|
| 322 |
+
"eval_runtime": 126.4739,
|
| 323 |
+
"eval_samples_per_second": 816.571,
|
| 324 |
+
"eval_steps_per_second": 51.038,
|
| 325 |
+
"step": 232000
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 1.96,
|
| 329 |
+
"learning_rate": 9.005253064287502e-06,
|
| 330 |
+
"loss": 2.7284,
|
| 331 |
+
"step": 240000
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 1.96,
|
| 335 |
+
"eval_loss": 2.721714973449707,
|
| 336 |
+
"eval_runtime": 126.401,
|
| 337 |
+
"eval_samples_per_second": 817.043,
|
| 338 |
+
"eval_steps_per_second": 51.068,
|
| 339 |
+
"step": 240000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 2.02,
|
| 343 |
+
"eval_loss": 2.725090265274048,
|
| 344 |
+
"eval_runtime": 127.0199,
|
| 345 |
+
"eval_samples_per_second": 813.061,
|
| 346 |
+
"eval_steps_per_second": 50.819,
|
| 347 |
+
"step": 248000
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 2.09,
|
| 351 |
+
"learning_rate": 8.93854748603352e-06,
|
| 352 |
+
"loss": 2.7242,
|
| 353 |
+
"step": 256000
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 2.09,
|
| 357 |
+
"eval_loss": 2.718090057373047,
|
| 358 |
+
"eval_runtime": 127.9402,
|
| 359 |
+
"eval_samples_per_second": 807.213,
|
| 360 |
+
"eval_steps_per_second": 50.453,
|
| 361 |
+
"step": 256000
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 2.15,
|
| 365 |
+
"eval_loss": 2.723750591278076,
|
| 366 |
+
"eval_runtime": 127.1287,
|
| 367 |
+
"eval_samples_per_second": 812.366,
|
| 368 |
+
"eval_steps_per_second": 50.775,
|
| 369 |
+
"step": 264000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 2.22,
|
| 373 |
+
"learning_rate": 8.871841907779539e-06,
|
| 374 |
+
"loss": 2.7171,
|
| 375 |
+
"step": 272000
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 2.22,
|
| 379 |
+
"eval_loss": 2.748772144317627,
|
| 380 |
+
"eval_runtime": 128.6406,
|
| 381 |
+
"eval_samples_per_second": 802.818,
|
| 382 |
+
"eval_steps_per_second": 50.179,
|
| 383 |
+
"step": 272000
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 2.28,
|
| 387 |
+
"eval_loss": 2.731541633605957,
|
| 388 |
+
"eval_runtime": 127.5368,
|
| 389 |
+
"eval_samples_per_second": 809.766,
|
| 390 |
+
"eval_steps_per_second": 50.613,
|
| 391 |
+
"step": 280000
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 2.35,
|
| 395 |
+
"learning_rate": 8.805136329525557e-06,
|
| 396 |
+
"loss": 2.7312,
|
| 397 |
+
"step": 288000
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 2.35,
|
| 401 |
+
"eval_loss": 2.746854305267334,
|
| 402 |
+
"eval_runtime": 127.6128,
|
| 403 |
+
"eval_samples_per_second": 809.284,
|
| 404 |
+
"eval_steps_per_second": 50.583,
|
| 405 |
+
"step": 288000
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 2.41,
|
| 409 |
+
"eval_loss": 2.7363078594207764,
|
| 410 |
+
"eval_runtime": 127.9259,
|
| 411 |
+
"eval_samples_per_second": 807.303,
|
| 412 |
+
"eval_steps_per_second": 50.459,
|
| 413 |
+
"step": 296000
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 2.48,
|
| 417 |
+
"learning_rate": 8.738430751271576e-06,
|
| 418 |
+
"loss": 2.7386,
|
| 419 |
+
"step": 304000
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 2.48,
|
| 423 |
+
"eval_loss": 2.7398250102996826,
|
| 424 |
+
"eval_runtime": 127.3013,
|
| 425 |
+
"eval_samples_per_second": 811.264,
|
| 426 |
+
"eval_steps_per_second": 50.706,
|
| 427 |
+
"step": 304000
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 2.54,
|
| 431 |
+
"eval_loss": 2.747743844985962,
|
| 432 |
+
"eval_runtime": 127.9865,
|
| 433 |
+
"eval_samples_per_second": 806.921,
|
| 434 |
+
"eval_steps_per_second": 50.435,
|
| 435 |
+
"step": 312000
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 2.61,
|
| 439 |
+
"learning_rate": 8.671725173017595e-06,
|
| 440 |
+
"loss": 2.7457,
|
| 441 |
+
"step": 320000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 2.61,
|
| 445 |
+
"eval_loss": 2.753558397293091,
|
| 446 |
+
"eval_runtime": 128.9208,
|
| 447 |
+
"eval_samples_per_second": 801.073,
|
| 448 |
+
"eval_steps_per_second": 50.069,
|
| 449 |
+
"step": 320000
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 2.67,
|
| 453 |
+
"eval_loss": 2.748337984085083,
|
| 454 |
+
"eval_runtime": 128.758,
|
| 455 |
+
"eval_samples_per_second": 802.086,
|
| 456 |
+
"eval_steps_per_second": 50.133,
|
| 457 |
+
"step": 328000
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 2.74,
|
| 461 |
+
"learning_rate": 8.605019594763613e-06,
|
| 462 |
+
"loss": 2.7496,
|
| 463 |
+
"step": 336000
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 2.74,
|
| 467 |
+
"eval_loss": 2.752856969833374,
|
| 468 |
+
"eval_runtime": 128.3684,
|
| 469 |
+
"eval_samples_per_second": 804.521,
|
| 470 |
+
"eval_steps_per_second": 50.285,
|
| 471 |
+
"step": 336000
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 2.8,
|
| 475 |
+
"eval_loss": 2.749178171157837,
|
| 476 |
+
"eval_runtime": 129.8422,
|
| 477 |
+
"eval_samples_per_second": 795.388,
|
| 478 |
+
"eval_steps_per_second": 49.714,
|
| 479 |
+
"step": 344000
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 2.87,
|
| 483 |
+
"learning_rate": 8.538314016509632e-06,
|
| 484 |
+
"loss": 2.7521,
|
| 485 |
+
"step": 352000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 2.87,
|
| 489 |
+
"eval_loss": 2.761200189590454,
|
| 490 |
+
"eval_runtime": 127.6309,
|
| 491 |
+
"eval_samples_per_second": 809.169,
|
| 492 |
+
"eval_steps_per_second": 50.576,
|
| 493 |
+
"step": 352000
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 2.94,
|
| 497 |
+
"eval_loss": 2.7700963020324707,
|
| 498 |
+
"eval_runtime": 128.3946,
|
| 499 |
+
"eval_samples_per_second": 804.356,
|
| 500 |
+
"eval_steps_per_second": 50.275,
|
| 501 |
+
"step": 360000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 3.0,
|
| 505 |
+
"learning_rate": 8.471608438255649e-06,
|
| 506 |
+
"loss": 2.7649,
|
| 507 |
+
"step": 368000
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 3.0,
|
| 511 |
+
"eval_loss": 2.7705161571502686,
|
| 512 |
+
"eval_runtime": 128.8577,
|
| 513 |
+
"eval_samples_per_second": 801.466,
|
| 514 |
+
"eval_steps_per_second": 50.094,
|
| 515 |
+
"step": 368000
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 3.07,
|
| 519 |
+
"eval_loss": 2.782761335372925,
|
| 520 |
+
"eval_runtime": 129.17,
|
| 521 |
+
"eval_samples_per_second": 799.528,
|
| 522 |
+
"eval_steps_per_second": 49.973,
|
| 523 |
+
"step": 376000
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 3.13,
|
| 527 |
+
"learning_rate": 8.404902860001667e-06,
|
| 528 |
+
"loss": 2.7516,
|
| 529 |
+
"step": 384000
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 3.13,
|
| 533 |
+
"eval_loss": 2.7680482864379883,
|
| 534 |
+
"eval_runtime": 128.9028,
|
| 535 |
+
"eval_samples_per_second": 801.185,
|
| 536 |
+
"eval_steps_per_second": 50.077,
|
| 537 |
+
"step": 384000
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 3.2,
|
| 541 |
+
"eval_loss": 2.784294605255127,
|
| 542 |
+
"eval_runtime": 128.4737,
|
| 543 |
+
"eval_samples_per_second": 803.861,
|
| 544 |
+
"eval_steps_per_second": 50.244,
|
| 545 |
+
"step": 392000
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 3.26,
|
| 549 |
+
"learning_rate": 8.338197281747686e-06,
|
| 550 |
+
"loss": 2.762,
|
| 551 |
+
"step": 400000
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 3.26,
|
| 555 |
+
"eval_loss": 2.7915961742401123,
|
| 556 |
+
"eval_runtime": 128.2651,
|
| 557 |
+
"eval_samples_per_second": 805.168,
|
| 558 |
+
"eval_steps_per_second": 50.325,
|
| 559 |
+
"step": 400000
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 3.33,
|
| 563 |
+
"eval_loss": 2.7691826820373535,
|
| 564 |
+
"eval_runtime": 128.6705,
|
| 565 |
+
"eval_samples_per_second": 802.632,
|
| 566 |
+
"eval_steps_per_second": 50.167,
|
| 567 |
+
"step": 408000
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 3.39,
|
| 571 |
+
"learning_rate": 8.271491703493705e-06,
|
| 572 |
+
"loss": 2.7789,
|
| 573 |
+
"step": 416000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 3.39,
|
| 577 |
+
"eval_loss": 2.783369302749634,
|
| 578 |
+
"eval_runtime": 128.6603,
|
| 579 |
+
"eval_samples_per_second": 802.695,
|
| 580 |
+
"eval_steps_per_second": 50.171,
|
| 581 |
+
"step": 416000
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 3.46,
|
| 585 |
+
"eval_loss": 2.7788405418395996,
|
| 586 |
+
"eval_runtime": 129.7209,
|
| 587 |
+
"eval_samples_per_second": 796.132,
|
| 588 |
+
"eval_steps_per_second": 49.761,
|
| 589 |
+
"step": 424000
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 3.52,
|
| 593 |
+
"learning_rate": 8.204786125239725e-06,
|
| 594 |
+
"loss": 2.7879,
|
| 595 |
+
"step": 432000
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 3.52,
|
| 599 |
+
"eval_loss": 2.803699493408203,
|
| 600 |
+
"eval_runtime": 128.2575,
|
| 601 |
+
"eval_samples_per_second": 805.216,
|
| 602 |
+
"eval_steps_per_second": 50.328,
|
| 603 |
+
"step": 432000
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 3.59,
|
| 607 |
+
"eval_loss": 2.791905403137207,
|
| 608 |
+
"eval_runtime": 129.4159,
|
| 609 |
+
"eval_samples_per_second": 798.009,
|
| 610 |
+
"eval_steps_per_second": 49.878,
|
| 611 |
+
"step": 440000
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 3.65,
|
| 615 |
+
"learning_rate": 8.138080546985743e-06,
|
| 616 |
+
"loss": 2.7853,
|
| 617 |
+
"step": 448000
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 3.65,
|
| 621 |
+
"eval_loss": 2.8077127933502197,
|
| 622 |
+
"eval_runtime": 127.9753,
|
| 623 |
+
"eval_samples_per_second": 806.992,
|
| 624 |
+
"eval_steps_per_second": 50.439,
|
| 625 |
+
"step": 448000
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 3.72,
|
| 629 |
+
"eval_loss": 2.7903032302856445,
|
| 630 |
+
"eval_runtime": 128.9005,
|
| 631 |
+
"eval_samples_per_second": 801.2,
|
| 632 |
+
"eval_steps_per_second": 50.077,
|
| 633 |
+
"step": 456000
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 3.78,
|
| 637 |
+
"learning_rate": 8.07137496873176e-06,
|
| 638 |
+
"loss": 2.7976,
|
| 639 |
+
"step": 464000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 3.78,
|
| 643 |
+
"eval_loss": 2.810896158218384,
|
| 644 |
+
"eval_runtime": 129.0626,
|
| 645 |
+
"eval_samples_per_second": 800.193,
|
| 646 |
+
"eval_steps_per_second": 50.014,
|
| 647 |
+
"step": 464000
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 3.85,
|
| 651 |
+
"eval_loss": 2.795713424682617,
|
| 652 |
+
"eval_runtime": 128.0638,
|
| 653 |
+
"eval_samples_per_second": 806.434,
|
| 654 |
+
"eval_steps_per_second": 50.405,
|
| 655 |
+
"step": 472000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 3.91,
|
| 659 |
+
"learning_rate": 8.004669390477779e-06,
|
| 660 |
+
"loss": 2.789,
|
| 661 |
+
"step": 480000
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 3.91,
|
| 665 |
+
"eval_loss": 2.8023178577423096,
|
| 666 |
+
"eval_runtime": 128.1962,
|
| 667 |
+
"eval_samples_per_second": 805.601,
|
| 668 |
+
"eval_steps_per_second": 50.353,
|
| 669 |
+
"step": 480000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 3.98,
|
| 673 |
+
"eval_loss": 2.8125839233398438,
|
| 674 |
+
"eval_runtime": 128.7992,
|
| 675 |
+
"eval_samples_per_second": 801.83,
|
| 676 |
+
"eval_steps_per_second": 50.117,
|
| 677 |
+
"step": 488000
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 4.04,
|
| 681 |
+
"learning_rate": 7.937963812223798e-06,
|
| 682 |
+
"loss": 2.8089,
|
| 683 |
+
"step": 496000
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 4.04,
|
| 687 |
+
"eval_loss": 2.815424919128418,
|
| 688 |
+
"eval_runtime": 128.7985,
|
| 689 |
+
"eval_samples_per_second": 801.834,
|
| 690 |
+
"eval_steps_per_second": 50.117,
|
| 691 |
+
"step": 496000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 4.11,
|
| 695 |
+
"eval_loss": 2.8122923374176025,
|
| 696 |
+
"eval_runtime": 127.4092,
|
| 697 |
+
"eval_samples_per_second": 810.577,
|
| 698 |
+
"eval_steps_per_second": 50.664,
|
| 699 |
+
"step": 504000
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 4.17,
|
| 703 |
+
"learning_rate": 7.871258233969816e-06,
|
| 704 |
+
"loss": 2.7915,
|
| 705 |
+
"step": 512000
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 4.17,
|
| 709 |
+
"eval_loss": 2.8145976066589355,
|
| 710 |
+
"eval_runtime": 128.9266,
|
| 711 |
+
"eval_samples_per_second": 801.037,
|
| 712 |
+
"eval_steps_per_second": 50.067,
|
| 713 |
+
"step": 512000
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 4.24,
|
| 717 |
+
"eval_loss": 2.8249683380126953,
|
| 718 |
+
"eval_runtime": 129.1348,
|
| 719 |
+
"eval_samples_per_second": 799.746,
|
| 720 |
+
"eval_steps_per_second": 49.987,
|
| 721 |
+
"step": 520000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 4.31,
|
| 725 |
+
"learning_rate": 7.804552655715835e-06,
|
| 726 |
+
"loss": 2.8094,
|
| 727 |
+
"step": 528000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 4.31,
|
| 731 |
+
"eval_loss": 2.820560932159424,
|
| 732 |
+
"eval_runtime": 129.6096,
|
| 733 |
+
"eval_samples_per_second": 796.816,
|
| 734 |
+
"eval_steps_per_second": 49.803,
|
| 735 |
+
"step": 528000
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 4.37,
|
| 739 |
+
"eval_loss": 2.818159341812134,
|
| 740 |
+
"eval_runtime": 128.5096,
|
| 741 |
+
"eval_samples_per_second": 803.637,
|
| 742 |
+
"eval_steps_per_second": 50.23,
|
| 743 |
+
"step": 536000
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 4.44,
|
| 747 |
+
"learning_rate": 7.737847077461853e-06,
|
| 748 |
+
"loss": 2.8196,
|
| 749 |
+
"step": 544000
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 4.44,
|
| 753 |
+
"eval_loss": 2.8351361751556396,
|
| 754 |
+
"eval_runtime": 129.1287,
|
| 755 |
+
"eval_samples_per_second": 799.783,
|
| 756 |
+
"eval_steps_per_second": 49.989,
|
| 757 |
+
"step": 544000
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 4.5,
|
| 761 |
+
"eval_loss": 2.839430570602417,
|
| 762 |
+
"eval_runtime": 129.5203,
|
| 763 |
+
"eval_samples_per_second": 797.365,
|
| 764 |
+
"eval_steps_per_second": 49.838,
|
| 765 |
+
"step": 552000
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 4.57,
|
| 769 |
+
"learning_rate": 7.671141499207872e-06,
|
| 770 |
+
"loss": 2.8316,
|
| 771 |
+
"step": 560000
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 4.57,
|
| 775 |
+
"eval_loss": 2.8396623134613037,
|
| 776 |
+
"eval_runtime": 128.6713,
|
| 777 |
+
"eval_samples_per_second": 802.627,
|
| 778 |
+
"eval_steps_per_second": 50.167,
|
| 779 |
+
"step": 560000
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 4.63,
|
| 783 |
+
"eval_loss": 2.8402562141418457,
|
| 784 |
+
"eval_runtime": 128.654,
|
| 785 |
+
"eval_samples_per_second": 802.735,
|
| 786 |
+
"eval_steps_per_second": 50.173,
|
| 787 |
+
"step": 568000
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 4.7,
|
| 791 |
+
"learning_rate": 7.604435920953891e-06,
|
| 792 |
+
"loss": 2.8444,
|
| 793 |
+
"step": 576000
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 4.7,
|
| 797 |
+
"eval_loss": 2.8350980281829834,
|
| 798 |
+
"eval_runtime": 129.3424,
|
| 799 |
+
"eval_samples_per_second": 798.462,
|
| 800 |
+
"eval_steps_per_second": 49.906,
|
| 801 |
+
"step": 576000
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 4.76,
|
| 805 |
+
"eval_loss": 2.8574254512786865,
|
| 806 |
+
"eval_runtime": 129.6206,
|
| 807 |
+
"eval_samples_per_second": 796.748,
|
| 808 |
+
"eval_steps_per_second": 49.799,
|
| 809 |
+
"step": 584000
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 4.83,
|
| 813 |
+
"learning_rate": 7.537730342699909e-06,
|
| 814 |
+
"loss": 2.833,
|
| 815 |
+
"step": 592000
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 4.83,
|
| 819 |
+
"eval_loss": 2.86171293258667,
|
| 820 |
+
"eval_runtime": 129.2684,
|
| 821 |
+
"eval_samples_per_second": 798.919,
|
| 822 |
+
"eval_steps_per_second": 49.935,
|
| 823 |
+
"step": 592000
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 4.89,
|
| 827 |
+
"eval_loss": 2.857750654220581,
|
| 828 |
+
"eval_runtime": 128.5027,
|
| 829 |
+
"eval_samples_per_second": 803.679,
|
| 830 |
+
"eval_steps_per_second": 50.232,
|
| 831 |
+
"step": 600000
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 4.96,
|
| 835 |
+
"learning_rate": 7.471024764445928e-06,
|
| 836 |
+
"loss": 2.839,
|
| 837 |
+
"step": 608000
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 4.96,
|
| 841 |
+
"eval_loss": 2.8577184677124023,
|
| 842 |
+
"eval_runtime": 128.7081,
|
| 843 |
+
"eval_samples_per_second": 802.397,
|
| 844 |
+
"eval_steps_per_second": 50.152,
|
| 845 |
+
"step": 608000
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 5.02,
|
| 849 |
+
"eval_loss": 2.8726649284362793,
|
| 850 |
+
"eval_runtime": 128.6474,
|
| 851 |
+
"eval_samples_per_second": 802.776,
|
| 852 |
+
"eval_steps_per_second": 50.176,
|
| 853 |
+
"step": 616000
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 5.09,
|
| 857 |
+
"learning_rate": 7.4043191861919465e-06,
|
| 858 |
+
"loss": 2.8427,
|
| 859 |
+
"step": 624000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 5.09,
|
| 863 |
+
"eval_loss": 2.858550786972046,
|
| 864 |
+
"eval_runtime": 129.0947,
|
| 865 |
+
"eval_samples_per_second": 799.994,
|
| 866 |
+
"eval_steps_per_second": 50.002,
|
| 867 |
+
"step": 624000
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 5.15,
|
| 871 |
+
"eval_loss": 2.880849599838257,
|
| 872 |
+
"eval_runtime": 128.221,
|
| 873 |
+
"eval_samples_per_second": 805.445,
|
| 874 |
+
"eval_steps_per_second": 50.343,
|
| 875 |
+
"step": 632000
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 5.22,
|
| 879 |
+
"learning_rate": 7.337613607937964e-06,
|
| 880 |
+
"loss": 2.8599,
|
| 881 |
+
"step": 640000
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 5.22,
|
| 885 |
+
"eval_loss": 2.8959789276123047,
|
| 886 |
+
"eval_runtime": 129.9831,
|
| 887 |
+
"eval_samples_per_second": 794.527,
|
| 888 |
+
"eval_steps_per_second": 49.66,
|
| 889 |
+
"step": 640000
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 5.28,
|
| 893 |
+
"eval_loss": 2.8883421421051025,
|
| 894 |
+
"eval_runtime": 129.4941,
|
| 895 |
+
"eval_samples_per_second": 797.527,
|
| 896 |
+
"eval_steps_per_second": 49.848,
|
| 897 |
+
"step": 648000
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 5.35,
|
| 901 |
+
"learning_rate": 7.270908029683983e-06,
|
| 902 |
+
"loss": 2.8694,
|
| 903 |
+
"step": 656000
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 5.35,
|
| 907 |
+
"eval_loss": 2.8884825706481934,
|
| 908 |
+
"eval_runtime": 129.3172,
|
| 909 |
+
"eval_samples_per_second": 798.618,
|
| 910 |
+
"eval_steps_per_second": 49.916,
|
| 911 |
+
"step": 656000
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 5.41,
|
| 915 |
+
"eval_loss": 2.887291431427002,
|
| 916 |
+
"eval_runtime": 129.1298,
|
| 917 |
+
"eval_samples_per_second": 799.777,
|
| 918 |
+
"eval_steps_per_second": 49.988,
|
| 919 |
+
"step": 664000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 5.48,
|
| 923 |
+
"learning_rate": 7.2042024514300015e-06,
|
| 924 |
+
"loss": 2.8626,
|
| 925 |
+
"step": 672000
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 5.48,
|
| 929 |
+
"eval_loss": 2.8929550647735596,
|
| 930 |
+
"eval_runtime": 129.4886,
|
| 931 |
+
"eval_samples_per_second": 797.56,
|
| 932 |
+
"eval_steps_per_second": 49.85,
|
| 933 |
+
"step": 672000
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 5.54,
|
| 937 |
+
"eval_loss": 2.8987772464752197,
|
| 938 |
+
"eval_runtime": 129.8683,
|
| 939 |
+
"eval_samples_per_second": 795.229,
|
| 940 |
+
"eval_steps_per_second": 49.704,
|
| 941 |
+
"step": 680000
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 5.61,
|
| 945 |
+
"learning_rate": 7.13749687317602e-06,
|
| 946 |
+
"loss": 2.8921,
|
| 947 |
+
"step": 688000
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 5.61,
|
| 951 |
+
"eval_loss": 2.9117259979248047,
|
| 952 |
+
"eval_runtime": 128.3205,
|
| 953 |
+
"eval_samples_per_second": 804.821,
|
| 954 |
+
"eval_steps_per_second": 50.304,
|
| 955 |
+
"step": 688000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 5.68,
|
| 959 |
+
"eval_loss": 2.912231206893921,
|
| 960 |
+
"eval_runtime": 128.7871,
|
| 961 |
+
"eval_samples_per_second": 801.905,
|
| 962 |
+
"eval_steps_per_second": 50.121,
|
| 963 |
+
"step": 696000
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 5.74,
|
| 967 |
+
"learning_rate": 7.070791294922038e-06,
|
| 968 |
+
"loss": 2.8884,
|
| 969 |
+
"step": 704000
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 5.74,
|
| 973 |
+
"eval_loss": 2.900118827819824,
|
| 974 |
+
"eval_runtime": 130.1834,
|
| 975 |
+
"eval_samples_per_second": 793.304,
|
| 976 |
+
"eval_steps_per_second": 49.584,
|
| 977 |
+
"step": 704000
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 5.81,
|
| 981 |
+
"eval_loss": 2.9093644618988037,
|
| 982 |
+
"eval_runtime": 129.4918,
|
| 983 |
+
"eval_samples_per_second": 797.541,
|
| 984 |
+
"eval_steps_per_second": 49.849,
|
| 985 |
+
"step": 712000
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 5.87,
|
| 989 |
+
"learning_rate": 7.0040857166680564e-06,
|
| 990 |
+
"loss": 2.8974,
|
| 991 |
+
"step": 720000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 5.87,
|
| 995 |
+
"eval_loss": 2.9110264778137207,
|
| 996 |
+
"eval_runtime": 129.9051,
|
| 997 |
+
"eval_samples_per_second": 795.003,
|
| 998 |
+
"eval_steps_per_second": 49.69,
|
| 999 |
+
"step": 720000
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 5.94,
|
| 1003 |
+
"eval_loss": 2.9044594764709473,
|
| 1004 |
+
"eval_runtime": 129.2324,
|
| 1005 |
+
"eval_samples_per_second": 799.141,
|
| 1006 |
+
"eval_steps_per_second": 49.949,
|
| 1007 |
+
"step": 728000
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 6.0,
|
| 1011 |
+
"learning_rate": 6.937380138414076e-06,
|
| 1012 |
+
"loss": 2.903,
|
| 1013 |
+
"step": 736000
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 6.0,
|
| 1017 |
+
"eval_loss": 2.933678388595581,
|
| 1018 |
+
"eval_runtime": 130.3644,
|
| 1019 |
+
"eval_samples_per_second": 792.202,
|
| 1020 |
+
"eval_steps_per_second": 49.515,
|
| 1021 |
+
"step": 736000
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 6.07,
|
| 1025 |
+
"eval_loss": 2.931581735610962,
|
| 1026 |
+
"eval_runtime": 128.3976,
|
| 1027 |
+
"eval_samples_per_second": 804.337,
|
| 1028 |
+
"eval_steps_per_second": 50.274,
|
| 1029 |
+
"step": 744000
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 6.13,
|
| 1033 |
+
"learning_rate": 6.8706745601600945e-06,
|
| 1034 |
+
"loss": 2.9057,
|
| 1035 |
+
"step": 752000
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 6.13,
|
| 1039 |
+
"eval_loss": 2.944746971130371,
|
| 1040 |
+
"eval_runtime": 128.9912,
|
| 1041 |
+
"eval_samples_per_second": 800.636,
|
| 1042 |
+
"eval_steps_per_second": 50.042,
|
| 1043 |
+
"step": 752000
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 6.2,
|
| 1047 |
+
"eval_loss": 2.936281681060791,
|
| 1048 |
+
"eval_runtime": 129.9533,
|
| 1049 |
+
"eval_samples_per_second": 794.709,
|
| 1050 |
+
"eval_steps_per_second": 49.672,
|
| 1051 |
+
"step": 760000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 6.26,
|
| 1055 |
+
"learning_rate": 6.803968981906113e-06,
|
| 1056 |
+
"loss": 2.9146,
|
| 1057 |
+
"step": 768000
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 6.26,
|
| 1061 |
+
"eval_loss": 2.943751096725464,
|
| 1062 |
+
"eval_runtime": 129.9494,
|
| 1063 |
+
"eval_samples_per_second": 794.732,
|
| 1064 |
+
"eval_steps_per_second": 49.673,
|
| 1065 |
+
"step": 768000
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 6.33,
|
| 1069 |
+
"eval_loss": 2.9474806785583496,
|
| 1070 |
+
"eval_runtime": 130.0993,
|
| 1071 |
+
"eval_samples_per_second": 793.817,
|
| 1072 |
+
"eval_steps_per_second": 49.616,
|
| 1073 |
+
"step": 776000
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 6.39,
|
| 1077 |
+
"learning_rate": 6.737263403652131e-06,
|
| 1078 |
+
"loss": 2.9221,
|
| 1079 |
+
"step": 784000
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 6.39,
|
| 1083 |
+
"eval_loss": 2.9394171237945557,
|
| 1084 |
+
"eval_runtime": 129.1928,
|
| 1085 |
+
"eval_samples_per_second": 799.387,
|
| 1086 |
+
"eval_steps_per_second": 49.964,
|
| 1087 |
+
"step": 784000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 6.46,
|
| 1091 |
+
"eval_loss": 2.937087297439575,
|
| 1092 |
+
"eval_runtime": 129.9118,
|
| 1093 |
+
"eval_samples_per_second": 794.963,
|
| 1094 |
+
"eval_steps_per_second": 49.688,
|
| 1095 |
+
"step": 792000
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 6.52,
|
| 1099 |
+
"learning_rate": 6.6705578253981495e-06,
|
| 1100 |
+
"loss": 2.9316,
|
| 1101 |
+
"step": 800000
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 6.52,
|
| 1105 |
+
"eval_loss": 2.949429512023926,
|
| 1106 |
+
"eval_runtime": 129.8602,
|
| 1107 |
+
"eval_samples_per_second": 795.278,
|
| 1108 |
+
"eval_steps_per_second": 49.707,
|
| 1109 |
+
"step": 800000
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 6.59,
|
| 1113 |
+
"eval_loss": 2.9727399349212646,
|
| 1114 |
+
"eval_runtime": 130.9441,
|
| 1115 |
+
"eval_samples_per_second": 788.695,
|
| 1116 |
+
"eval_steps_per_second": 49.296,
|
| 1117 |
+
"step": 808000
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 6.65,
|
| 1121 |
+
"learning_rate": 6.603852247144168e-06,
|
| 1122 |
+
"loss": 2.9421,
|
| 1123 |
+
"step": 816000
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 6.65,
|
| 1127 |
+
"eval_loss": 2.9758830070495605,
|
| 1128 |
+
"eval_runtime": 129.8861,
|
| 1129 |
+
"eval_samples_per_second": 795.12,
|
| 1130 |
+
"eval_steps_per_second": 49.697,
|
| 1131 |
+
"step": 816000
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 6.72,
|
| 1135 |
+
"eval_loss": 2.966480016708374,
|
| 1136 |
+
"eval_runtime": 129.44,
|
| 1137 |
+
"eval_samples_per_second": 797.86,
|
| 1138 |
+
"eval_steps_per_second": 49.869,
|
| 1139 |
+
"step": 824000
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 6.78,
|
| 1143 |
+
"learning_rate": 6.537146668890187e-06,
|
| 1144 |
+
"loss": 2.9538,
|
| 1145 |
+
"step": 832000
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 6.78,
|
| 1149 |
+
"eval_loss": 2.9650251865386963,
|
| 1150 |
+
"eval_runtime": 129.4919,
|
| 1151 |
+
"eval_samples_per_second": 797.54,
|
| 1152 |
+
"eval_steps_per_second": 49.849,
|
| 1153 |
+
"step": 832000
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 6.85,
|
| 1157 |
+
"eval_loss": 2.976144313812256,
|
| 1158 |
+
"eval_runtime": 129.8294,
|
| 1159 |
+
"eval_samples_per_second": 795.467,
|
| 1160 |
+
"eval_steps_per_second": 49.719,
|
| 1161 |
+
"step": 840000
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 6.91,
|
| 1165 |
+
"learning_rate": 6.4704410906362044e-06,
|
| 1166 |
+
"loss": 2.9594,
|
| 1167 |
+
"step": 848000
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 6.91,
|
| 1171 |
+
"eval_loss": 2.990086317062378,
|
| 1172 |
+
"eval_runtime": 129.827,
|
| 1173 |
+
"eval_samples_per_second": 795.482,
|
| 1174 |
+
"eval_steps_per_second": 49.72,
|
| 1175 |
+
"step": 848000
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 6.98,
|
| 1179 |
+
"eval_loss": 2.973181962966919,
|
| 1180 |
+
"eval_runtime": 131.5126,
|
| 1181 |
+
"eval_samples_per_second": 785.286,
|
| 1182 |
+
"eval_steps_per_second": 49.083,
|
| 1183 |
+
"step": 856000
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 7.05,
|
| 1187 |
+
"learning_rate": 6.403735512382223e-06,
|
| 1188 |
+
"loss": 2.9564,
|
| 1189 |
+
"step": 864000
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 7.05,
|
| 1193 |
+
"eval_loss": 2.9896528720855713,
|
| 1194 |
+
"eval_runtime": 129.878,
|
| 1195 |
+
"eval_samples_per_second": 795.169,
|
| 1196 |
+
"eval_steps_per_second": 49.7,
|
| 1197 |
+
"step": 864000
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 7.11,
|
| 1201 |
+
"eval_loss": 2.980059862136841,
|
| 1202 |
+
"eval_runtime": 129.5351,
|
| 1203 |
+
"eval_samples_per_second": 797.274,
|
| 1204 |
+
"eval_steps_per_second": 49.832,
|
| 1205 |
+
"step": 872000
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 7.18,
|
| 1209 |
+
"learning_rate": 6.337029934128242e-06,
|
| 1210 |
+
"loss": 2.9561,
|
| 1211 |
+
"step": 880000
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 7.18,
|
| 1215 |
+
"eval_loss": 2.983869791030884,
|
| 1216 |
+
"eval_runtime": 130.0357,
|
| 1217 |
+
"eval_samples_per_second": 794.205,
|
| 1218 |
+
"eval_steps_per_second": 49.64,
|
| 1219 |
+
"step": 880000
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 7.24,
|
| 1223 |
+
"eval_loss": 2.9887585639953613,
|
| 1224 |
+
"eval_runtime": 130.015,
|
| 1225 |
+
"eval_samples_per_second": 794.331,
|
| 1226 |
+
"eval_steps_per_second": 49.648,
|
| 1227 |
+
"step": 888000
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 7.31,
|
| 1231 |
+
"learning_rate": 6.270324355874261e-06,
|
| 1232 |
+
"loss": 2.9669,
|
| 1233 |
+
"step": 896000
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 7.31,
|
| 1237 |
+
"eval_loss": 2.99999737739563,
|
| 1238 |
+
"eval_runtime": 130.6345,
|
| 1239 |
+
"eval_samples_per_second": 790.564,
|
| 1240 |
+
"eval_steps_per_second": 49.413,
|
| 1241 |
+
"step": 896000
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 7.37,
|
| 1245 |
+
"eval_loss": 2.9786183834075928,
|
| 1246 |
+
"eval_runtime": 129.9739,
|
| 1247 |
+
"eval_samples_per_second": 794.582,
|
| 1248 |
+
"eval_steps_per_second": 49.664,
|
| 1249 |
+
"step": 904000
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 7.44,
|
| 1253 |
+
"learning_rate": 6.20361877762028e-06,
|
| 1254 |
+
"loss": 2.9649,
|
| 1255 |
+
"step": 912000
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 7.44,
|
| 1259 |
+
"eval_loss": 2.994581460952759,
|
| 1260 |
+
"eval_runtime": 131.0156,
|
| 1261 |
+
"eval_samples_per_second": 788.265,
|
| 1262 |
+
"eval_steps_per_second": 49.269,
|
| 1263 |
+
"step": 912000
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 7.5,
|
| 1267 |
+
"eval_loss": 3.0002031326293945,
|
| 1268 |
+
"eval_runtime": 131.7355,
|
| 1269 |
+
"eval_samples_per_second": 783.957,
|
| 1270 |
+
"eval_steps_per_second": 49.0,
|
| 1271 |
+
"step": 920000
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 7.57,
|
| 1275 |
+
"learning_rate": 6.1369131993662975e-06,
|
| 1276 |
+
"loss": 2.9665,
|
| 1277 |
+
"step": 928000
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 7.57,
|
| 1281 |
+
"eval_loss": 2.9960474967956543,
|
| 1282 |
+
"eval_runtime": 131.6559,
|
| 1283 |
+
"eval_samples_per_second": 784.431,
|
| 1284 |
+
"eval_steps_per_second": 49.029,
|
| 1285 |
+
"step": 928000
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 7.63,
|
| 1289 |
+
"eval_loss": 3.0067989826202393,
|
| 1290 |
+
"eval_runtime": 131.8152,
|
| 1291 |
+
"eval_samples_per_second": 783.483,
|
| 1292 |
+
"eval_steps_per_second": 48.97,
|
| 1293 |
+
"step": 936000
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 7.7,
|
| 1297 |
+
"learning_rate": 6.070207621112316e-06,
|
| 1298 |
+
"loss": 2.9708,
|
| 1299 |
+
"step": 944000
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 7.7,
|
| 1303 |
+
"eval_loss": 2.993788242340088,
|
| 1304 |
+
"eval_runtime": 130.3799,
|
| 1305 |
+
"eval_samples_per_second": 792.108,
|
| 1306 |
+
"eval_steps_per_second": 49.509,
|
| 1307 |
+
"step": 944000
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 7.76,
|
| 1311 |
+
"eval_loss": 3.0126230716705322,
|
| 1312 |
+
"eval_runtime": 130.4447,
|
| 1313 |
+
"eval_samples_per_second": 791.715,
|
| 1314 |
+
"eval_steps_per_second": 49.485,
|
| 1315 |
+
"step": 952000
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 7.83,
|
| 1319 |
+
"learning_rate": 6.003502042858335e-06,
|
| 1320 |
+
"loss": 2.981,
|
| 1321 |
+
"step": 960000
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 7.83,
|
| 1325 |
+
"eval_loss": 2.9959194660186768,
|
| 1326 |
+
"eval_runtime": 132.0738,
|
| 1327 |
+
"eval_samples_per_second": 781.949,
|
| 1328 |
+
"eval_steps_per_second": 48.874,
|
| 1329 |
+
"step": 960000
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 7.89,
|
| 1333 |
+
"eval_loss": 2.995976448059082,
|
| 1334 |
+
"eval_runtime": 130.9412,
|
| 1335 |
+
"eval_samples_per_second": 788.713,
|
| 1336 |
+
"eval_steps_per_second": 49.297,
|
| 1337 |
+
"step": 968000
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 7.96,
|
| 1341 |
+
"learning_rate": 5.936796464604353e-06,
|
| 1342 |
+
"loss": 2.9805,
|
| 1343 |
+
"step": 976000
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 7.96,
|
| 1347 |
+
"eval_loss": 2.991947889328003,
|
| 1348 |
+
"eval_runtime": 130.0819,
|
| 1349 |
+
"eval_samples_per_second": 793.923,
|
| 1350 |
+
"eval_steps_per_second": 49.623,
|
| 1351 |
+
"step": 976000
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 8.02,
|
| 1355 |
+
"eval_loss": 3.0058255195617676,
|
| 1356 |
+
"eval_runtime": 130.7007,
|
| 1357 |
+
"eval_samples_per_second": 790.164,
|
| 1358 |
+
"eval_steps_per_second": 49.388,
|
| 1359 |
+
"step": 984000
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 8.09,
|
| 1363 |
+
"learning_rate": 5.870090886350371e-06,
|
| 1364 |
+
"loss": 2.9705,
|
| 1365 |
+
"step": 992000
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 8.09,
|
| 1369 |
+
"eval_loss": 3.0232017040252686,
|
| 1370 |
+
"eval_runtime": 129.9163,
|
| 1371 |
+
"eval_samples_per_second": 794.935,
|
| 1372 |
+
"eval_steps_per_second": 49.686,
|
| 1373 |
+
"step": 992000
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 8.15,
|
| 1377 |
+
"eval_loss": 3.0046939849853516,
|
| 1378 |
+
"eval_runtime": 130.7903,
|
| 1379 |
+
"eval_samples_per_second": 789.623,
|
| 1380 |
+
"eval_steps_per_second": 49.354,
|
| 1381 |
+
"step": 1000000
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 8.22,
|
| 1385 |
+
"learning_rate": 5.80338530809639e-06,
|
| 1386 |
+
"loss": 2.9715,
|
| 1387 |
+
"step": 1008000
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 8.22,
|
| 1391 |
+
"eval_loss": 3.0068600177764893,
|
| 1392 |
+
"eval_runtime": 131.6119,
|
| 1393 |
+
"eval_samples_per_second": 784.693,
|
| 1394 |
+
"eval_steps_per_second": 49.046,
|
| 1395 |
+
"step": 1008000
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 8.28,
|
| 1399 |
+
"eval_loss": 3.0018742084503174,
|
| 1400 |
+
"eval_runtime": 131.7567,
|
| 1401 |
+
"eval_samples_per_second": 783.831,
|
| 1402 |
+
"eval_steps_per_second": 48.992,
|
| 1403 |
+
"step": 1016000
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 8.35,
|
| 1407 |
+
"learning_rate": 5.736679729842408e-06,
|
| 1408 |
+
"loss": 2.9695,
|
| 1409 |
+
"step": 1024000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 8.35,
|
| 1413 |
+
"eval_loss": 3.021596670150757,
|
| 1414 |
+
"eval_runtime": 131.2334,
|
| 1415 |
+
"eval_samples_per_second": 786.956,
|
| 1416 |
+
"eval_steps_per_second": 49.187,
|
| 1417 |
+
"step": 1024000
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 8.41,
|
| 1421 |
+
"eval_loss": 3.0219063758850098,
|
| 1422 |
+
"eval_runtime": 131.6228,
|
| 1423 |
+
"eval_samples_per_second": 784.629,
|
| 1424 |
+
"eval_steps_per_second": 49.042,
|
| 1425 |
+
"step": 1032000
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 8.48,
|
| 1429 |
+
"learning_rate": 5.669974151588427e-06,
|
| 1430 |
+
"loss": 2.9762,
|
| 1431 |
+
"step": 1040000
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 8.48,
|
| 1435 |
+
"eval_loss": 3.018242597579956,
|
| 1436 |
+
"eval_runtime": 131.898,
|
| 1437 |
+
"eval_samples_per_second": 782.991,
|
| 1438 |
+
"eval_steps_per_second": 48.939,
|
| 1439 |
+
"step": 1040000
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 8.55,
|
| 1443 |
+
"eval_loss": 3.0332210063934326,
|
| 1444 |
+
"eval_runtime": 132.3771,
|
| 1445 |
+
"eval_samples_per_second": 780.158,
|
| 1446 |
+
"eval_steps_per_second": 48.762,
|
| 1447 |
+
"step": 1048000
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 8.61,
|
| 1451 |
+
"learning_rate": 5.603268573334446e-06,
|
| 1452 |
+
"loss": 2.9786,
|
| 1453 |
+
"step": 1056000
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 8.61,
|
| 1457 |
+
"eval_loss": 3.001666307449341,
|
| 1458 |
+
"eval_runtime": 131.4368,
|
| 1459 |
+
"eval_samples_per_second": 785.739,
|
| 1460 |
+
"eval_steps_per_second": 49.111,
|
| 1461 |
+
"step": 1056000
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 8.68,
|
| 1465 |
+
"eval_loss": 3.0236458778381348,
|
| 1466 |
+
"eval_runtime": 130.9562,
|
| 1467 |
+
"eval_samples_per_second": 788.622,
|
| 1468 |
+
"eval_steps_per_second": 49.291,
|
| 1469 |
+
"step": 1064000
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 8.74,
|
| 1473 |
+
"learning_rate": 5.536562995080464e-06,
|
| 1474 |
+
"loss": 2.9889,
|
| 1475 |
+
"step": 1072000
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 8.74,
|
| 1479 |
+
"eval_loss": 3.0273077487945557,
|
| 1480 |
+
"eval_runtime": 131.9047,
|
| 1481 |
+
"eval_samples_per_second": 782.952,
|
| 1482 |
+
"eval_steps_per_second": 48.937,
|
| 1483 |
+
"step": 1072000
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 8.81,
|
| 1487 |
+
"eval_loss": 3.01967191696167,
|
| 1488 |
+
"eval_runtime": 131.9615,
|
| 1489 |
+
"eval_samples_per_second": 782.615,
|
| 1490 |
+
"eval_steps_per_second": 48.916,
|
| 1491 |
+
"step": 1080000
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 8.87,
|
| 1495 |
+
"learning_rate": 5.469857416826483e-06,
|
| 1496 |
+
"loss": 2.9842,
|
| 1497 |
+
"step": 1088000
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 8.87,
|
| 1501 |
+
"eval_loss": 3.037600040435791,
|
| 1502 |
+
"eval_runtime": 131.9507,
|
| 1503 |
+
"eval_samples_per_second": 782.679,
|
| 1504 |
+
"eval_steps_per_second": 48.92,
|
| 1505 |
+
"step": 1088000
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 8.94,
|
| 1509 |
+
"eval_loss": 3.032285213470459,
|
| 1510 |
+
"eval_runtime": 131.7234,
|
| 1511 |
+
"eval_samples_per_second": 784.029,
|
| 1512 |
+
"eval_steps_per_second": 49.004,
|
| 1513 |
+
"step": 1096000
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 9.0,
|
| 1517 |
+
"learning_rate": 5.403151838572501e-06,
|
| 1518 |
+
"loss": 2.9912,
|
| 1519 |
+
"step": 1104000
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 9.0,
|
| 1523 |
+
"eval_loss": 3.031731367111206,
|
| 1524 |
+
"eval_runtime": 131.8868,
|
| 1525 |
+
"eval_samples_per_second": 783.058,
|
| 1526 |
+
"eval_steps_per_second": 48.944,
|
| 1527 |
+
"step": 1104000
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 9.07,
|
| 1531 |
+
"eval_loss": 3.022475481033325,
|
| 1532 |
+
"eval_runtime": 131.0568,
|
| 1533 |
+
"eval_samples_per_second": 788.017,
|
| 1534 |
+
"eval_steps_per_second": 49.253,
|
| 1535 |
+
"step": 1112000
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 9.13,
|
| 1539 |
+
"learning_rate": 5.33644626031852e-06,
|
| 1540 |
+
"loss": 2.9919,
|
| 1541 |
+
"step": 1120000
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 9.13,
|
| 1545 |
+
"eval_loss": 3.036106824874878,
|
| 1546 |
+
"eval_runtime": 132.2182,
|
| 1547 |
+
"eval_samples_per_second": 781.095,
|
| 1548 |
+
"eval_steps_per_second": 48.821,
|
| 1549 |
+
"step": 1120000
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 9.2,
|
| 1553 |
+
"eval_loss": 3.0432300567626953,
|
| 1554 |
+
"eval_runtime": 131.9088,
|
| 1555 |
+
"eval_samples_per_second": 782.927,
|
| 1556 |
+
"eval_steps_per_second": 48.935,
|
| 1557 |
+
"step": 1128000
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 9.26,
|
| 1561 |
+
"learning_rate": 5.269740682064538e-06,
|
| 1562 |
+
"loss": 2.9872,
|
| 1563 |
+
"step": 1136000
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 9.26,
|
| 1567 |
+
"eval_loss": 3.0306613445281982,
|
| 1568 |
+
"eval_runtime": 131.2348,
|
| 1569 |
+
"eval_samples_per_second": 786.948,
|
| 1570 |
+
"eval_steps_per_second": 49.187,
|
| 1571 |
+
"step": 1136000
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 9.33,
|
| 1575 |
+
"eval_loss": 3.0481879711151123,
|
| 1576 |
+
"eval_runtime": 131.7205,
|
| 1577 |
+
"eval_samples_per_second": 784.046,
|
| 1578 |
+
"eval_steps_per_second": 49.005,
|
| 1579 |
+
"step": 1144000
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 9.39,
|
| 1583 |
+
"learning_rate": 5.203035103810556e-06,
|
| 1584 |
+
"loss": 2.9823,
|
| 1585 |
+
"step": 1152000
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 9.39,
|
| 1589 |
+
"eval_loss": 3.035399913787842,
|
| 1590 |
+
"eval_runtime": 131.2188,
|
| 1591 |
+
"eval_samples_per_second": 787.044,
|
| 1592 |
+
"eval_steps_per_second": 49.193,
|
| 1593 |
+
"step": 1152000
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 9.46,
|
| 1597 |
+
"eval_loss": 3.0419015884399414,
|
| 1598 |
+
"eval_runtime": 131.8024,
|
| 1599 |
+
"eval_samples_per_second": 783.559,
|
| 1600 |
+
"eval_steps_per_second": 48.975,
|
| 1601 |
+
"step": 1160000
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 9.52,
|
| 1605 |
+
"learning_rate": 5.136329525556575e-06,
|
| 1606 |
+
"loss": 2.9882,
|
| 1607 |
+
"step": 1168000
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 9.52,
|
| 1611 |
+
"eval_loss": 3.0567431449890137,
|
| 1612 |
+
"eval_runtime": 132.7773,
|
| 1613 |
+
"eval_samples_per_second": 777.806,
|
| 1614 |
+
"eval_steps_per_second": 48.615,
|
| 1615 |
+
"step": 1168000
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 9.59,
|
| 1619 |
+
"eval_loss": 3.0395400524139404,
|
| 1620 |
+
"eval_runtime": 131.6554,
|
| 1621 |
+
"eval_samples_per_second": 784.434,
|
| 1622 |
+
"eval_steps_per_second": 49.03,
|
| 1623 |
+
"step": 1176000
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 9.65,
|
| 1627 |
+
"learning_rate": 5.0696239473025935e-06,
|
| 1628 |
+
"loss": 3.0079,
|
| 1629 |
+
"step": 1184000
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 9.65,
|
| 1633 |
+
"eval_loss": 3.0572261810302734,
|
| 1634 |
+
"eval_runtime": 132.0184,
|
| 1635 |
+
"eval_samples_per_second": 782.278,
|
| 1636 |
+
"eval_steps_per_second": 48.895,
|
| 1637 |
+
"step": 1184000
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 9.72,
|
| 1641 |
+
"eval_loss": 3.04028058052063,
|
| 1642 |
+
"eval_runtime": 131.8056,
|
| 1643 |
+
"eval_samples_per_second": 783.54,
|
| 1644 |
+
"eval_steps_per_second": 48.974,
|
| 1645 |
+
"step": 1192000
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 9.78,
|
| 1649 |
+
"learning_rate": 5.002918369048611e-06,
|
| 1650 |
+
"loss": 3.0243,
|
| 1651 |
+
"step": 1200000
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 9.78,
|
| 1655 |
+
"eval_loss": 3.047227621078491,
|
| 1656 |
+
"eval_runtime": 131.9863,
|
| 1657 |
+
"eval_samples_per_second": 782.467,
|
| 1658 |
+
"eval_steps_per_second": 48.907,
|
| 1659 |
+
"step": 1200000
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 9.85,
|
| 1663 |
+
"eval_loss": 3.052279472351074,
|
| 1664 |
+
"eval_runtime": 132.2017,
|
| 1665 |
+
"eval_samples_per_second": 781.193,
|
| 1666 |
+
"eval_steps_per_second": 48.827,
|
| 1667 |
+
"step": 1208000
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 9.92,
|
| 1671 |
+
"learning_rate": 4.936212790794631e-06,
|
| 1672 |
+
"loss": 3.0127,
|
| 1673 |
+
"step": 1216000
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 9.92,
|
| 1677 |
+
"eval_loss": 3.053439140319824,
|
| 1678 |
+
"eval_runtime": 131.3363,
|
| 1679 |
+
"eval_samples_per_second": 786.34,
|
| 1680 |
+
"eval_steps_per_second": 49.149,
|
| 1681 |
+
"step": 1216000
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 9.98,
|
| 1685 |
+
"eval_loss": 3.0434141159057617,
|
| 1686 |
+
"eval_runtime": 131.7363,
|
| 1687 |
+
"eval_samples_per_second": 783.952,
|
| 1688 |
+
"eval_steps_per_second": 48.999,
|
| 1689 |
+
"step": 1224000
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 10.05,
|
| 1693 |
+
"learning_rate": 4.869507212540649e-06,
|
| 1694 |
+
"loss": 3.0106,
|
| 1695 |
+
"step": 1232000
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 10.05,
|
| 1699 |
+
"eval_loss": 3.0687036514282227,
|
| 1700 |
+
"eval_runtime": 131.4287,
|
| 1701 |
+
"eval_samples_per_second": 785.788,
|
| 1702 |
+
"eval_steps_per_second": 49.114,
|
| 1703 |
+
"step": 1232000
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 10.11,
|
| 1707 |
+
"eval_loss": 3.0677733421325684,
|
| 1708 |
+
"eval_runtime": 132.6312,
|
| 1709 |
+
"eval_samples_per_second": 778.663,
|
| 1710 |
+
"eval_steps_per_second": 48.669,
|
| 1711 |
+
"step": 1240000
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 10.18,
|
| 1715 |
+
"learning_rate": 4.802801634286667e-06,
|
| 1716 |
+
"loss": 3.0063,
|
| 1717 |
+
"step": 1248000
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 10.18,
|
| 1721 |
+
"eval_loss": 3.0652401447296143,
|
| 1722 |
+
"eval_runtime": 132.5035,
|
| 1723 |
+
"eval_samples_per_second": 779.413,
|
| 1724 |
+
"eval_steps_per_second": 48.716,
|
| 1725 |
+
"step": 1248000
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 10.24,
|
| 1729 |
+
"eval_loss": 3.0768234729766846,
|
| 1730 |
+
"eval_runtime": 131.7104,
|
| 1731 |
+
"eval_samples_per_second": 784.107,
|
| 1732 |
+
"eval_steps_per_second": 49.009,
|
| 1733 |
+
"step": 1256000
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 10.31,
|
| 1737 |
+
"learning_rate": 4.7360960560326865e-06,
|
| 1738 |
+
"loss": 3.0187,
|
| 1739 |
+
"step": 1264000
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 10.31,
|
| 1743 |
+
"eval_loss": 3.069179058074951,
|
| 1744 |
+
"eval_runtime": 132.7895,
|
| 1745 |
+
"eval_samples_per_second": 777.735,
|
| 1746 |
+
"eval_steps_per_second": 48.611,
|
| 1747 |
+
"step": 1264000
|
| 1748 |
+
},
|
| 1749 |
+
{
|
| 1750 |
+
"epoch": 10.37,
|
| 1751 |
+
"eval_loss": 3.0621213912963867,
|
| 1752 |
+
"eval_runtime": 132.041,
|
| 1753 |
+
"eval_samples_per_second": 782.144,
|
| 1754 |
+
"eval_steps_per_second": 48.886,
|
| 1755 |
+
"step": 1272000
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 10.44,
|
| 1759 |
+
"learning_rate": 4.669390477778704e-06,
|
| 1760 |
+
"loss": 3.0202,
|
| 1761 |
+
"step": 1280000
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 10.44,
|
| 1765 |
+
"eval_loss": 3.0663187503814697,
|
| 1766 |
+
"eval_runtime": 132.2635,
|
| 1767 |
+
"eval_samples_per_second": 780.828,
|
| 1768 |
+
"eval_steps_per_second": 48.804,
|
| 1769 |
+
"step": 1280000
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 10.5,
|
| 1773 |
+
"eval_loss": 3.0537171363830566,
|
| 1774 |
+
"eval_runtime": 132.2536,
|
| 1775 |
+
"eval_samples_per_second": 780.886,
|
| 1776 |
+
"eval_steps_per_second": 48.808,
|
| 1777 |
+
"step": 1288000
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"epoch": 10.57,
|
| 1781 |
+
"learning_rate": 4.602684899524723e-06,
|
| 1782 |
+
"loss": 3.0219,
|
| 1783 |
+
"step": 1296000
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 10.57,
|
| 1787 |
+
"eval_loss": 3.072500705718994,
|
| 1788 |
+
"eval_runtime": 132.0295,
|
| 1789 |
+
"eval_samples_per_second": 782.212,
|
| 1790 |
+
"eval_steps_per_second": 48.891,
|
| 1791 |
+
"step": 1296000
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 10.63,
|
| 1795 |
+
"eval_loss": 3.0664169788360596,
|
| 1796 |
+
"eval_runtime": 131.9651,
|
| 1797 |
+
"eval_samples_per_second": 782.593,
|
| 1798 |
+
"eval_steps_per_second": 48.914,
|
| 1799 |
+
"step": 1304000
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 10.7,
|
| 1803 |
+
"learning_rate": 4.5359793212707415e-06,
|
| 1804 |
+
"loss": 3.0232,
|
| 1805 |
+
"step": 1312000
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 10.7,
|
| 1809 |
+
"eval_loss": 3.0724074840545654,
|
| 1810 |
+
"eval_runtime": 133.2104,
|
| 1811 |
+
"eval_samples_per_second": 775.277,
|
| 1812 |
+
"eval_steps_per_second": 48.457,
|
| 1813 |
+
"step": 1312000
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 10.76,
|
| 1817 |
+
"eval_loss": 3.0476126670837402,
|
| 1818 |
+
"eval_runtime": 132.7171,
|
| 1819 |
+
"eval_samples_per_second": 778.159,
|
| 1820 |
+
"eval_steps_per_second": 48.637,
|
| 1821 |
+
"step": 1320000
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 10.83,
|
| 1825 |
+
"learning_rate": 4.46927374301676e-06,
|
| 1826 |
+
"loss": 3.0247,
|
| 1827 |
+
"step": 1328000
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 10.83,
|
| 1831 |
+
"eval_loss": 3.0729353427886963,
|
| 1832 |
+
"eval_runtime": 132.4018,
|
| 1833 |
+
"eval_samples_per_second": 780.012,
|
| 1834 |
+
"eval_steps_per_second": 48.753,
|
| 1835 |
+
"step": 1328000
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 10.89,
|
| 1839 |
+
"eval_loss": 3.0645902156829834,
|
| 1840 |
+
"eval_runtime": 133.3334,
|
| 1841 |
+
"eval_samples_per_second": 774.562,
|
| 1842 |
+
"eval_steps_per_second": 48.412,
|
| 1843 |
+
"step": 1336000
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 10.96,
|
| 1847 |
+
"learning_rate": 4.402568164762779e-06,
|
| 1848 |
+
"loss": 3.0335,
|
| 1849 |
+
"step": 1344000
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 10.96,
|
| 1853 |
+
"eval_loss": 3.0603559017181396,
|
| 1854 |
+
"eval_runtime": 131.9232,
|
| 1855 |
+
"eval_samples_per_second": 782.842,
|
| 1856 |
+
"eval_steps_per_second": 48.93,
|
| 1857 |
+
"step": 1344000
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 11.02,
|
| 1861 |
+
"eval_loss": 3.0630509853363037,
|
| 1862 |
+
"eval_runtime": 132.4502,
|
| 1863 |
+
"eval_samples_per_second": 779.727,
|
| 1864 |
+
"eval_steps_per_second": 48.735,
|
| 1865 |
+
"step": 1352000
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 11.09,
|
| 1869 |
+
"learning_rate": 4.335862586508797e-06,
|
| 1870 |
+
"loss": 3.0182,
|
| 1871 |
+
"step": 1360000
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 11.09,
|
| 1875 |
+
"eval_loss": 3.0669026374816895,
|
| 1876 |
+
"eval_runtime": 133.3499,
|
| 1877 |
+
"eval_samples_per_second": 774.466,
|
| 1878 |
+
"eval_steps_per_second": 48.406,
|
| 1879 |
+
"step": 1360000
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 11.15,
|
| 1883 |
+
"eval_loss": 3.0626471042633057,
|
| 1884 |
+
"eval_runtime": 133.0041,
|
| 1885 |
+
"eval_samples_per_second": 776.48,
|
| 1886 |
+
"eval_steps_per_second": 48.532,
|
| 1887 |
+
"step": 1368000
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 11.22,
|
| 1891 |
+
"learning_rate": 4.269157008254816e-06,
|
| 1892 |
+
"loss": 3.0124,
|
| 1893 |
+
"step": 1376000
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 11.22,
|
| 1897 |
+
"eval_loss": 3.053469181060791,
|
| 1898 |
+
"eval_runtime": 133.5969,
|
| 1899 |
+
"eval_samples_per_second": 773.034,
|
| 1900 |
+
"eval_steps_per_second": 48.317,
|
| 1901 |
+
"step": 1376000
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 11.29,
|
| 1905 |
+
"eval_loss": 3.076792001724243,
|
| 1906 |
+
"eval_runtime": 133.0672,
|
| 1907 |
+
"eval_samples_per_second": 776.112,
|
| 1908 |
+
"eval_steps_per_second": 48.509,
|
| 1909 |
+
"step": 1384000
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 11.35,
|
| 1913 |
+
"learning_rate": 4.202451430000834e-06,
|
| 1914 |
+
"loss": 3.016,
|
| 1915 |
+
"step": 1392000
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 11.35,
|
| 1919 |
+
"eval_loss": 3.0615081787109375,
|
| 1920 |
+
"eval_runtime": 133.9693,
|
| 1921 |
+
"eval_samples_per_second": 770.886,
|
| 1922 |
+
"eval_steps_per_second": 48.183,
|
| 1923 |
+
"step": 1392000
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 11.42,
|
| 1927 |
+
"eval_loss": 3.0689148902893066,
|
| 1928 |
+
"eval_runtime": 134.418,
|
| 1929 |
+
"eval_samples_per_second": 768.312,
|
| 1930 |
+
"eval_steps_per_second": 48.022,
|
| 1931 |
+
"step": 1400000
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 11.48,
|
| 1935 |
+
"learning_rate": 4.135745851746852e-06,
|
| 1936 |
+
"loss": 3.0133,
|
| 1937 |
+
"step": 1408000
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 11.48,
|
| 1941 |
+
"eval_loss": 3.069943428039551,
|
| 1942 |
+
"eval_runtime": 133.7409,
|
| 1943 |
+
"eval_samples_per_second": 772.202,
|
| 1944 |
+
"eval_steps_per_second": 48.265,
|
| 1945 |
+
"step": 1408000
|
| 1946 |
+
},
|
| 1947 |
+
{
|
| 1948 |
+
"epoch": 11.55,
|
| 1949 |
+
"eval_loss": 3.0647213459014893,
|
| 1950 |
+
"eval_runtime": 134.5422,
|
| 1951 |
+
"eval_samples_per_second": 767.603,
|
| 1952 |
+
"eval_steps_per_second": 47.977,
|
| 1953 |
+
"step": 1416000
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 11.61,
|
| 1957 |
+
"learning_rate": 4.069040273492872e-06,
|
| 1958 |
+
"loss": 3.0227,
|
| 1959 |
+
"step": 1424000
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 11.61,
|
| 1963 |
+
"eval_loss": 3.0704684257507324,
|
| 1964 |
+
"eval_runtime": 135.8934,
|
| 1965 |
+
"eval_samples_per_second": 759.97,
|
| 1966 |
+
"eval_steps_per_second": 47.5,
|
| 1967 |
+
"step": 1424000
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 11.68,
|
| 1971 |
+
"eval_loss": 3.0705504417419434,
|
| 1972 |
+
"eval_runtime": 133.4155,
|
| 1973 |
+
"eval_samples_per_second": 774.086,
|
| 1974 |
+
"eval_steps_per_second": 48.383,
|
| 1975 |
+
"step": 1432000
|
| 1976 |
+
},
|
| 1977 |
+
{
|
| 1978 |
+
"epoch": 11.74,
|
| 1979 |
+
"learning_rate": 4.0023346952388895e-06,
|
| 1980 |
+
"loss": 3.0267,
|
| 1981 |
+
"step": 1440000
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 11.74,
|
| 1985 |
+
"eval_loss": 3.069384813308716,
|
| 1986 |
+
"eval_runtime": 133.2021,
|
| 1987 |
+
"eval_samples_per_second": 775.326,
|
| 1988 |
+
"eval_steps_per_second": 48.46,
|
| 1989 |
+
"step": 1440000
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 11.81,
|
| 1993 |
+
"eval_loss": 3.0720527172088623,
|
| 1994 |
+
"eval_runtime": 133.9349,
|
| 1995 |
+
"eval_samples_per_second": 771.083,
|
| 1996 |
+
"eval_steps_per_second": 48.195,
|
| 1997 |
+
"step": 1448000
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 11.87,
|
| 2001 |
+
"learning_rate": 3.935629116984908e-06,
|
| 2002 |
+
"loss": 3.021,
|
| 2003 |
+
"step": 1456000
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 11.87,
|
| 2007 |
+
"eval_loss": 3.068966865539551,
|
| 2008 |
+
"eval_runtime": 132.597,
|
| 2009 |
+
"eval_samples_per_second": 778.864,
|
| 2010 |
+
"eval_steps_per_second": 48.681,
|
| 2011 |
+
"step": 1456000
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 11.94,
|
| 2015 |
+
"eval_loss": 3.060349702835083,
|
| 2016 |
+
"eval_runtime": 134.1972,
|
| 2017 |
+
"eval_samples_per_second": 769.576,
|
| 2018 |
+
"eval_steps_per_second": 48.101,
|
| 2019 |
+
"step": 1464000
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 12.0,
|
| 2023 |
+
"learning_rate": 3.868923538730927e-06,
|
| 2024 |
+
"loss": 3.0144,
|
| 2025 |
+
"step": 1472000
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 12.0,
|
| 2029 |
+
"eval_loss": 3.065760374069214,
|
| 2030 |
+
"eval_runtime": 134.4544,
|
| 2031 |
+
"eval_samples_per_second": 768.104,
|
| 2032 |
+
"eval_steps_per_second": 48.009,
|
| 2033 |
+
"step": 1472000
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 12.07,
|
| 2037 |
+
"eval_loss": 3.0719916820526123,
|
| 2038 |
+
"eval_runtime": 133.6199,
|
| 2039 |
+
"eval_samples_per_second": 772.902,
|
| 2040 |
+
"eval_steps_per_second": 48.309,
|
| 2041 |
+
"step": 1480000
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 12.13,
|
| 2045 |
+
"learning_rate": 3.8022179604769453e-06,
|
| 2046 |
+
"loss": 3.0204,
|
| 2047 |
+
"step": 1488000
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 12.13,
|
| 2051 |
+
"eval_loss": 3.066779851913452,
|
| 2052 |
+
"eval_runtime": 133.3793,
|
| 2053 |
+
"eval_samples_per_second": 774.296,
|
| 2054 |
+
"eval_steps_per_second": 48.396,
|
| 2055 |
+
"step": 1488000
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 12.2,
|
| 2059 |
+
"eval_loss": 3.0773117542266846,
|
| 2060 |
+
"eval_runtime": 135.2249,
|
| 2061 |
+
"eval_samples_per_second": 763.728,
|
| 2062 |
+
"eval_steps_per_second": 47.735,
|
| 2063 |
+
"step": 1496000
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 12.26,
|
| 2067 |
+
"learning_rate": 3.735512382222964e-06,
|
| 2068 |
+
"loss": 3.0085,
|
| 2069 |
+
"step": 1504000
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 12.26,
|
| 2073 |
+
"eval_loss": 3.0847675800323486,
|
| 2074 |
+
"eval_runtime": 133.4406,
|
| 2075 |
+
"eval_samples_per_second": 773.94,
|
| 2076 |
+
"eval_steps_per_second": 48.374,
|
| 2077 |
+
"step": 1504000
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 12.33,
|
| 2081 |
+
"eval_loss": 3.0567853450775146,
|
| 2082 |
+
"eval_runtime": 136.5184,
|
| 2083 |
+
"eval_samples_per_second": 756.492,
|
| 2084 |
+
"eval_steps_per_second": 47.283,
|
| 2085 |
+
"step": 1512000
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 12.39,
|
| 2089 |
+
"learning_rate": 3.668806803968982e-06,
|
| 2090 |
+
"loss": 3.0146,
|
| 2091 |
+
"step": 1520000
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 12.39,
|
| 2095 |
+
"eval_loss": 3.0783281326293945,
|
| 2096 |
+
"eval_runtime": 134.8805,
|
| 2097 |
+
"eval_samples_per_second": 765.678,
|
| 2098 |
+
"eval_steps_per_second": 47.857,
|
| 2099 |
+
"step": 1520000
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 12.46,
|
| 2103 |
+
"eval_loss": 3.073552370071411,
|
| 2104 |
+
"eval_runtime": 133.8542,
|
| 2105 |
+
"eval_samples_per_second": 771.549,
|
| 2106 |
+
"eval_steps_per_second": 48.224,
|
| 2107 |
+
"step": 1528000
|
| 2108 |
+
},
|
| 2109 |
+
{
|
| 2110 |
+
"epoch": 12.52,
|
| 2111 |
+
"learning_rate": 3.6021012257150007e-06,
|
| 2112 |
+
"loss": 3.02,
|
| 2113 |
+
"step": 1536000
|
| 2114 |
+
},
|
| 2115 |
+
{
|
| 2116 |
+
"epoch": 12.52,
|
| 2117 |
+
"eval_loss": 3.0533952713012695,
|
| 2118 |
+
"eval_runtime": 133.5934,
|
| 2119 |
+
"eval_samples_per_second": 773.055,
|
| 2120 |
+
"eval_steps_per_second": 48.318,
|
| 2121 |
+
"step": 1536000
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 12.59,
|
| 2125 |
+
"eval_loss": 3.0684494972229004,
|
| 2126 |
+
"eval_runtime": 133.2901,
|
| 2127 |
+
"eval_samples_per_second": 774.814,
|
| 2128 |
+
"eval_steps_per_second": 48.428,
|
| 2129 |
+
"step": 1544000
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 12.65,
|
| 2133 |
+
"learning_rate": 3.535395647461019e-06,
|
| 2134 |
+
"loss": 3.0229,
|
| 2135 |
+
"step": 1552000
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 12.65,
|
| 2139 |
+
"eval_loss": 3.07673978805542,
|
| 2140 |
+
"eval_runtime": 134.115,
|
| 2141 |
+
"eval_samples_per_second": 770.048,
|
| 2142 |
+
"eval_steps_per_second": 48.13,
|
| 2143 |
+
"step": 1552000
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 12.72,
|
| 2147 |
+
"eval_loss": 3.0568747520446777,
|
| 2148 |
+
"eval_runtime": 134.3484,
|
| 2149 |
+
"eval_samples_per_second": 768.71,
|
| 2150 |
+
"eval_steps_per_second": 48.047,
|
| 2151 |
+
"step": 1560000
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 12.79,
|
| 2155 |
+
"learning_rate": 3.468690069207038e-06,
|
| 2156 |
+
"loss": 3.0152,
|
| 2157 |
+
"step": 1568000
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 12.79,
|
| 2161 |
+
"eval_loss": 3.0787863731384277,
|
| 2162 |
+
"eval_runtime": 133.764,
|
| 2163 |
+
"eval_samples_per_second": 772.068,
|
| 2164 |
+
"eval_steps_per_second": 48.257,
|
| 2165 |
+
"step": 1568000
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 12.85,
|
| 2169 |
+
"eval_loss": 3.066344738006592,
|
| 2170 |
+
"eval_runtime": 133.9216,
|
| 2171 |
+
"eval_samples_per_second": 771.16,
|
| 2172 |
+
"eval_steps_per_second": 48.2,
|
| 2173 |
+
"step": 1576000
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 12.92,
|
| 2177 |
+
"learning_rate": 3.4019844909530565e-06,
|
| 2178 |
+
"loss": 3.02,
|
| 2179 |
+
"step": 1584000
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 12.92,
|
| 2183 |
+
"eval_loss": 3.067016839981079,
|
| 2184 |
+
"eval_runtime": 133.9971,
|
| 2185 |
+
"eval_samples_per_second": 770.725,
|
| 2186 |
+
"eval_steps_per_second": 48.173,
|
| 2187 |
+
"step": 1584000
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 12.98,
|
| 2191 |
+
"eval_loss": 3.0683343410491943,
|
| 2192 |
+
"eval_runtime": 134.2208,
|
| 2193 |
+
"eval_samples_per_second": 769.441,
|
| 2194 |
+
"eval_steps_per_second": 48.092,
|
| 2195 |
+
"step": 1592000
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 13.05,
|
| 2199 |
+
"learning_rate": 3.3352789126990747e-06,
|
| 2200 |
+
"loss": 3.0128,
|
| 2201 |
+
"step": 1600000
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 13.05,
|
| 2205 |
+
"eval_loss": 3.071779489517212,
|
| 2206 |
+
"eval_runtime": 134.2033,
|
| 2207 |
+
"eval_samples_per_second": 769.541,
|
| 2208 |
+
"eval_steps_per_second": 48.099,
|
| 2209 |
+
"step": 1600000
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 13.11,
|
| 2213 |
+
"eval_loss": 3.0846707820892334,
|
| 2214 |
+
"eval_runtime": 134.6625,
|
| 2215 |
+
"eval_samples_per_second": 766.917,
|
| 2216 |
+
"eval_steps_per_second": 47.935,
|
| 2217 |
+
"step": 1608000
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 13.18,
|
| 2221 |
+
"learning_rate": 3.2685733344450933e-06,
|
| 2222 |
+
"loss": 3.016,
|
| 2223 |
+
"step": 1616000
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 13.18,
|
| 2227 |
+
"eval_loss": 3.066356897354126,
|
| 2228 |
+
"eval_runtime": 134.4556,
|
| 2229 |
+
"eval_samples_per_second": 768.097,
|
| 2230 |
+
"eval_steps_per_second": 48.008,
|
| 2231 |
+
"step": 1616000
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 13.24,
|
| 2235 |
+
"eval_loss": 3.0687520503997803,
|
| 2236 |
+
"eval_runtime": 134.3299,
|
| 2237 |
+
"eval_samples_per_second": 768.816,
|
| 2238 |
+
"eval_steps_per_second": 48.053,
|
| 2239 |
+
"step": 1624000
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 13.31,
|
| 2243 |
+
"learning_rate": 3.2018677561911115e-06,
|
| 2244 |
+
"loss": 3.0007,
|
| 2245 |
+
"step": 1632000
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 13.31,
|
| 2249 |
+
"eval_loss": 3.0740671157836914,
|
| 2250 |
+
"eval_runtime": 134.4424,
|
| 2251 |
+
"eval_samples_per_second": 768.173,
|
| 2252 |
+
"eval_steps_per_second": 48.013,
|
| 2253 |
+
"step": 1632000
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 13.37,
|
| 2257 |
+
"eval_loss": 3.0663323402404785,
|
| 2258 |
+
"eval_runtime": 134.2383,
|
| 2259 |
+
"eval_samples_per_second": 769.341,
|
| 2260 |
+
"eval_steps_per_second": 48.086,
|
| 2261 |
+
"step": 1640000
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 13.44,
|
| 2265 |
+
"learning_rate": 3.1351621779371306e-06,
|
| 2266 |
+
"loss": 3.0241,
|
| 2267 |
+
"step": 1648000
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 13.44,
|
| 2271 |
+
"eval_loss": 3.0607213973999023,
|
| 2272 |
+
"eval_runtime": 134.0502,
|
| 2273 |
+
"eval_samples_per_second": 770.42,
|
| 2274 |
+
"eval_steps_per_second": 48.154,
|
| 2275 |
+
"step": 1648000
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 13.5,
|
| 2279 |
+
"eval_loss": 3.0635085105895996,
|
| 2280 |
+
"eval_runtime": 133.9453,
|
| 2281 |
+
"eval_samples_per_second": 771.024,
|
| 2282 |
+
"eval_steps_per_second": 48.191,
|
| 2283 |
+
"step": 1656000
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 13.57,
|
| 2287 |
+
"learning_rate": 3.0684565996831487e-06,
|
| 2288 |
+
"loss": 3.0103,
|
| 2289 |
+
"step": 1664000
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 13.57,
|
| 2293 |
+
"eval_loss": 3.0730724334716797,
|
| 2294 |
+
"eval_runtime": 135.0683,
|
| 2295 |
+
"eval_samples_per_second": 764.613,
|
| 2296 |
+
"eval_steps_per_second": 47.791,
|
| 2297 |
+
"step": 1664000
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 13.63,
|
| 2301 |
+
"eval_loss": 3.0649466514587402,
|
| 2302 |
+
"eval_runtime": 134.138,
|
| 2303 |
+
"eval_samples_per_second": 769.916,
|
| 2304 |
+
"eval_steps_per_second": 48.122,
|
| 2305 |
+
"step": 1672000
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 13.7,
|
| 2309 |
+
"learning_rate": 3.0017510214291673e-06,
|
| 2310 |
+
"loss": 3.0188,
|
| 2311 |
+
"step": 1680000
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 13.7,
|
| 2315 |
+
"eval_loss": 3.058675765991211,
|
| 2316 |
+
"eval_runtime": 134.4659,
|
| 2317 |
+
"eval_samples_per_second": 768.039,
|
| 2318 |
+
"eval_steps_per_second": 48.005,
|
| 2319 |
+
"step": 1680000
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 13.76,
|
| 2323 |
+
"eval_loss": 3.0703861713409424,
|
| 2324 |
+
"eval_runtime": 134.1628,
|
| 2325 |
+
"eval_samples_per_second": 769.773,
|
| 2326 |
+
"eval_steps_per_second": 48.113,
|
| 2327 |
+
"step": 1688000
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 13.83,
|
| 2331 |
+
"learning_rate": 2.9350454431751855e-06,
|
| 2332 |
+
"loss": 3.0217,
|
| 2333 |
+
"step": 1696000
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 13.83,
|
| 2337 |
+
"eval_loss": 3.066443920135498,
|
| 2338 |
+
"eval_runtime": 135.8944,
|
| 2339 |
+
"eval_samples_per_second": 759.965,
|
| 2340 |
+
"eval_steps_per_second": 47.5,
|
| 2341 |
+
"step": 1696000
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 13.89,
|
| 2345 |
+
"eval_loss": 3.0626626014709473,
|
| 2346 |
+
"eval_runtime": 135.45,
|
| 2347 |
+
"eval_samples_per_second": 762.458,
|
| 2348 |
+
"eval_steps_per_second": 47.656,
|
| 2349 |
+
"step": 1704000
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 13.96,
|
| 2353 |
+
"learning_rate": 2.868339864921204e-06,
|
| 2354 |
+
"loss": 3.0282,
|
| 2355 |
+
"step": 1712000
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 13.96,
|
| 2359 |
+
"eval_loss": 3.071357488632202,
|
| 2360 |
+
"eval_runtime": 134.3182,
|
| 2361 |
+
"eval_samples_per_second": 768.883,
|
| 2362 |
+
"eval_steps_per_second": 48.058,
|
| 2363 |
+
"step": 1712000
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 14.02,
|
| 2367 |
+
"eval_loss": 3.0688371658325195,
|
| 2368 |
+
"eval_runtime": 135.2782,
|
| 2369 |
+
"eval_samples_per_second": 763.427,
|
| 2370 |
+
"eval_steps_per_second": 47.716,
|
| 2371 |
+
"step": 1720000
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 14.09,
|
| 2375 |
+
"learning_rate": 2.801634286667223e-06,
|
| 2376 |
+
"loss": 3.0166,
|
| 2377 |
+
"step": 1728000
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 14.09,
|
| 2381 |
+
"eval_loss": 3.05212664604187,
|
| 2382 |
+
"eval_runtime": 135.0648,
|
| 2383 |
+
"eval_samples_per_second": 764.633,
|
| 2384 |
+
"eval_steps_per_second": 47.792,
|
| 2385 |
+
"step": 1728000
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 14.16,
|
| 2389 |
+
"eval_loss": 3.0538179874420166,
|
| 2390 |
+
"eval_runtime": 134.2844,
|
| 2391 |
+
"eval_samples_per_second": 769.076,
|
| 2392 |
+
"eval_steps_per_second": 48.07,
|
| 2393 |
+
"step": 1736000
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 14.22,
|
| 2397 |
+
"learning_rate": 2.7349287084132413e-06,
|
| 2398 |
+
"loss": 3.0134,
|
| 2399 |
+
"step": 1744000
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 14.22,
|
| 2403 |
+
"eval_loss": 3.064086437225342,
|
| 2404 |
+
"eval_runtime": 135.4053,
|
| 2405 |
+
"eval_samples_per_second": 762.71,
|
| 2406 |
+
"eval_steps_per_second": 47.672,
|
| 2407 |
+
"step": 1744000
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 14.29,
|
| 2411 |
+
"eval_loss": 3.063884735107422,
|
| 2412 |
+
"eval_runtime": 134.2537,
|
| 2413 |
+
"eval_samples_per_second": 769.253,
|
| 2414 |
+
"eval_steps_per_second": 48.081,
|
| 2415 |
+
"step": 1752000
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 14.35,
|
| 2419 |
+
"learning_rate": 2.66822313015926e-06,
|
| 2420 |
+
"loss": 3.0032,
|
| 2421 |
+
"step": 1760000
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 14.35,
|
| 2425 |
+
"eval_loss": 3.0587823390960693,
|
| 2426 |
+
"eval_runtime": 135.0451,
|
| 2427 |
+
"eval_samples_per_second": 764.745,
|
| 2428 |
+
"eval_steps_per_second": 47.799,
|
| 2429 |
+
"step": 1760000
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 14.42,
|
| 2433 |
+
"eval_loss": 3.064620018005371,
|
| 2434 |
+
"eval_runtime": 134.9837,
|
| 2435 |
+
"eval_samples_per_second": 765.092,
|
| 2436 |
+
"eval_steps_per_second": 47.821,
|
| 2437 |
+
"step": 1768000
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 14.48,
|
| 2441 |
+
"learning_rate": 2.601517551905278e-06,
|
| 2442 |
+
"loss": 3.0136,
|
| 2443 |
+
"step": 1776000
|
| 2444 |
+
},
|
| 2445 |
+
{
|
| 2446 |
+
"epoch": 14.48,
|
| 2447 |
+
"eval_loss": 3.062889337539673,
|
| 2448 |
+
"eval_runtime": 134.9119,
|
| 2449 |
+
"eval_samples_per_second": 765.499,
|
| 2450 |
+
"eval_steps_per_second": 47.846,
|
| 2451 |
+
"step": 1776000
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 14.55,
|
| 2455 |
+
"eval_loss": 3.0578110218048096,
|
| 2456 |
+
"eval_runtime": 136.5221,
|
| 2457 |
+
"eval_samples_per_second": 756.471,
|
| 2458 |
+
"eval_steps_per_second": 47.282,
|
| 2459 |
+
"step": 1784000
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 14.61,
|
| 2463 |
+
"learning_rate": 2.5348119736512967e-06,
|
| 2464 |
+
"loss": 3.0086,
|
| 2465 |
+
"step": 1792000
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 14.61,
|
| 2469 |
+
"eval_loss": 3.0528934001922607,
|
| 2470 |
+
"eval_runtime": 135.6145,
|
| 2471 |
+
"eval_samples_per_second": 761.534,
|
| 2472 |
+
"eval_steps_per_second": 47.598,
|
| 2473 |
+
"step": 1792000
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 14.68,
|
| 2477 |
+
"eval_loss": 3.0615251064300537,
|
| 2478 |
+
"eval_runtime": 135.3281,
|
| 2479 |
+
"eval_samples_per_second": 763.145,
|
| 2480 |
+
"eval_steps_per_second": 47.699,
|
| 2481 |
+
"step": 1800000
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 14.74,
|
| 2485 |
+
"learning_rate": 2.4681063953973154e-06,
|
| 2486 |
+
"loss": 3.019,
|
| 2487 |
+
"step": 1808000
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 14.74,
|
| 2491 |
+
"eval_loss": 3.0565857887268066,
|
| 2492 |
+
"eval_runtime": 134.9377,
|
| 2493 |
+
"eval_samples_per_second": 765.353,
|
| 2494 |
+
"eval_steps_per_second": 47.837,
|
| 2495 |
+
"step": 1808000
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 14.81,
|
| 2499 |
+
"eval_loss": 3.0658679008483887,
|
| 2500 |
+
"eval_runtime": 135.2159,
|
| 2501 |
+
"eval_samples_per_second": 763.778,
|
| 2502 |
+
"eval_steps_per_second": 47.738,
|
| 2503 |
+
"step": 1816000
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 14.87,
|
| 2507 |
+
"learning_rate": 2.4014008171433335e-06,
|
| 2508 |
+
"loss": 3.024,
|
| 2509 |
+
"step": 1824000
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 14.87,
|
| 2513 |
+
"eval_loss": 3.061464786529541,
|
| 2514 |
+
"eval_runtime": 135.2789,
|
| 2515 |
+
"eval_samples_per_second": 763.423,
|
| 2516 |
+
"eval_steps_per_second": 47.716,
|
| 2517 |
+
"step": 1824000
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 14.94,
|
| 2521 |
+
"eval_loss": 3.0530033111572266,
|
| 2522 |
+
"eval_runtime": 135.9081,
|
| 2523 |
+
"eval_samples_per_second": 759.889,
|
| 2524 |
+
"eval_steps_per_second": 47.495,
|
| 2525 |
+
"step": 1832000
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 15.0,
|
| 2529 |
+
"learning_rate": 2.334695238889352e-06,
|
| 2530 |
+
"loss": 3.0089,
|
| 2531 |
+
"step": 1840000
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 15.0,
|
| 2535 |
+
"eval_loss": 3.0796985626220703,
|
| 2536 |
+
"eval_runtime": 135.2715,
|
| 2537 |
+
"eval_samples_per_second": 763.465,
|
| 2538 |
+
"eval_steps_per_second": 47.719,
|
| 2539 |
+
"step": 1840000
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 15.07,
|
| 2543 |
+
"eval_loss": 3.0700411796569824,
|
| 2544 |
+
"eval_runtime": 136.6273,
|
| 2545 |
+
"eval_samples_per_second": 755.888,
|
| 2546 |
+
"eval_steps_per_second": 47.245,
|
| 2547 |
+
"step": 1848000
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 15.13,
|
| 2551 |
+
"learning_rate": 2.2679896606353707e-06,
|
| 2552 |
+
"loss": 3.0174,
|
| 2553 |
+
"step": 1856000
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 15.13,
|
| 2557 |
+
"eval_loss": 3.0748071670532227,
|
| 2558 |
+
"eval_runtime": 136.44,
|
| 2559 |
+
"eval_samples_per_second": 756.926,
|
| 2560 |
+
"eval_steps_per_second": 47.31,
|
| 2561 |
+
"step": 1856000
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 15.2,
|
| 2565 |
+
"eval_loss": 3.064267635345459,
|
| 2566 |
+
"eval_runtime": 135.3728,
|
| 2567 |
+
"eval_samples_per_second": 762.894,
|
| 2568 |
+
"eval_steps_per_second": 47.683,
|
| 2569 |
+
"step": 1864000
|
| 2570 |
+
},
|
| 2571 |
+
{
|
| 2572 |
+
"epoch": 15.26,
|
| 2573 |
+
"learning_rate": 2.2012840823813894e-06,
|
| 2574 |
+
"loss": 3.0176,
|
| 2575 |
+
"step": 1872000
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 15.26,
|
| 2579 |
+
"eval_loss": 3.0627517700195312,
|
| 2580 |
+
"eval_runtime": 135.5713,
|
| 2581 |
+
"eval_samples_per_second": 761.776,
|
| 2582 |
+
"eval_steps_per_second": 47.613,
|
| 2583 |
+
"step": 1872000
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"epoch": 15.33,
|
| 2587 |
+
"eval_loss": 3.0629563331604004,
|
| 2588 |
+
"eval_runtime": 135.894,
|
| 2589 |
+
"eval_samples_per_second": 759.967,
|
| 2590 |
+
"eval_steps_per_second": 47.5,
|
| 2591 |
+
"step": 1880000
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 15.39,
|
| 2595 |
+
"learning_rate": 2.134578504127408e-06,
|
| 2596 |
+
"loss": 3.0164,
|
| 2597 |
+
"step": 1888000
|
| 2598 |
+
},
|
| 2599 |
+
{
|
| 2600 |
+
"epoch": 15.39,
|
| 2601 |
+
"eval_loss": 3.0721538066864014,
|
| 2602 |
+
"eval_runtime": 135.9329,
|
| 2603 |
+
"eval_samples_per_second": 759.75,
|
| 2604 |
+
"eval_steps_per_second": 47.487,
|
| 2605 |
+
"step": 1888000
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 15.46,
|
| 2609 |
+
"eval_loss": 3.0744197368621826,
|
| 2610 |
+
"eval_runtime": 135.4506,
|
| 2611 |
+
"eval_samples_per_second": 762.455,
|
| 2612 |
+
"eval_steps_per_second": 47.656,
|
| 2613 |
+
"step": 1896000
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 15.53,
|
| 2617 |
+
"learning_rate": 2.067872925873426e-06,
|
| 2618 |
+
"loss": 3.0302,
|
| 2619 |
+
"step": 1904000
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 15.53,
|
| 2623 |
+
"eval_loss": 3.0739452838897705,
|
| 2624 |
+
"eval_runtime": 135.8281,
|
| 2625 |
+
"eval_samples_per_second": 760.336,
|
| 2626 |
+
"eval_steps_per_second": 47.523,
|
| 2627 |
+
"step": 1904000
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 15.59,
|
| 2631 |
+
"eval_loss": 3.0700225830078125,
|
| 2632 |
+
"eval_runtime": 136.0724,
|
| 2633 |
+
"eval_samples_per_second": 758.971,
|
| 2634 |
+
"eval_steps_per_second": 47.438,
|
| 2635 |
+
"step": 1912000
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 15.66,
|
| 2639 |
+
"learning_rate": 2.0011673476194448e-06,
|
| 2640 |
+
"loss": 3.0204,
|
| 2641 |
+
"step": 1920000
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 15.66,
|
| 2645 |
+
"eval_loss": 3.0751476287841797,
|
| 2646 |
+
"eval_runtime": 136.119,
|
| 2647 |
+
"eval_samples_per_second": 758.711,
|
| 2648 |
+
"eval_steps_per_second": 47.422,
|
| 2649 |
+
"step": 1920000
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 15.72,
|
| 2653 |
+
"eval_loss": 3.0597870349884033,
|
| 2654 |
+
"eval_runtime": 136.3427,
|
| 2655 |
+
"eval_samples_per_second": 757.466,
|
| 2656 |
+
"eval_steps_per_second": 47.344,
|
| 2657 |
+
"step": 1928000
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 15.79,
|
| 2661 |
+
"learning_rate": 1.9344617693654634e-06,
|
| 2662 |
+
"loss": 3.0147,
|
| 2663 |
+
"step": 1936000
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 15.79,
|
| 2667 |
+
"eval_loss": 3.0522122383117676,
|
| 2668 |
+
"eval_runtime": 136.0082,
|
| 2669 |
+
"eval_samples_per_second": 759.329,
|
| 2670 |
+
"eval_steps_per_second": 47.46,
|
| 2671 |
+
"step": 1936000
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 15.85,
|
| 2675 |
+
"eval_loss": 3.065509557723999,
|
| 2676 |
+
"eval_runtime": 136.1421,
|
| 2677 |
+
"eval_samples_per_second": 758.582,
|
| 2678 |
+
"eval_steps_per_second": 47.414,
|
| 2679 |
+
"step": 1944000
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 15.92,
|
| 2683 |
+
"learning_rate": 1.867756191111482e-06,
|
| 2684 |
+
"loss": 3.0245,
|
| 2685 |
+
"step": 1952000
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 15.92,
|
| 2689 |
+
"eval_loss": 3.0568597316741943,
|
| 2690 |
+
"eval_runtime": 136.6776,
|
| 2691 |
+
"eval_samples_per_second": 755.61,
|
| 2692 |
+
"eval_steps_per_second": 47.228,
|
| 2693 |
+
"step": 1952000
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 15.98,
|
| 2697 |
+
"eval_loss": 3.062300205230713,
|
| 2698 |
+
"eval_runtime": 136.0258,
|
| 2699 |
+
"eval_samples_per_second": 759.231,
|
| 2700 |
+
"eval_steps_per_second": 47.454,
|
| 2701 |
+
"step": 1960000
|
| 2702 |
+
},
|
| 2703 |
+
{
|
| 2704 |
+
"epoch": 16.05,
|
| 2705 |
+
"learning_rate": 1.8010506128575004e-06,
|
| 2706 |
+
"loss": 3.0069,
|
| 2707 |
+
"step": 1968000
|
| 2708 |
+
},
|
| 2709 |
+
{
|
| 2710 |
+
"epoch": 16.05,
|
| 2711 |
+
"eval_loss": 3.059983730316162,
|
| 2712 |
+
"eval_runtime": 136.4638,
|
| 2713 |
+
"eval_samples_per_second": 756.794,
|
| 2714 |
+
"eval_steps_per_second": 47.302,
|
| 2715 |
+
"step": 1968000
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 16.11,
|
| 2719 |
+
"eval_loss": 3.0638678073883057,
|
| 2720 |
+
"eval_runtime": 137.569,
|
| 2721 |
+
"eval_samples_per_second": 750.714,
|
| 2722 |
+
"eval_steps_per_second": 46.922,
|
| 2723 |
+
"step": 1976000
|
| 2724 |
+
},
|
| 2725 |
+
{
|
| 2726 |
+
"epoch": 16.18,
|
| 2727 |
+
"learning_rate": 1.734345034603519e-06,
|
| 2728 |
+
"loss": 3.0068,
|
| 2729 |
+
"step": 1984000
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 16.18,
|
| 2733 |
+
"eval_loss": 3.077465534210205,
|
| 2734 |
+
"eval_runtime": 136.0507,
|
| 2735 |
+
"eval_samples_per_second": 759.092,
|
| 2736 |
+
"eval_steps_per_second": 47.446,
|
| 2737 |
+
"step": 1984000
|
| 2738 |
+
},
|
| 2739 |
+
{
|
| 2740 |
+
"epoch": 16.24,
|
| 2741 |
+
"eval_loss": 3.0668864250183105,
|
| 2742 |
+
"eval_runtime": 136.2552,
|
| 2743 |
+
"eval_samples_per_second": 757.953,
|
| 2744 |
+
"eval_steps_per_second": 47.374,
|
| 2745 |
+
"step": 1992000
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 16.31,
|
| 2749 |
+
"learning_rate": 1.6676394563495374e-06,
|
| 2750 |
+
"loss": 3.0275,
|
| 2751 |
+
"step": 2000000
|
| 2752 |
+
},
|
| 2753 |
+
{
|
| 2754 |
+
"epoch": 16.31,
|
| 2755 |
+
"eval_loss": 3.062725782394409,
|
| 2756 |
+
"eval_runtime": 136.3436,
|
| 2757 |
+
"eval_samples_per_second": 757.461,
|
| 2758 |
+
"eval_steps_per_second": 47.344,
|
| 2759 |
+
"step": 2000000
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 16.37,
|
| 2763 |
+
"eval_loss": 3.0644514560699463,
|
| 2764 |
+
"eval_runtime": 137.752,
|
| 2765 |
+
"eval_samples_per_second": 749.717,
|
| 2766 |
+
"eval_steps_per_second": 46.86,
|
| 2767 |
+
"step": 2008000
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 16.44,
|
| 2771 |
+
"learning_rate": 1.6009338780955558e-06,
|
| 2772 |
+
"loss": 3.0164,
|
| 2773 |
+
"step": 2016000
|
| 2774 |
+
},
|
| 2775 |
+
{
|
| 2776 |
+
"epoch": 16.44,
|
| 2777 |
+
"eval_loss": 3.0666866302490234,
|
| 2778 |
+
"eval_runtime": 135.9171,
|
| 2779 |
+
"eval_samples_per_second": 759.838,
|
| 2780 |
+
"eval_steps_per_second": 47.492,
|
| 2781 |
+
"step": 2016000
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 16.5,
|
| 2785 |
+
"eval_loss": 3.048987627029419,
|
| 2786 |
+
"eval_runtime": 136.0156,
|
| 2787 |
+
"eval_samples_per_second": 759.288,
|
| 2788 |
+
"eval_steps_per_second": 47.458,
|
| 2789 |
+
"step": 2024000
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 16.57,
|
| 2793 |
+
"learning_rate": 1.5342282998415744e-06,
|
| 2794 |
+
"loss": 3.0148,
|
| 2795 |
+
"step": 2032000
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 16.57,
|
| 2799 |
+
"eval_loss": 3.061800003051758,
|
| 2800 |
+
"eval_runtime": 137.187,
|
| 2801 |
+
"eval_samples_per_second": 752.805,
|
| 2802 |
+
"eval_steps_per_second": 47.053,
|
| 2803 |
+
"step": 2032000
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 16.63,
|
| 2807 |
+
"eval_loss": 3.0544731616973877,
|
| 2808 |
+
"eval_runtime": 137.5014,
|
| 2809 |
+
"eval_samples_per_second": 751.083,
|
| 2810 |
+
"eval_steps_per_second": 46.945,
|
| 2811 |
+
"step": 2040000
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 16.7,
|
| 2815 |
+
"learning_rate": 1.4675227215875928e-06,
|
| 2816 |
+
"loss": 3.022,
|
| 2817 |
+
"step": 2048000
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 16.7,
|
| 2821 |
+
"eval_loss": 3.0651352405548096,
|
| 2822 |
+
"eval_runtime": 137.0124,
|
| 2823 |
+
"eval_samples_per_second": 753.764,
|
| 2824 |
+
"eval_steps_per_second": 47.113,
|
| 2825 |
+
"step": 2048000
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 16.76,
|
| 2829 |
+
"eval_loss": 3.068650484085083,
|
| 2830 |
+
"eval_runtime": 137.324,
|
| 2831 |
+
"eval_samples_per_second": 752.053,
|
| 2832 |
+
"eval_steps_per_second": 47.006,
|
| 2833 |
+
"step": 2056000
|
| 2834 |
+
},
|
| 2835 |
+
{
|
| 2836 |
+
"epoch": 16.83,
|
| 2837 |
+
"learning_rate": 1.4008171433336116e-06,
|
| 2838 |
+
"loss": 3.0235,
|
| 2839 |
+
"step": 2064000
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 16.83,
|
| 2843 |
+
"eval_loss": 3.0515873432159424,
|
| 2844 |
+
"eval_runtime": 137.8405,
|
| 2845 |
+
"eval_samples_per_second": 749.235,
|
| 2846 |
+
"eval_steps_per_second": 46.829,
|
| 2847 |
+
"step": 2064000
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 16.89,
|
| 2851 |
+
"eval_loss": 3.0761473178863525,
|
| 2852 |
+
"eval_runtime": 137.435,
|
| 2853 |
+
"eval_samples_per_second": 751.446,
|
| 2854 |
+
"eval_steps_per_second": 46.968,
|
| 2855 |
+
"step": 2072000
|
| 2856 |
+
},
|
| 2857 |
+
{
|
| 2858 |
+
"epoch": 16.96,
|
| 2859 |
+
"learning_rate": 1.33411156507963e-06,
|
| 2860 |
+
"loss": 3.0194,
|
| 2861 |
+
"step": 2080000
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 16.96,
|
| 2865 |
+
"eval_loss": 3.0807414054870605,
|
| 2866 |
+
"eval_runtime": 136.8928,
|
| 2867 |
+
"eval_samples_per_second": 754.423,
|
| 2868 |
+
"eval_steps_per_second": 47.154,
|
| 2869 |
+
"step": 2080000
|
| 2870 |
+
},
|
| 2871 |
+
{
|
| 2872 |
+
"epoch": 17.03,
|
| 2873 |
+
"eval_loss": 3.060075283050537,
|
| 2874 |
+
"eval_runtime": 136.6441,
|
| 2875 |
+
"eval_samples_per_second": 755.796,
|
| 2876 |
+
"eval_steps_per_second": 47.24,
|
| 2877 |
+
"step": 2088000
|
| 2878 |
+
},
|
| 2879 |
+
{
|
| 2880 |
+
"epoch": 17.09,
|
| 2881 |
+
"learning_rate": 1.2674059868256484e-06,
|
| 2882 |
+
"loss": 3.0142,
|
| 2883 |
+
"step": 2096000
|
| 2884 |
+
},
|
| 2885 |
+
{
|
| 2886 |
+
"epoch": 17.09,
|
| 2887 |
+
"eval_loss": 3.0721395015716553,
|
| 2888 |
+
"eval_runtime": 136.5201,
|
| 2889 |
+
"eval_samples_per_second": 756.482,
|
| 2890 |
+
"eval_steps_per_second": 47.282,
|
| 2891 |
+
"step": 2096000
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 17.16,
|
| 2895 |
+
"eval_loss": 3.0653316974639893,
|
| 2896 |
+
"eval_runtime": 138.2812,
|
| 2897 |
+
"eval_samples_per_second": 746.848,
|
| 2898 |
+
"eval_steps_per_second": 46.68,
|
| 2899 |
+
"step": 2104000
|
| 2900 |
+
},
|
| 2901 |
+
{
|
| 2902 |
+
"epoch": 17.22,
|
| 2903 |
+
"learning_rate": 1.2007004085716668e-06,
|
| 2904 |
+
"loss": 3.0183,
|
| 2905 |
+
"step": 2112000
|
| 2906 |
+
},
|
| 2907 |
+
{
|
| 2908 |
+
"epoch": 17.22,
|
| 2909 |
+
"eval_loss": 3.061683416366577,
|
| 2910 |
+
"eval_runtime": 136.6654,
|
| 2911 |
+
"eval_samples_per_second": 755.678,
|
| 2912 |
+
"eval_steps_per_second": 47.232,
|
| 2913 |
+
"step": 2112000
|
| 2914 |
+
},
|
| 2915 |
+
{
|
| 2916 |
+
"epoch": 17.29,
|
| 2917 |
+
"eval_loss": 3.062178373336792,
|
| 2918 |
+
"eval_runtime": 137.9621,
|
| 2919 |
+
"eval_samples_per_second": 748.575,
|
| 2920 |
+
"eval_steps_per_second": 46.788,
|
| 2921 |
+
"step": 2120000
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 17.35,
|
| 2925 |
+
"learning_rate": 1.1339948303176854e-06,
|
| 2926 |
+
"loss": 3.0092,
|
| 2927 |
+
"step": 2128000
|
| 2928 |
+
},
|
| 2929 |
+
{
|
| 2930 |
+
"epoch": 17.35,
|
| 2931 |
+
"eval_loss": 3.068242311477661,
|
| 2932 |
+
"eval_runtime": 137.4752,
|
| 2933 |
+
"eval_samples_per_second": 751.227,
|
| 2934 |
+
"eval_steps_per_second": 46.954,
|
| 2935 |
+
"step": 2128000
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 17.42,
|
| 2939 |
+
"eval_loss": 3.073157787322998,
|
| 2940 |
+
"eval_runtime": 136.5003,
|
| 2941 |
+
"eval_samples_per_second": 756.592,
|
| 2942 |
+
"eval_steps_per_second": 47.289,
|
| 2943 |
+
"step": 2136000
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 17.48,
|
| 2947 |
+
"learning_rate": 1.067289252063704e-06,
|
| 2948 |
+
"loss": 3.0071,
|
| 2949 |
+
"step": 2144000
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 17.48,
|
| 2953 |
+
"eval_loss": 3.0763022899627686,
|
| 2954 |
+
"eval_runtime": 137.95,
|
| 2955 |
+
"eval_samples_per_second": 748.641,
|
| 2956 |
+
"eval_steps_per_second": 46.792,
|
| 2957 |
+
"step": 2144000
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 17.55,
|
| 2961 |
+
"eval_loss": 3.0674524307250977,
|
| 2962 |
+
"eval_runtime": 137.3106,
|
| 2963 |
+
"eval_samples_per_second": 752.127,
|
| 2964 |
+
"eval_steps_per_second": 47.01,
|
| 2965 |
+
"step": 2152000
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 17.61,
|
| 2969 |
+
"learning_rate": 1.0005836738097224e-06,
|
| 2970 |
+
"loss": 3.0272,
|
| 2971 |
+
"step": 2160000
|
| 2972 |
+
},
|
| 2973 |
+
{
|
| 2974 |
+
"epoch": 17.61,
|
| 2975 |
+
"eval_loss": 3.0671498775482178,
|
| 2976 |
+
"eval_runtime": 138.0717,
|
| 2977 |
+
"eval_samples_per_second": 747.981,
|
| 2978 |
+
"eval_steps_per_second": 46.751,
|
| 2979 |
+
"step": 2160000
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 17.68,
|
| 2983 |
+
"eval_loss": 3.062239170074463,
|
| 2984 |
+
"eval_runtime": 138.0499,
|
| 2985 |
+
"eval_samples_per_second": 748.099,
|
| 2986 |
+
"eval_steps_per_second": 46.758,
|
| 2987 |
+
"step": 2168000
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 17.74,
|
| 2991 |
+
"learning_rate": 9.33878095555741e-07,
|
| 2992 |
+
"loss": 3.0235,
|
| 2993 |
+
"step": 2176000
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 17.74,
|
| 2997 |
+
"eval_loss": 3.0789263248443604,
|
| 2998 |
+
"eval_runtime": 137.5626,
|
| 2999 |
+
"eval_samples_per_second": 750.749,
|
| 3000 |
+
"eval_steps_per_second": 46.924,
|
| 3001 |
+
"step": 2176000
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 17.81,
|
| 3005 |
+
"eval_loss": 3.062295436859131,
|
| 3006 |
+
"eval_runtime": 138.8694,
|
| 3007 |
+
"eval_samples_per_second": 743.684,
|
| 3008 |
+
"eval_steps_per_second": 46.483,
|
| 3009 |
+
"step": 2184000
|
| 3010 |
+
},
|
| 3011 |
+
{
|
| 3012 |
+
"epoch": 17.87,
|
| 3013 |
+
"learning_rate": 8.671725173017595e-07,
|
| 3014 |
+
"loss": 3.0179,
|
| 3015 |
+
"step": 2192000
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 17.87,
|
| 3019 |
+
"eval_loss": 3.078376054763794,
|
| 3020 |
+
"eval_runtime": 136.985,
|
| 3021 |
+
"eval_samples_per_second": 753.914,
|
| 3022 |
+
"eval_steps_per_second": 47.122,
|
| 3023 |
+
"step": 2192000
|
| 3024 |
+
},
|
| 3025 |
+
{
|
| 3026 |
+
"epoch": 17.94,
|
| 3027 |
+
"eval_loss": 3.062905788421631,
|
| 3028 |
+
"eval_runtime": 137.7472,
|
| 3029 |
+
"eval_samples_per_second": 749.743,
|
| 3030 |
+
"eval_steps_per_second": 46.861,
|
| 3031 |
+
"step": 2200000
|
| 3032 |
+
},
|
| 3033 |
+
{
|
| 3034 |
+
"epoch": 18.0,
|
| 3035 |
+
"learning_rate": 8.004669390477779e-07,
|
| 3036 |
+
"loss": 3.0209,
|
| 3037 |
+
"step": 2208000
|
| 3038 |
+
},
|
| 3039 |
+
{
|
| 3040 |
+
"epoch": 18.0,
|
| 3041 |
+
"eval_loss": 3.0731070041656494,
|
| 3042 |
+
"eval_runtime": 138.0906,
|
| 3043 |
+
"eval_samples_per_second": 747.878,
|
| 3044 |
+
"eval_steps_per_second": 46.745,
|
| 3045 |
+
"step": 2208000
|
| 3046 |
+
},
|
| 3047 |
+
{
|
| 3048 |
+
"epoch": 18.07,
|
| 3049 |
+
"eval_loss": 3.0945563316345215,
|
| 3050 |
+
"eval_runtime": 137.4959,
|
| 3051 |
+
"eval_samples_per_second": 751.113,
|
| 3052 |
+
"eval_steps_per_second": 46.947,
|
| 3053 |
+
"step": 2216000
|
| 3054 |
+
},
|
| 3055 |
+
{
|
| 3056 |
+
"epoch": 18.13,
|
| 3057 |
+
"learning_rate": 7.337613607937964e-07,
|
| 3058 |
+
"loss": 3.0237,
|
| 3059 |
+
"step": 2224000
|
| 3060 |
+
},
|
| 3061 |
+
{
|
| 3062 |
+
"epoch": 18.13,
|
| 3063 |
+
"eval_loss": 3.065315008163452,
|
| 3064 |
+
"eval_runtime": 138.0159,
|
| 3065 |
+
"eval_samples_per_second": 748.283,
|
| 3066 |
+
"eval_steps_per_second": 46.77,
|
| 3067 |
+
"step": 2224000
|
| 3068 |
+
},
|
| 3069 |
+
{
|
| 3070 |
+
"epoch": 18.2,
|
| 3071 |
+
"eval_loss": 3.0589962005615234,
|
| 3072 |
+
"eval_runtime": 137.6387,
|
| 3073 |
+
"eval_samples_per_second": 750.334,
|
| 3074 |
+
"eval_steps_per_second": 46.898,
|
| 3075 |
+
"step": 2232000
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 18.26,
|
| 3079 |
+
"learning_rate": 6.67055782539815e-07,
|
| 3080 |
+
"loss": 3.0164,
|
| 3081 |
+
"step": 2240000
|
| 3082 |
+
},
|
| 3083 |
+
{
|
| 3084 |
+
"epoch": 18.26,
|
| 3085 |
+
"eval_loss": 3.070741891860962,
|
| 3086 |
+
"eval_runtime": 138.4523,
|
| 3087 |
+
"eval_samples_per_second": 745.925,
|
| 3088 |
+
"eval_steps_per_second": 46.623,
|
| 3089 |
+
"step": 2240000
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 18.33,
|
| 3093 |
+
"eval_loss": 3.0545763969421387,
|
| 3094 |
+
"eval_runtime": 138.1194,
|
| 3095 |
+
"eval_samples_per_second": 747.723,
|
| 3096 |
+
"eval_steps_per_second": 46.735,
|
| 3097 |
+
"step": 2248000
|
| 3098 |
+
},
|
| 3099 |
+
{
|
| 3100 |
+
"epoch": 18.4,
|
| 3101 |
+
"learning_rate": 6.003502042858334e-07,
|
| 3102 |
+
"loss": 3.0206,
|
| 3103 |
+
"step": 2256000
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 18.4,
|
| 3107 |
+
"eval_loss": 3.0741806030273438,
|
| 3108 |
+
"eval_runtime": 138.8634,
|
| 3109 |
+
"eval_samples_per_second": 743.717,
|
| 3110 |
+
"eval_steps_per_second": 46.485,
|
| 3111 |
+
"step": 2256000
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 18.46,
|
| 3115 |
+
"eval_loss": 3.0793333053588867,
|
| 3116 |
+
"eval_runtime": 138.6181,
|
| 3117 |
+
"eval_samples_per_second": 745.032,
|
| 3118 |
+
"eval_steps_per_second": 46.567,
|
| 3119 |
+
"step": 2264000
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 18.53,
|
| 3123 |
+
"learning_rate": 5.33644626031852e-07,
|
| 3124 |
+
"loss": 3.0138,
|
| 3125 |
+
"step": 2272000
|
| 3126 |
+
},
|
| 3127 |
+
{
|
| 3128 |
+
"epoch": 18.53,
|
| 3129 |
+
"eval_loss": 3.05604887008667,
|
| 3130 |
+
"eval_runtime": 139.1325,
|
| 3131 |
+
"eval_samples_per_second": 742.278,
|
| 3132 |
+
"eval_steps_per_second": 46.395,
|
| 3133 |
+
"step": 2272000
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 18.59,
|
| 3137 |
+
"eval_loss": 3.086977958679199,
|
| 3138 |
+
"eval_runtime": 137.8163,
|
| 3139 |
+
"eval_samples_per_second": 749.367,
|
| 3140 |
+
"eval_steps_per_second": 46.838,
|
| 3141 |
+
"step": 2280000
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 18.66,
|
| 3145 |
+
"learning_rate": 4.669390477778705e-07,
|
| 3146 |
+
"loss": 3.0377,
|
| 3147 |
+
"step": 2288000
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 18.66,
|
| 3151 |
+
"eval_loss": 3.07423996925354,
|
| 3152 |
+
"eval_runtime": 137.4738,
|
| 3153 |
+
"eval_samples_per_second": 751.234,
|
| 3154 |
+
"eval_steps_per_second": 46.954,
|
| 3155 |
+
"step": 2288000
|
| 3156 |
+
},
|
| 3157 |
+
{
|
| 3158 |
+
"epoch": 18.72,
|
| 3159 |
+
"eval_loss": 3.0675508975982666,
|
| 3160 |
+
"eval_runtime": 138.0596,
|
| 3161 |
+
"eval_samples_per_second": 748.046,
|
| 3162 |
+
"eval_steps_per_second": 46.755,
|
| 3163 |
+
"step": 2296000
|
| 3164 |
+
},
|
| 3165 |
+
{
|
| 3166 |
+
"epoch": 18.79,
|
| 3167 |
+
"learning_rate": 4.0023346952388894e-07,
|
| 3168 |
+
"loss": 3.0227,
|
| 3169 |
+
"step": 2304000
|
| 3170 |
+
},
|
| 3171 |
+
{
|
| 3172 |
+
"epoch": 18.79,
|
| 3173 |
+
"eval_loss": 3.06254506111145,
|
| 3174 |
+
"eval_runtime": 139.3504,
|
| 3175 |
+
"eval_samples_per_second": 741.117,
|
| 3176 |
+
"eval_steps_per_second": 46.322,
|
| 3177 |
+
"step": 2304000
|
| 3178 |
+
},
|
| 3179 |
+
{
|
| 3180 |
+
"epoch": 18.85,
|
| 3181 |
+
"eval_loss": 3.0736207962036133,
|
| 3182 |
+
"eval_runtime": 139.5433,
|
| 3183 |
+
"eval_samples_per_second": 740.093,
|
| 3184 |
+
"eval_steps_per_second": 46.258,
|
| 3185 |
+
"step": 2312000
|
| 3186 |
+
},
|
| 3187 |
+
{
|
| 3188 |
+
"epoch": 18.92,
|
| 3189 |
+
"learning_rate": 3.335278912699075e-07,
|
| 3190 |
+
"loss": 3.0359,
|
| 3191 |
+
"step": 2320000
|
| 3192 |
+
},
|
| 3193 |
+
{
|
| 3194 |
+
"epoch": 18.92,
|
| 3195 |
+
"eval_loss": 3.0800607204437256,
|
| 3196 |
+
"eval_runtime": 138.2846,
|
| 3197 |
+
"eval_samples_per_second": 746.829,
|
| 3198 |
+
"eval_steps_per_second": 46.679,
|
| 3199 |
+
"step": 2320000
|
| 3200 |
+
},
|
| 3201 |
+
{
|
| 3202 |
+
"epoch": 18.98,
|
| 3203 |
+
"eval_loss": 3.0709972381591797,
|
| 3204 |
+
"eval_runtime": 139.0505,
|
| 3205 |
+
"eval_samples_per_second": 742.716,
|
| 3206 |
+
"eval_steps_per_second": 46.422,
|
| 3207 |
+
"step": 2328000
|
| 3208 |
+
},
|
| 3209 |
+
{
|
| 3210 |
+
"epoch": 19.05,
|
| 3211 |
+
"learning_rate": 2.66822313015926e-07,
|
| 3212 |
+
"loss": 3.0248,
|
| 3213 |
+
"step": 2336000
|
| 3214 |
+
},
|
| 3215 |
+
{
|
| 3216 |
+
"epoch": 19.05,
|
| 3217 |
+
"eval_loss": 3.069218158721924,
|
| 3218 |
+
"eval_runtime": 138.9779,
|
| 3219 |
+
"eval_samples_per_second": 743.104,
|
| 3220 |
+
"eval_steps_per_second": 46.446,
|
| 3221 |
+
"step": 2336000
|
| 3222 |
+
},
|
| 3223 |
+
{
|
| 3224 |
+
"epoch": 19.11,
|
| 3225 |
+
"eval_loss": 3.067660331726074,
|
| 3226 |
+
"eval_runtime": 138.4099,
|
| 3227 |
+
"eval_samples_per_second": 746.154,
|
| 3228 |
+
"eval_steps_per_second": 46.637,
|
| 3229 |
+
"step": 2344000
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 19.18,
|
| 3233 |
+
"learning_rate": 2.0011673476194447e-07,
|
| 3234 |
+
"loss": 3.0235,
|
| 3235 |
+
"step": 2352000
|
| 3236 |
+
},
|
| 3237 |
+
{
|
| 3238 |
+
"epoch": 19.18,
|
| 3239 |
+
"eval_loss": 3.089552879333496,
|
| 3240 |
+
"eval_runtime": 138.573,
|
| 3241 |
+
"eval_samples_per_second": 745.275,
|
| 3242 |
+
"eval_steps_per_second": 46.582,
|
| 3243 |
+
"step": 2352000
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 19.24,
|
| 3247 |
+
"eval_loss": 3.0777699947357178,
|
| 3248 |
+
"eval_runtime": 140.4362,
|
| 3249 |
+
"eval_samples_per_second": 735.388,
|
| 3250 |
+
"eval_steps_per_second": 45.964,
|
| 3251 |
+
"step": 2360000
|
| 3252 |
+
},
|
| 3253 |
+
{
|
| 3254 |
+
"epoch": 19.31,
|
| 3255 |
+
"learning_rate": 1.33411156507963e-07,
|
| 3256 |
+
"loss": 3.0187,
|
| 3257 |
+
"step": 2368000
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 19.31,
|
| 3261 |
+
"eval_loss": 3.069951295852661,
|
| 3262 |
+
"eval_runtime": 140.0319,
|
| 3263 |
+
"eval_samples_per_second": 737.511,
|
| 3264 |
+
"eval_steps_per_second": 46.097,
|
| 3265 |
+
"step": 2368000
|
| 3266 |
+
},
|
| 3267 |
+
{
|
| 3268 |
+
"epoch": 19.37,
|
| 3269 |
+
"eval_loss": 3.0742506980895996,
|
| 3270 |
+
"eval_runtime": 139.3192,
|
| 3271 |
+
"eval_samples_per_second": 741.283,
|
| 3272 |
+
"eval_steps_per_second": 46.332,
|
| 3273 |
+
"step": 2376000
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 19.44,
|
| 3277 |
+
"learning_rate": 6.67055782539815e-08,
|
| 3278 |
+
"loss": 3.0189,
|
| 3279 |
+
"step": 2384000
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 19.44,
|
| 3283 |
+
"eval_loss": 3.0780065059661865,
|
| 3284 |
+
"eval_runtime": 138.4114,
|
| 3285 |
+
"eval_samples_per_second": 746.145,
|
| 3286 |
+
"eval_steps_per_second": 46.636,
|
| 3287 |
+
"step": 2384000
|
| 3288 |
+
},
|
| 3289 |
+
{
|
| 3290 |
+
"epoch": 19.5,
|
| 3291 |
+
"eval_loss": 3.0866599082946777,
|
| 3292 |
+
"eval_runtime": 138.3665,
|
| 3293 |
+
"eval_samples_per_second": 746.387,
|
| 3294 |
+
"eval_steps_per_second": 46.651,
|
| 3295 |
+
"step": 2392000
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 19.57,
|
| 3299 |
+
"learning_rate": 0.0,
|
| 3300 |
+
"loss": 3.0184,
|
| 3301 |
+
"step": 2400000
|
| 3302 |
+
},
|
| 3303 |
+
{
|
| 3304 |
+
"epoch": 19.57,
|
| 3305 |
+
"eval_loss": 3.079288959503174,
|
| 3306 |
+
"eval_runtime": 138.4519,
|
| 3307 |
+
"eval_samples_per_second": 745.927,
|
| 3308 |
+
"eval_steps_per_second": 46.623,
|
| 3309 |
+
"step": 2400000
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 19.57,
|
| 3313 |
+
"step": 2400000,
|
| 3314 |
+
"total_flos": 7.178820925216543e+17,
|
| 3315 |
+
"train_loss": 2.9400340771484377,
|
| 3316 |
+
"train_runtime": 198144.865,
|
| 3317 |
+
"train_samples_per_second": 193.798,
|
| 3318 |
+
"train_steps_per_second": 12.112
|
| 3319 |
+
}
|
| 3320 |
+
],
|
| 3321 |
+
"logging_steps": 16000,
|
| 3322 |
+
"max_steps": 2400000,
|
| 3323 |
+
"num_train_epochs": 20,
|
| 3324 |
+
"save_steps": 32000,
|
| 3325 |
+
"total_flos": 7.178820925216543e+17,
|
| 3326 |
+
"trial_name": null,
|
| 3327 |
+
"trial_params": null
|
| 3328 |
+
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b6531054f40a076cf48758b3c5a78e56d07f7d3992f3e900341a389820e51c63
|
| 3 |
+
size 4219
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|