martimfasantos commited on
Commit
3751d9c
1 Parent(s): c619324

Model save

Browse files
README.md ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T
4
+ tags:
5
+ - trl
6
+ - sft
7
+ - generated_from_trainer
8
+ datasets:
9
+ - generator
10
+ model-index:
11
+ - name: tinyllama-1.1b-sum-sft-full_new
12
+ results: []
13
+ ---
14
+
15
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
16
+ should probably proofread and complete it, then remove this comment. -->
17
+
18
+ # tinyllama-1.1b-sum-sft-full_new
19
+
20
+ This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T](https://huggingface.co/TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T) on the generator dataset.
21
+ It achieves the following results on the evaluation set:
22
+ - Loss: 2.1350
23
+
24
+ ## Model description
25
+
26
+ More information needed
27
+
28
+ ## Intended uses & limitations
29
+
30
+ More information needed
31
+
32
+ ## Training and evaluation data
33
+
34
+ More information needed
35
+
36
+ ## Training procedure
37
+
38
+ ### Training hyperparameters
39
+
40
+ The following hyperparameters were used during training:
41
+ - learning_rate: 2e-05
42
+ - train_batch_size: 8
43
+ - eval_batch_size: 8
44
+ - seed: 42
45
+ - distributed_type: multi-GPU
46
+ - gradient_accumulation_steps: 2
47
+ - total_train_batch_size: 16
48
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
49
+ - lr_scheduler_type: cosine
50
+ - lr_scheduler_warmup_ratio: 0.1
51
+ - num_epochs: 1
52
+
53
+ ### Training results
54
+
55
+ | Training Loss | Epoch | Step | Validation Loss |
56
+ |:-------------:|:------:|:----:|:---------------:|
57
+ | 2.1307 | 0.9997 | 1476 | 2.1350 |
58
+
59
+
60
+ ### Framework versions
61
+
62
+ - Transformers 4.41.2
63
+ - Pytorch 2.1.2
64
+ - Datasets 2.19.2
65
+ - Tokenizers 0.19.1
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.9996613613274636,
3
+ "total_flos": 3.002086743311647e+17,
4
+ "train_loss": 2.154247896135015,
5
+ "train_runtime": 6599.6501,
6
+ "train_samples": 116722,
7
+ "train_samples_per_second": 3.579,
8
+ "train_steps_per_second": 0.224
9
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 1,
3
+ "eos_token_id": 2,
4
+ "max_length": 2048,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.41.2"
7
+ }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:614aafaa064cf07df3e391835e1c9161848a50bcf88bf83733a85b1a8c24a44c
3
  size 2200119864
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32ed8d9bca7ce2b203bf7188da32232b035aaf956a7022e8e38b959992ab8696
3
  size 2200119864
runs/Jun07_00-32-35_poseidon/events.out.tfevents.1717720368.poseidon.3182296.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6e4eaa08a9ce6f0d557261a36ae449d066568f395f0408d34913680b828c9b61
3
- size 64234
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:48a8eb237cef6f6c917d284848ccc1a84251281639ad0d902f086e5558903b64
3
+ size 68024
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.9996613613274636,
3
+ "total_flos": 3.002086743311647e+17,
4
+ "train_loss": 2.154247896135015,
5
+ "train_runtime": 6599.6501,
6
+ "train_samples": 116722,
7
+ "train_samples_per_second": 3.579,
8
+ "train_steps_per_second": 0.224
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,2122 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9996613613274636,
5
+ "eval_steps": 500,
6
+ "global_step": 1476,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0006772773450728074,
13
+ "grad_norm": 11.0625,
14
+ "learning_rate": 1.3513513513513515e-07,
15
+ "loss": 2.4477,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.0033863867253640365,
20
+ "grad_norm": 11.75,
21
+ "learning_rate": 6.756756756756758e-07,
22
+ "loss": 2.4743,
23
+ "step": 5
24
+ },
25
+ {
26
+ "epoch": 0.006772773450728073,
27
+ "grad_norm": 12.5625,
28
+ "learning_rate": 1.3513513513513515e-06,
29
+ "loss": 2.465,
30
+ "step": 10
31
+ },
32
+ {
33
+ "epoch": 0.01015916017609211,
34
+ "grad_norm": 11.375,
35
+ "learning_rate": 2.0270270270270273e-06,
36
+ "loss": 2.4495,
37
+ "step": 15
38
+ },
39
+ {
40
+ "epoch": 0.013545546901456146,
41
+ "grad_norm": 7.78125,
42
+ "learning_rate": 2.702702702702703e-06,
43
+ "loss": 2.4107,
44
+ "step": 20
45
+ },
46
+ {
47
+ "epoch": 0.016931933626820182,
48
+ "grad_norm": 4.28125,
49
+ "learning_rate": 3.3783783783783788e-06,
50
+ "loss": 2.4134,
51
+ "step": 25
52
+ },
53
+ {
54
+ "epoch": 0.02031832035218422,
55
+ "grad_norm": 2.96875,
56
+ "learning_rate": 4.0540540540540545e-06,
57
+ "loss": 2.3642,
58
+ "step": 30
59
+ },
60
+ {
61
+ "epoch": 0.023704707077548254,
62
+ "grad_norm": 2.25,
63
+ "learning_rate": 4.72972972972973e-06,
64
+ "loss": 2.3561,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.027091093802912292,
69
+ "grad_norm": 1.9453125,
70
+ "learning_rate": 5.405405405405406e-06,
71
+ "loss": 2.3428,
72
+ "step": 40
73
+ },
74
+ {
75
+ "epoch": 0.03047748052827633,
76
+ "grad_norm": 1.765625,
77
+ "learning_rate": 6.081081081081082e-06,
78
+ "loss": 2.2938,
79
+ "step": 45
80
+ },
81
+ {
82
+ "epoch": 0.033863867253640365,
83
+ "grad_norm": 1.6484375,
84
+ "learning_rate": 6.7567567567567575e-06,
85
+ "loss": 2.2739,
86
+ "step": 50
87
+ },
88
+ {
89
+ "epoch": 0.0372502539790044,
90
+ "grad_norm": 1.546875,
91
+ "learning_rate": 7.4324324324324324e-06,
92
+ "loss": 2.2748,
93
+ "step": 55
94
+ },
95
+ {
96
+ "epoch": 0.04063664070436844,
97
+ "grad_norm": 1.5078125,
98
+ "learning_rate": 8.108108108108109e-06,
99
+ "loss": 2.2662,
100
+ "step": 60
101
+ },
102
+ {
103
+ "epoch": 0.04402302742973248,
104
+ "grad_norm": 1.421875,
105
+ "learning_rate": 8.783783783783785e-06,
106
+ "loss": 2.2491,
107
+ "step": 65
108
+ },
109
+ {
110
+ "epoch": 0.04740941415509651,
111
+ "grad_norm": 1.4140625,
112
+ "learning_rate": 9.45945945945946e-06,
113
+ "loss": 2.2262,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.05079580088046055,
118
+ "grad_norm": 1.390625,
119
+ "learning_rate": 1.0135135135135136e-05,
120
+ "loss": 2.2425,
121
+ "step": 75
122
+ },
123
+ {
124
+ "epoch": 0.054182187605824585,
125
+ "grad_norm": 1.28125,
126
+ "learning_rate": 1.0810810810810812e-05,
127
+ "loss": 2.224,
128
+ "step": 80
129
+ },
130
+ {
131
+ "epoch": 0.05756857433118862,
132
+ "grad_norm": 1.328125,
133
+ "learning_rate": 1.1486486486486488e-05,
134
+ "loss": 2.1884,
135
+ "step": 85
136
+ },
137
+ {
138
+ "epoch": 0.06095496105655266,
139
+ "grad_norm": 1.2734375,
140
+ "learning_rate": 1.2162162162162164e-05,
141
+ "loss": 2.2151,
142
+ "step": 90
143
+ },
144
+ {
145
+ "epoch": 0.06434134778191669,
146
+ "grad_norm": 1.3046875,
147
+ "learning_rate": 1.283783783783784e-05,
148
+ "loss": 2.1939,
149
+ "step": 95
150
+ },
151
+ {
152
+ "epoch": 0.06772773450728073,
153
+ "grad_norm": 1.3359375,
154
+ "learning_rate": 1.3513513513513515e-05,
155
+ "loss": 2.1887,
156
+ "step": 100
157
+ },
158
+ {
159
+ "epoch": 0.07111412123264477,
160
+ "grad_norm": 1.2578125,
161
+ "learning_rate": 1.4189189189189189e-05,
162
+ "loss": 2.1825,
163
+ "step": 105
164
+ },
165
+ {
166
+ "epoch": 0.0745005079580088,
167
+ "grad_norm": 1.265625,
168
+ "learning_rate": 1.4864864864864865e-05,
169
+ "loss": 2.1925,
170
+ "step": 110
171
+ },
172
+ {
173
+ "epoch": 0.07788689468337284,
174
+ "grad_norm": 1.2734375,
175
+ "learning_rate": 1.554054054054054e-05,
176
+ "loss": 2.1969,
177
+ "step": 115
178
+ },
179
+ {
180
+ "epoch": 0.08127328140873688,
181
+ "grad_norm": 1.2578125,
182
+ "learning_rate": 1.6216216216216218e-05,
183
+ "loss": 2.1783,
184
+ "step": 120
185
+ },
186
+ {
187
+ "epoch": 0.08465966813410092,
188
+ "grad_norm": 1.2421875,
189
+ "learning_rate": 1.6891891891891896e-05,
190
+ "loss": 2.1641,
191
+ "step": 125
192
+ },
193
+ {
194
+ "epoch": 0.08804605485946496,
195
+ "grad_norm": 1.296875,
196
+ "learning_rate": 1.756756756756757e-05,
197
+ "loss": 2.2032,
198
+ "step": 130
199
+ },
200
+ {
201
+ "epoch": 0.091432441584829,
202
+ "grad_norm": 1.2734375,
203
+ "learning_rate": 1.8243243243243244e-05,
204
+ "loss": 2.2021,
205
+ "step": 135
206
+ },
207
+ {
208
+ "epoch": 0.09481882831019302,
209
+ "grad_norm": 1.2421875,
210
+ "learning_rate": 1.891891891891892e-05,
211
+ "loss": 2.1589,
212
+ "step": 140
213
+ },
214
+ {
215
+ "epoch": 0.09820521503555706,
216
+ "grad_norm": 1.2265625,
217
+ "learning_rate": 1.9594594594594595e-05,
218
+ "loss": 2.1598,
219
+ "step": 145
220
+ },
221
+ {
222
+ "epoch": 0.1015916017609211,
223
+ "grad_norm": 1.21875,
224
+ "learning_rate": 1.9999888073536733e-05,
225
+ "loss": 2.1588,
226
+ "step": 150
227
+ },
228
+ {
229
+ "epoch": 0.10497798848628513,
230
+ "grad_norm": 1.2578125,
231
+ "learning_rate": 1.9998628929598956e-05,
232
+ "loss": 2.1711,
233
+ "step": 155
234
+ },
235
+ {
236
+ "epoch": 0.10836437521164917,
237
+ "grad_norm": 1.203125,
238
+ "learning_rate": 1.9995970910394228e-05,
239
+ "loss": 2.1901,
240
+ "step": 160
241
+ },
242
+ {
243
+ "epoch": 0.11175076193701321,
244
+ "grad_norm": 1.2265625,
245
+ "learning_rate": 1.9991914387797266e-05,
246
+ "loss": 2.1554,
247
+ "step": 165
248
+ },
249
+ {
250
+ "epoch": 0.11513714866237724,
251
+ "grad_norm": 1.2265625,
252
+ "learning_rate": 1.99864599293428e-05,
253
+ "loss": 2.1558,
254
+ "step": 170
255
+ },
256
+ {
257
+ "epoch": 0.11852353538774128,
258
+ "grad_norm": 1.2265625,
259
+ "learning_rate": 1.997960829814616e-05,
260
+ "loss": 2.1718,
261
+ "step": 175
262
+ },
263
+ {
264
+ "epoch": 0.12190992211310532,
265
+ "grad_norm": 1.2421875,
266
+ "learning_rate": 1.9971360452796523e-05,
267
+ "loss": 2.171,
268
+ "step": 180
269
+ },
270
+ {
271
+ "epoch": 0.12529630883846934,
272
+ "grad_norm": 1.2109375,
273
+ "learning_rate": 1.9961717547222775e-05,
274
+ "loss": 2.1686,
275
+ "step": 185
276
+ },
277
+ {
278
+ "epoch": 0.12868269556383338,
279
+ "grad_norm": 1.265625,
280
+ "learning_rate": 1.9950680930532107e-05,
281
+ "loss": 2.1569,
282
+ "step": 190
283
+ },
284
+ {
285
+ "epoch": 0.13206908228919742,
286
+ "grad_norm": 1.234375,
287
+ "learning_rate": 1.9938252146821236e-05,
288
+ "loss": 2.1832,
289
+ "step": 195
290
+ },
291
+ {
292
+ "epoch": 0.13545546901456146,
293
+ "grad_norm": 1.21875,
294
+ "learning_rate": 1.9924432934960384e-05,
295
+ "loss": 2.1616,
296
+ "step": 200
297
+ },
298
+ {
299
+ "epoch": 0.1388418557399255,
300
+ "grad_norm": 1.1796875,
301
+ "learning_rate": 1.9909225228350007e-05,
302
+ "loss": 2.1798,
303
+ "step": 205
304
+ },
305
+ {
306
+ "epoch": 0.14222824246528953,
307
+ "grad_norm": 1.203125,
308
+ "learning_rate": 1.989263115465028e-05,
309
+ "loss": 2.1645,
310
+ "step": 210
311
+ },
312
+ {
313
+ "epoch": 0.14561462919065357,
314
+ "grad_norm": 1.2109375,
315
+ "learning_rate": 1.987465303548345e-05,
316
+ "loss": 2.1599,
317
+ "step": 215
318
+ },
319
+ {
320
+ "epoch": 0.1490010159160176,
321
+ "grad_norm": 1.203125,
322
+ "learning_rate": 1.9855293386108995e-05,
323
+ "loss": 2.1525,
324
+ "step": 220
325
+ },
326
+ {
327
+ "epoch": 0.15238740264138165,
328
+ "grad_norm": 1.265625,
329
+ "learning_rate": 1.9834554915071745e-05,
330
+ "loss": 2.1711,
331
+ "step": 225
332
+ },
333
+ {
334
+ "epoch": 0.15577378936674569,
335
+ "grad_norm": 1.21875,
336
+ "learning_rate": 1.981244052382293e-05,
337
+ "loss": 2.1597,
338
+ "step": 230
339
+ },
340
+ {
341
+ "epoch": 0.15916017609210972,
342
+ "grad_norm": 1.2109375,
343
+ "learning_rate": 1.978895330631425e-05,
344
+ "loss": 2.1511,
345
+ "step": 235
346
+ },
347
+ {
348
+ "epoch": 0.16254656281747376,
349
+ "grad_norm": 1.2109375,
350
+ "learning_rate": 1.976409654856501e-05,
351
+ "loss": 2.1579,
352
+ "step": 240
353
+ },
354
+ {
355
+ "epoch": 0.1659329495428378,
356
+ "grad_norm": 1.1953125,
357
+ "learning_rate": 1.9737873728202376e-05,
358
+ "loss": 2.1586,
359
+ "step": 245
360
+ },
361
+ {
362
+ "epoch": 0.16931933626820184,
363
+ "grad_norm": 1.1953125,
364
+ "learning_rate": 1.9710288513974846e-05,
365
+ "loss": 2.1634,
366
+ "step": 250
367
+ },
368
+ {
369
+ "epoch": 0.17270572299356587,
370
+ "grad_norm": 1.1953125,
371
+ "learning_rate": 1.9681344765238958e-05,
372
+ "loss": 2.1667,
373
+ "step": 255
374
+ },
375
+ {
376
+ "epoch": 0.1760921097189299,
377
+ "grad_norm": 1.25,
378
+ "learning_rate": 1.9651046531419335e-05,
379
+ "loss": 2.1732,
380
+ "step": 260
381
+ },
382
+ {
383
+ "epoch": 0.17947849644429395,
384
+ "grad_norm": 1.203125,
385
+ "learning_rate": 1.961939805144216e-05,
386
+ "loss": 2.1726,
387
+ "step": 265
388
+ },
389
+ {
390
+ "epoch": 0.182864883169658,
391
+ "grad_norm": 1.234375,
392
+ "learning_rate": 1.9586403753142104e-05,
393
+ "loss": 2.1538,
394
+ "step": 270
395
+ },
396
+ {
397
+ "epoch": 0.186251269895022,
398
+ "grad_norm": 1.2109375,
399
+ "learning_rate": 1.9552068252642858e-05,
400
+ "loss": 2.1728,
401
+ "step": 275
402
+ },
403
+ {
404
+ "epoch": 0.18963765662038604,
405
+ "grad_norm": 1.1953125,
406
+ "learning_rate": 1.9516396353711297e-05,
407
+ "loss": 2.1467,
408
+ "step": 280
409
+ },
410
+ {
411
+ "epoch": 0.19302404334575007,
412
+ "grad_norm": 1.1953125,
413
+ "learning_rate": 1.9479393047085392e-05,
414
+ "loss": 2.1676,
415
+ "step": 285
416
+ },
417
+ {
418
+ "epoch": 0.1964104300711141,
419
+ "grad_norm": 1.2109375,
420
+ "learning_rate": 1.9441063509776003e-05,
421
+ "loss": 2.1476,
422
+ "step": 290
423
+ },
424
+ {
425
+ "epoch": 0.19979681679647815,
426
+ "grad_norm": 1.1796875,
427
+ "learning_rate": 1.9401413104342535e-05,
428
+ "loss": 2.1346,
429
+ "step": 295
430
+ },
431
+ {
432
+ "epoch": 0.2031832035218422,
433
+ "grad_norm": 1.2265625,
434
+ "learning_rate": 1.936044737814273e-05,
435
+ "loss": 2.1629,
436
+ "step": 300
437
+ },
438
+ {
439
+ "epoch": 0.20656959024720623,
440
+ "grad_norm": 1.1796875,
441
+ "learning_rate": 1.931817206255651e-05,
442
+ "loss": 2.1322,
443
+ "step": 305
444
+ },
445
+ {
446
+ "epoch": 0.20995597697257026,
447
+ "grad_norm": 1.1953125,
448
+ "learning_rate": 1.9274593072184152e-05,
449
+ "loss": 2.1842,
450
+ "step": 310
451
+ },
452
+ {
453
+ "epoch": 0.2133423636979343,
454
+ "grad_norm": 1.1953125,
455
+ "learning_rate": 1.922971650401877e-05,
456
+ "loss": 2.133,
457
+ "step": 315
458
+ },
459
+ {
460
+ "epoch": 0.21672875042329834,
461
+ "grad_norm": 1.1953125,
462
+ "learning_rate": 1.9183548636593322e-05,
463
+ "loss": 2.1499,
464
+ "step": 320
465
+ },
466
+ {
467
+ "epoch": 0.22011513714866238,
468
+ "grad_norm": 1.2109375,
469
+ "learning_rate": 1.9136095929102204e-05,
470
+ "loss": 2.1444,
471
+ "step": 325
472
+ },
473
+ {
474
+ "epoch": 0.22350152387402641,
475
+ "grad_norm": 1.203125,
476
+ "learning_rate": 1.908736502049754e-05,
477
+ "loss": 2.1458,
478
+ "step": 330
479
+ },
480
+ {
481
+ "epoch": 0.22688791059939045,
482
+ "grad_norm": 1.203125,
483
+ "learning_rate": 1.903736272856038e-05,
484
+ "loss": 2.1738,
485
+ "step": 335
486
+ },
487
+ {
488
+ "epoch": 0.2302742973247545,
489
+ "grad_norm": 1.1875,
490
+ "learning_rate": 1.8986096048946826e-05,
491
+ "loss": 2.1413,
492
+ "step": 340
493
+ },
494
+ {
495
+ "epoch": 0.23366068405011853,
496
+ "grad_norm": 1.15625,
497
+ "learning_rate": 1.893357215420929e-05,
498
+ "loss": 2.1614,
499
+ "step": 345
500
+ },
501
+ {
502
+ "epoch": 0.23704707077548257,
503
+ "grad_norm": 1.21875,
504
+ "learning_rate": 1.8879798392793033e-05,
505
+ "loss": 2.1582,
506
+ "step": 350
507
+ },
508
+ {
509
+ "epoch": 0.2404334575008466,
510
+ "grad_norm": 1.171875,
511
+ "learning_rate": 1.8824782288008038e-05,
512
+ "loss": 2.1545,
513
+ "step": 355
514
+ },
515
+ {
516
+ "epoch": 0.24381984422621064,
517
+ "grad_norm": 1.2421875,
518
+ "learning_rate": 1.8768531536976452e-05,
519
+ "loss": 2.1497,
520
+ "step": 360
521
+ },
522
+ {
523
+ "epoch": 0.24720623095157468,
524
+ "grad_norm": 1.2109375,
525
+ "learning_rate": 1.8711054009555736e-05,
526
+ "loss": 2.1677,
527
+ "step": 365
528
+ },
529
+ {
530
+ "epoch": 0.2505926176769387,
531
+ "grad_norm": 1.203125,
532
+ "learning_rate": 1.865235774723758e-05,
533
+ "loss": 2.1475,
534
+ "step": 370
535
+ },
536
+ {
537
+ "epoch": 0.25397900440230275,
538
+ "grad_norm": 1.15625,
539
+ "learning_rate": 1.8592450962022872e-05,
540
+ "loss": 2.1422,
541
+ "step": 375
542
+ },
543
+ {
544
+ "epoch": 0.25736539112766676,
545
+ "grad_norm": 1.1875,
546
+ "learning_rate": 1.8531342035272768e-05,
547
+ "loss": 2.1679,
548
+ "step": 380
549
+ },
550
+ {
551
+ "epoch": 0.26075177785303083,
552
+ "grad_norm": 1.234375,
553
+ "learning_rate": 1.84690395165361e-05,
554
+ "loss": 2.1473,
555
+ "step": 385
556
+ },
557
+ {
558
+ "epoch": 0.26413816457839484,
559
+ "grad_norm": 1.1953125,
560
+ "learning_rate": 1.8405552122353212e-05,
561
+ "loss": 2.1614,
562
+ "step": 390
563
+ },
564
+ {
565
+ "epoch": 0.2675245513037589,
566
+ "grad_norm": 1.203125,
567
+ "learning_rate": 1.8340888735036485e-05,
568
+ "loss": 2.1299,
569
+ "step": 395
570
+ },
571
+ {
572
+ "epoch": 0.2709109380291229,
573
+ "grad_norm": 1.2109375,
574
+ "learning_rate": 1.8275058401427622e-05,
575
+ "loss": 2.1553,
576
+ "step": 400
577
+ },
578
+ {
579
+ "epoch": 0.274297324754487,
580
+ "grad_norm": 1.1875,
581
+ "learning_rate": 1.8208070331631943e-05,
582
+ "loss": 2.1445,
583
+ "step": 405
584
+ },
585
+ {
586
+ "epoch": 0.277683711479851,
587
+ "grad_norm": 1.21875,
588
+ "learning_rate": 1.8139933897729833e-05,
589
+ "loss": 2.1562,
590
+ "step": 410
591
+ },
592
+ {
593
+ "epoch": 0.28107009820521506,
594
+ "grad_norm": 1.1796875,
595
+ "learning_rate": 1.807065863246551e-05,
596
+ "loss": 2.1343,
597
+ "step": 415
598
+ },
599
+ {
600
+ "epoch": 0.28445648493057907,
601
+ "grad_norm": 1.203125,
602
+ "learning_rate": 1.8000254227913346e-05,
603
+ "loss": 2.1468,
604
+ "step": 420
605
+ },
606
+ {
607
+ "epoch": 0.28784287165594313,
608
+ "grad_norm": 1.1484375,
609
+ "learning_rate": 1.7928730534121872e-05,
610
+ "loss": 2.1492,
611
+ "step": 425
612
+ },
613
+ {
614
+ "epoch": 0.29122925838130714,
615
+ "grad_norm": 1.171875,
616
+ "learning_rate": 1.7856097557735697e-05,
617
+ "loss": 2.1326,
618
+ "step": 430
619
+ },
620
+ {
621
+ "epoch": 0.2946156451066712,
622
+ "grad_norm": 1.2421875,
623
+ "learning_rate": 1.77823654605955e-05,
624
+ "loss": 2.1491,
625
+ "step": 435
626
+ },
627
+ {
628
+ "epoch": 0.2980020318320352,
629
+ "grad_norm": 1.203125,
630
+ "learning_rate": 1.7707544558316332e-05,
631
+ "loss": 2.1425,
632
+ "step": 440
633
+ },
634
+ {
635
+ "epoch": 0.30138841855739923,
636
+ "grad_norm": 1.171875,
637
+ "learning_rate": 1.763164531884439e-05,
638
+ "loss": 2.1356,
639
+ "step": 445
640
+ },
641
+ {
642
+ "epoch": 0.3047748052827633,
643
+ "grad_norm": 1.1640625,
644
+ "learning_rate": 1.7554678360992475e-05,
645
+ "loss": 2.1459,
646
+ "step": 450
647
+ },
648
+ {
649
+ "epoch": 0.3081611920081273,
650
+ "grad_norm": 1.21875,
651
+ "learning_rate": 1.7476654452954354e-05,
652
+ "loss": 2.1232,
653
+ "step": 455
654
+ },
655
+ {
656
+ "epoch": 0.31154757873349137,
657
+ "grad_norm": 1.1875,
658
+ "learning_rate": 1.7397584510798208e-05,
659
+ "loss": 2.1333,
660
+ "step": 460
661
+ },
662
+ {
663
+ "epoch": 0.3149339654588554,
664
+ "grad_norm": 1.1796875,
665
+ "learning_rate": 1.7317479596939426e-05,
666
+ "loss": 2.145,
667
+ "step": 465
668
+ },
669
+ {
670
+ "epoch": 0.31832035218421945,
671
+ "grad_norm": 1.234375,
672
+ "learning_rate": 1.723635091859287e-05,
673
+ "loss": 2.1452,
674
+ "step": 470
675
+ },
676
+ {
677
+ "epoch": 0.32170673890958346,
678
+ "grad_norm": 1.234375,
679
+ "learning_rate": 1.7154209826204916e-05,
680
+ "loss": 2.155,
681
+ "step": 475
682
+ },
683
+ {
684
+ "epoch": 0.3250931256349475,
685
+ "grad_norm": 1.1796875,
686
+ "learning_rate": 1.7071067811865477e-05,
687
+ "loss": 2.15,
688
+ "step": 480
689
+ },
690
+ {
691
+ "epoch": 0.32847951236031153,
692
+ "grad_norm": 1.203125,
693
+ "learning_rate": 1.698693650770014e-05,
694
+ "loss": 2.1436,
695
+ "step": 485
696
+ },
697
+ {
698
+ "epoch": 0.3318658990856756,
699
+ "grad_norm": 1.1875,
700
+ "learning_rate": 1.690182768424279e-05,
701
+ "loss": 2.1471,
702
+ "step": 490
703
+ },
704
+ {
705
+ "epoch": 0.3352522858110396,
706
+ "grad_norm": 1.140625,
707
+ "learning_rate": 1.6815753248788814e-05,
708
+ "loss": 2.1616,
709
+ "step": 495
710
+ },
711
+ {
712
+ "epoch": 0.3386386725364037,
713
+ "grad_norm": 1.203125,
714
+ "learning_rate": 1.672872524372919e-05,
715
+ "loss": 2.1325,
716
+ "step": 500
717
+ },
718
+ {
719
+ "epoch": 0.3420250592617677,
720
+ "grad_norm": 1.2109375,
721
+ "learning_rate": 1.6640755844865693e-05,
722
+ "loss": 2.1278,
723
+ "step": 505
724
+ },
725
+ {
726
+ "epoch": 0.34541144598713175,
727
+ "grad_norm": 1.1796875,
728
+ "learning_rate": 1.6551857359707407e-05,
729
+ "loss": 2.1371,
730
+ "step": 510
731
+ },
732
+ {
733
+ "epoch": 0.34879783271249576,
734
+ "grad_norm": 1.203125,
735
+ "learning_rate": 1.6462042225748803e-05,
736
+ "loss": 2.1385,
737
+ "step": 515
738
+ },
739
+ {
740
+ "epoch": 0.3521842194378598,
741
+ "grad_norm": 1.21875,
742
+ "learning_rate": 1.637132300872969e-05,
743
+ "loss": 2.1161,
744
+ "step": 520
745
+ },
746
+ {
747
+ "epoch": 0.35557060616322383,
748
+ "grad_norm": 1.21875,
749
+ "learning_rate": 1.627971240087715e-05,
750
+ "loss": 2.1194,
751
+ "step": 525
752
+ },
753
+ {
754
+ "epoch": 0.3589569928885879,
755
+ "grad_norm": 1.2109375,
756
+ "learning_rate": 1.6187223219129825e-05,
757
+ "loss": 2.1456,
758
+ "step": 530
759
+ },
760
+ {
761
+ "epoch": 0.3623433796139519,
762
+ "grad_norm": 1.1796875,
763
+ "learning_rate": 1.609386840334475e-05,
764
+ "loss": 2.1475,
765
+ "step": 535
766
+ },
767
+ {
768
+ "epoch": 0.365729766339316,
769
+ "grad_norm": 1.1640625,
770
+ "learning_rate": 1.5999661014486956e-05,
771
+ "loss": 2.144,
772
+ "step": 540
773
+ },
774
+ {
775
+ "epoch": 0.36911615306468,
776
+ "grad_norm": 1.1953125,
777
+ "learning_rate": 1.5904614232802173e-05,
778
+ "loss": 2.1698,
779
+ "step": 545
780
+ },
781
+ {
782
+ "epoch": 0.372502539790044,
783
+ "grad_norm": 1.1328125,
784
+ "learning_rate": 1.5808741355972808e-05,
785
+ "loss": 2.143,
786
+ "step": 550
787
+ },
788
+ {
789
+ "epoch": 0.37588892651540806,
790
+ "grad_norm": 1.140625,
791
+ "learning_rate": 1.5712055797257542e-05,
792
+ "loss": 2.1519,
793
+ "step": 555
794
+ },
795
+ {
796
+ "epoch": 0.37927531324077207,
797
+ "grad_norm": 1.1796875,
798
+ "learning_rate": 1.5614571083614683e-05,
799
+ "loss": 2.1537,
800
+ "step": 560
801
+ },
802
+ {
803
+ "epoch": 0.38266169996613614,
804
+ "grad_norm": 1.1953125,
805
+ "learning_rate": 1.5516300853809688e-05,
806
+ "loss": 2.1435,
807
+ "step": 565
808
+ },
809
+ {
810
+ "epoch": 0.38604808669150015,
811
+ "grad_norm": 1.15625,
812
+ "learning_rate": 1.5417258856506995e-05,
813
+ "loss": 2.151,
814
+ "step": 570
815
+ },
816
+ {
817
+ "epoch": 0.3894344734168642,
818
+ "grad_norm": 1.1875,
819
+ "learning_rate": 1.5317458948346482e-05,
820
+ "loss": 2.1489,
821
+ "step": 575
822
+ },
823
+ {
824
+ "epoch": 0.3928208601422282,
825
+ "grad_norm": 1.2109375,
826
+ "learning_rate": 1.5216915092004847e-05,
827
+ "loss": 2.136,
828
+ "step": 580
829
+ },
830
+ {
831
+ "epoch": 0.3962072468675923,
832
+ "grad_norm": 1.1875,
833
+ "learning_rate": 1.5115641354242135e-05,
834
+ "loss": 2.1468,
835
+ "step": 585
836
+ },
837
+ {
838
+ "epoch": 0.3995936335929563,
839
+ "grad_norm": 1.1640625,
840
+ "learning_rate": 1.5013651903933686e-05,
841
+ "loss": 2.1242,
842
+ "step": 590
843
+ },
844
+ {
845
+ "epoch": 0.40298002031832036,
846
+ "grad_norm": 1.1796875,
847
+ "learning_rate": 1.491096101008784e-05,
848
+ "loss": 2.1387,
849
+ "step": 595
850
+ },
851
+ {
852
+ "epoch": 0.4063664070436844,
853
+ "grad_norm": 1.203125,
854
+ "learning_rate": 1.4807583039849589e-05,
855
+ "loss": 2.1374,
856
+ "step": 600
857
+ },
858
+ {
859
+ "epoch": 0.40975279376904844,
860
+ "grad_norm": 1.1875,
861
+ "learning_rate": 1.4703532456490518e-05,
862
+ "loss": 2.1485,
863
+ "step": 605
864
+ },
865
+ {
866
+ "epoch": 0.41313918049441245,
867
+ "grad_norm": 1.171875,
868
+ "learning_rate": 1.45988238173853e-05,
869
+ "loss": 2.1299,
870
+ "step": 610
871
+ },
872
+ {
873
+ "epoch": 0.4165255672197765,
874
+ "grad_norm": 1.15625,
875
+ "learning_rate": 1.449347177197503e-05,
876
+ "loss": 2.1299,
877
+ "step": 615
878
+ },
879
+ {
880
+ "epoch": 0.4199119539451405,
881
+ "grad_norm": 1.1875,
882
+ "learning_rate": 1.4387491059717653e-05,
883
+ "loss": 2.1355,
884
+ "step": 620
885
+ },
886
+ {
887
+ "epoch": 0.4232983406705046,
888
+ "grad_norm": 1.1875,
889
+ "learning_rate": 1.4280896508025825e-05,
890
+ "loss": 2.1481,
891
+ "step": 625
892
+ },
893
+ {
894
+ "epoch": 0.4266847273958686,
895
+ "grad_norm": 1.171875,
896
+ "learning_rate": 1.4173703030192467e-05,
897
+ "loss": 2.1442,
898
+ "step": 630
899
+ },
900
+ {
901
+ "epoch": 0.43007111412123267,
902
+ "grad_norm": 1.203125,
903
+ "learning_rate": 1.4065925623304267e-05,
904
+ "loss": 2.1465,
905
+ "step": 635
906
+ },
907
+ {
908
+ "epoch": 0.4334575008465967,
909
+ "grad_norm": 1.2265625,
910
+ "learning_rate": 1.3957579366143521e-05,
911
+ "loss": 2.1553,
912
+ "step": 640
913
+ },
914
+ {
915
+ "epoch": 0.43684388757196074,
916
+ "grad_norm": 1.1484375,
917
+ "learning_rate": 1.3848679417078487e-05,
918
+ "loss": 2.144,
919
+ "step": 645
920
+ },
921
+ {
922
+ "epoch": 0.44023027429732475,
923
+ "grad_norm": 1.203125,
924
+ "learning_rate": 1.373924101194264e-05,
925
+ "loss": 2.1277,
926
+ "step": 650
927
+ },
928
+ {
929
+ "epoch": 0.4436166610226888,
930
+ "grad_norm": 1.1875,
931
+ "learning_rate": 1.3629279461903064e-05,
932
+ "loss": 2.1534,
933
+ "step": 655
934
+ },
935
+ {
936
+ "epoch": 0.44700304774805283,
937
+ "grad_norm": 1.1875,
938
+ "learning_rate": 1.351881015131833e-05,
939
+ "loss": 2.1327,
940
+ "step": 660
941
+ },
942
+ {
943
+ "epoch": 0.45038943447341684,
944
+ "grad_norm": 1.1640625,
945
+ "learning_rate": 1.3407848535586122e-05,
946
+ "loss": 2.1315,
947
+ "step": 665
948
+ },
949
+ {
950
+ "epoch": 0.4537758211987809,
951
+ "grad_norm": 1.2265625,
952
+ "learning_rate": 1.3296410138980903e-05,
953
+ "loss": 2.1231,
954
+ "step": 670
955
+ },
956
+ {
957
+ "epoch": 0.4571622079241449,
958
+ "grad_norm": 1.1953125,
959
+ "learning_rate": 1.3184510552481988e-05,
960
+ "loss": 2.1503,
961
+ "step": 675
962
+ },
963
+ {
964
+ "epoch": 0.460548594649509,
965
+ "grad_norm": 1.1796875,
966
+ "learning_rate": 1.307216543159225e-05,
967
+ "loss": 2.1274,
968
+ "step": 680
969
+ },
970
+ {
971
+ "epoch": 0.463934981374873,
972
+ "grad_norm": 1.1484375,
973
+ "learning_rate": 1.2959390494147814e-05,
974
+ "loss": 2.1279,
975
+ "step": 685
976
+ },
977
+ {
978
+ "epoch": 0.46732136810023706,
979
+ "grad_norm": 1.171875,
980
+ "learning_rate": 1.2846201518119019e-05,
981
+ "loss": 2.1436,
982
+ "step": 690
983
+ },
984
+ {
985
+ "epoch": 0.47070775482560107,
986
+ "grad_norm": 1.1875,
987
+ "learning_rate": 1.273261433940299e-05,
988
+ "loss": 2.1639,
989
+ "step": 695
990
+ },
991
+ {
992
+ "epoch": 0.47409414155096513,
993
+ "grad_norm": 1.1796875,
994
+ "learning_rate": 1.2618644849608068e-05,
995
+ "loss": 2.1305,
996
+ "step": 700
997
+ },
998
+ {
999
+ "epoch": 0.47748052827632914,
1000
+ "grad_norm": 1.1953125,
1001
+ "learning_rate": 1.2504308993830488e-05,
1002
+ "loss": 2.1595,
1003
+ "step": 705
1004
+ },
1005
+ {
1006
+ "epoch": 0.4808669150016932,
1007
+ "grad_norm": 1.1953125,
1008
+ "learning_rate": 1.2389622768423537e-05,
1009
+ "loss": 2.1304,
1010
+ "step": 710
1011
+ },
1012
+ {
1013
+ "epoch": 0.4842533017270572,
1014
+ "grad_norm": 1.1796875,
1015
+ "learning_rate": 1.2274602218759569e-05,
1016
+ "loss": 2.1382,
1017
+ "step": 715
1018
+ },
1019
+ {
1020
+ "epoch": 0.4876396884524213,
1021
+ "grad_norm": 1.1640625,
1022
+ "learning_rate": 1.2159263436985139e-05,
1023
+ "loss": 2.1279,
1024
+ "step": 720
1025
+ },
1026
+ {
1027
+ "epoch": 0.4910260751777853,
1028
+ "grad_norm": 1.1953125,
1029
+ "learning_rate": 1.2043622559769614e-05,
1030
+ "loss": 2.1422,
1031
+ "step": 725
1032
+ },
1033
+ {
1034
+ "epoch": 0.49441246190314936,
1035
+ "grad_norm": 1.1875,
1036
+ "learning_rate": 1.192769576604754e-05,
1037
+ "loss": 2.1352,
1038
+ "step": 730
1039
+ },
1040
+ {
1041
+ "epoch": 0.49779884862851337,
1042
+ "grad_norm": 1.234375,
1043
+ "learning_rate": 1.1811499274755099e-05,
1044
+ "loss": 2.1342,
1045
+ "step": 735
1046
+ },
1047
+ {
1048
+ "epoch": 0.5011852353538774,
1049
+ "grad_norm": 1.1875,
1050
+ "learning_rate": 1.1695049342560969e-05,
1051
+ "loss": 2.149,
1052
+ "step": 740
1053
+ },
1054
+ {
1055
+ "epoch": 0.5045716220792414,
1056
+ "grad_norm": 1.1796875,
1057
+ "learning_rate": 1.1578362261591916e-05,
1058
+ "loss": 2.1446,
1059
+ "step": 745
1060
+ },
1061
+ {
1062
+ "epoch": 0.5079580088046055,
1063
+ "grad_norm": 1.1796875,
1064
+ "learning_rate": 1.1461454357153407e-05,
1065
+ "loss": 2.166,
1066
+ "step": 750
1067
+ },
1068
+ {
1069
+ "epoch": 0.5113443955299696,
1070
+ "grad_norm": 1.140625,
1071
+ "learning_rate": 1.1344341985445594e-05,
1072
+ "loss": 2.1116,
1073
+ "step": 755
1074
+ },
1075
+ {
1076
+ "epoch": 0.5147307822553335,
1077
+ "grad_norm": 1.1640625,
1078
+ "learning_rate": 1.1227041531274978e-05,
1079
+ "loss": 2.1335,
1080
+ "step": 760
1081
+ },
1082
+ {
1083
+ "epoch": 0.5181171689806976,
1084
+ "grad_norm": 1.1796875,
1085
+ "learning_rate": 1.1109569405762051e-05,
1086
+ "loss": 2.1368,
1087
+ "step": 765
1088
+ },
1089
+ {
1090
+ "epoch": 0.5215035557060617,
1091
+ "grad_norm": 1.1640625,
1092
+ "learning_rate": 1.0991942044045277e-05,
1093
+ "loss": 2.1468,
1094
+ "step": 770
1095
+ },
1096
+ {
1097
+ "epoch": 0.5248899424314256,
1098
+ "grad_norm": 1.1796875,
1099
+ "learning_rate": 1.0874175902981711e-05,
1100
+ "loss": 2.1288,
1101
+ "step": 775
1102
+ },
1103
+ {
1104
+ "epoch": 0.5282763291567897,
1105
+ "grad_norm": 1.171875,
1106
+ "learning_rate": 1.075628745884457e-05,
1107
+ "loss": 2.1461,
1108
+ "step": 780
1109
+ },
1110
+ {
1111
+ "epoch": 0.5316627158821537,
1112
+ "grad_norm": 1.1640625,
1113
+ "learning_rate": 1.0638293205018084e-05,
1114
+ "loss": 2.1475,
1115
+ "step": 785
1116
+ },
1117
+ {
1118
+ "epoch": 0.5350491026075178,
1119
+ "grad_norm": 1.15625,
1120
+ "learning_rate": 1.0520209649689979e-05,
1121
+ "loss": 2.1375,
1122
+ "step": 790
1123
+ },
1124
+ {
1125
+ "epoch": 0.5384354893328818,
1126
+ "grad_norm": 1.1640625,
1127
+ "learning_rate": 1.0402053313541854e-05,
1128
+ "loss": 2.1267,
1129
+ "step": 795
1130
+ },
1131
+ {
1132
+ "epoch": 0.5418218760582458,
1133
+ "grad_norm": 1.1796875,
1134
+ "learning_rate": 1.0283840727437832e-05,
1135
+ "loss": 2.1394,
1136
+ "step": 800
1137
+ },
1138
+ {
1139
+ "epoch": 0.5452082627836099,
1140
+ "grad_norm": 1.1953125,
1141
+ "learning_rate": 1.0165588430111787e-05,
1142
+ "loss": 2.1366,
1143
+ "step": 805
1144
+ },
1145
+ {
1146
+ "epoch": 0.548594649508974,
1147
+ "grad_norm": 1.1640625,
1148
+ "learning_rate": 1.0047312965853456e-05,
1149
+ "loss": 2.1366,
1150
+ "step": 810
1151
+ },
1152
+ {
1153
+ "epoch": 0.5519810362343379,
1154
+ "grad_norm": 1.1953125,
1155
+ "learning_rate": 9.92903088219378e-06,
1156
+ "loss": 2.1493,
1157
+ "step": 815
1158
+ },
1159
+ {
1160
+ "epoch": 0.555367422959702,
1161
+ "grad_norm": 1.1875,
1162
+ "learning_rate": 9.810758727589814e-06,
1163
+ "loss": 2.1522,
1164
+ "step": 820
1165
+ },
1166
+ {
1167
+ "epoch": 0.558753809685066,
1168
+ "grad_norm": 1.2109375,
1169
+ "learning_rate": 9.69251304910945e-06,
1170
+ "loss": 2.1298,
1171
+ "step": 825
1172
+ },
1173
+ {
1174
+ "epoch": 0.5621401964104301,
1175
+ "grad_norm": 1.2109375,
1176
+ "learning_rate": 9.574310390116421e-06,
1177
+ "loss": 2.1347,
1178
+ "step": 830
1179
+ },
1180
+ {
1181
+ "epoch": 0.5655265831357941,
1182
+ "grad_norm": 1.203125,
1183
+ "learning_rate": 9.45616728795571e-06,
1184
+ "loss": 2.1401,
1185
+ "step": 835
1186
+ },
1187
+ {
1188
+ "epoch": 0.5689129698611581,
1189
+ "grad_norm": 1.171875,
1190
+ "learning_rate": 9.338100271639932e-06,
1191
+ "loss": 2.128,
1192
+ "step": 840
1193
+ },
1194
+ {
1195
+ "epoch": 0.5722993565865222,
1196
+ "grad_norm": 1.15625,
1197
+ "learning_rate": 9.220125859536766e-06,
1198
+ "loss": 2.1333,
1199
+ "step": 845
1200
+ },
1201
+ {
1202
+ "epoch": 0.5756857433118863,
1203
+ "grad_norm": 1.1796875,
1204
+ "learning_rate": 9.102260557057936e-06,
1205
+ "loss": 2.1386,
1206
+ "step": 850
1207
+ },
1208
+ {
1209
+ "epoch": 0.5790721300372502,
1210
+ "grad_norm": 1.15625,
1211
+ "learning_rate": 8.984520854350014e-06,
1212
+ "loss": 2.1107,
1213
+ "step": 855
1214
+ },
1215
+ {
1216
+ "epoch": 0.5824585167626143,
1217
+ "grad_norm": 1.1875,
1218
+ "learning_rate": 8.866923223987303e-06,
1219
+ "loss": 2.1386,
1220
+ "step": 860
1221
+ },
1222
+ {
1223
+ "epoch": 0.5858449034879784,
1224
+ "grad_norm": 1.1953125,
1225
+ "learning_rate": 8.749484118667242e-06,
1226
+ "loss": 2.1313,
1227
+ "step": 865
1228
+ },
1229
+ {
1230
+ "epoch": 0.5892312902133424,
1231
+ "grad_norm": 1.1953125,
1232
+ "learning_rate": 8.632219968908556e-06,
1233
+ "loss": 2.1424,
1234
+ "step": 870
1235
+ },
1236
+ {
1237
+ "epoch": 0.5926176769387064,
1238
+ "grad_norm": 1.1796875,
1239
+ "learning_rate": 8.515147180752504e-06,
1240
+ "loss": 2.1406,
1241
+ "step": 875
1242
+ },
1243
+ {
1244
+ "epoch": 0.5960040636640704,
1245
+ "grad_norm": 1.1953125,
1246
+ "learning_rate": 8.398282133467579e-06,
1247
+ "loss": 2.1324,
1248
+ "step": 880
1249
+ },
1250
+ {
1251
+ "epoch": 0.5993904503894345,
1252
+ "grad_norm": 1.140625,
1253
+ "learning_rate": 8.281641177257929e-06,
1254
+ "loss": 2.128,
1255
+ "step": 885
1256
+ },
1257
+ {
1258
+ "epoch": 0.6027768371147985,
1259
+ "grad_norm": 1.1640625,
1260
+ "learning_rate": 8.165240630975863e-06,
1261
+ "loss": 2.1153,
1262
+ "step": 890
1263
+ },
1264
+ {
1265
+ "epoch": 0.6061632238401625,
1266
+ "grad_norm": 1.1953125,
1267
+ "learning_rate": 8.04909677983872e-06,
1268
+ "loss": 2.1682,
1269
+ "step": 895
1270
+ },
1271
+ {
1272
+ "epoch": 0.6095496105655266,
1273
+ "grad_norm": 1.1640625,
1274
+ "learning_rate": 7.93322587315047e-06,
1275
+ "loss": 2.1152,
1276
+ "step": 900
1277
+ },
1278
+ {
1279
+ "epoch": 0.6129359972908907,
1280
+ "grad_norm": 1.1484375,
1281
+ "learning_rate": 7.81764412202833e-06,
1282
+ "loss": 2.1311,
1283
+ "step": 905
1284
+ },
1285
+ {
1286
+ "epoch": 0.6163223840162546,
1287
+ "grad_norm": 1.140625,
1288
+ "learning_rate": 7.702367697134701e-06,
1289
+ "loss": 2.1408,
1290
+ "step": 910
1291
+ },
1292
+ {
1293
+ "epoch": 0.6197087707416187,
1294
+ "grad_norm": 1.1484375,
1295
+ "learning_rate": 7.5874127264148075e-06,
1296
+ "loss": 2.145,
1297
+ "step": 915
1298
+ },
1299
+ {
1300
+ "epoch": 0.6230951574669827,
1301
+ "grad_norm": 1.15625,
1302
+ "learning_rate": 7.4727952928402695e-06,
1303
+ "loss": 2.1318,
1304
+ "step": 920
1305
+ },
1306
+ {
1307
+ "epoch": 0.6264815441923468,
1308
+ "grad_norm": 1.15625,
1309
+ "learning_rate": 7.358531432159002e-06,
1310
+ "loss": 2.1231,
1311
+ "step": 925
1312
+ },
1313
+ {
1314
+ "epoch": 0.6298679309177108,
1315
+ "grad_norm": 1.15625,
1316
+ "learning_rate": 7.2446371306516935e-06,
1317
+ "loss": 2.1285,
1318
+ "step": 930
1319
+ },
1320
+ {
1321
+ "epoch": 0.6332543176430748,
1322
+ "grad_norm": 1.1640625,
1323
+ "learning_rate": 7.1311283228952345e-06,
1324
+ "loss": 2.1384,
1325
+ "step": 935
1326
+ },
1327
+ {
1328
+ "epoch": 0.6366407043684389,
1329
+ "grad_norm": 1.1953125,
1330
+ "learning_rate": 7.018020889533348e-06,
1331
+ "loss": 2.1409,
1332
+ "step": 940
1333
+ },
1334
+ {
1335
+ "epoch": 0.640027091093803,
1336
+ "grad_norm": 1.140625,
1337
+ "learning_rate": 6.905330655054782e-06,
1338
+ "loss": 2.1417,
1339
+ "step": 945
1340
+ },
1341
+ {
1342
+ "epoch": 0.6434134778191669,
1343
+ "grad_norm": 1.15625,
1344
+ "learning_rate": 6.793073385579373e-06,
1345
+ "loss": 2.1375,
1346
+ "step": 950
1347
+ },
1348
+ {
1349
+ "epoch": 0.646799864544531,
1350
+ "grad_norm": 1.1953125,
1351
+ "learning_rate": 6.681264786652236e-06,
1352
+ "loss": 2.1219,
1353
+ "step": 955
1354
+ },
1355
+ {
1356
+ "epoch": 0.650186251269895,
1357
+ "grad_norm": 1.203125,
1358
+ "learning_rate": 6.569920501046474e-06,
1359
+ "loss": 2.1363,
1360
+ "step": 960
1361
+ },
1362
+ {
1363
+ "epoch": 0.6535726379952591,
1364
+ "grad_norm": 1.171875,
1365
+ "learning_rate": 6.459056106574646e-06,
1366
+ "loss": 2.1209,
1367
+ "step": 965
1368
+ },
1369
+ {
1370
+ "epoch": 0.6569590247206231,
1371
+ "grad_norm": 1.1796875,
1372
+ "learning_rate": 6.348687113909304e-06,
1373
+ "loss": 2.1254,
1374
+ "step": 970
1375
+ },
1376
+ {
1377
+ "epoch": 0.6603454114459871,
1378
+ "grad_norm": 1.1796875,
1379
+ "learning_rate": 6.238828964412989e-06,
1380
+ "loss": 2.1466,
1381
+ "step": 975
1382
+ },
1383
+ {
1384
+ "epoch": 0.6637317981713512,
1385
+ "grad_norm": 1.2109375,
1386
+ "learning_rate": 6.129497027977829e-06,
1387
+ "loss": 2.1163,
1388
+ "step": 980
1389
+ },
1390
+ {
1391
+ "epoch": 0.6671181848967151,
1392
+ "grad_norm": 1.15625,
1393
+ "learning_rate": 6.020706600875237e-06,
1394
+ "loss": 2.1078,
1395
+ "step": 985
1396
+ },
1397
+ {
1398
+ "epoch": 0.6705045716220792,
1399
+ "grad_norm": 1.140625,
1400
+ "learning_rate": 5.912472903615821e-06,
1401
+ "loss": 2.1486,
1402
+ "step": 990
1403
+ },
1404
+ {
1405
+ "epoch": 0.6738909583474433,
1406
+ "grad_norm": 1.1953125,
1407
+ "learning_rate": 5.804811078819956e-06,
1408
+ "loss": 2.1399,
1409
+ "step": 995
1410
+ },
1411
+ {
1412
+ "epoch": 0.6772773450728073,
1413
+ "grad_norm": 1.1875,
1414
+ "learning_rate": 5.69773618909923e-06,
1415
+ "loss": 2.137,
1416
+ "step": 1000
1417
+ },
1418
+ {
1419
+ "epoch": 0.6806637317981713,
1420
+ "grad_norm": 1.1875,
1421
+ "learning_rate": 5.591263214949059e-06,
1422
+ "loss": 2.1506,
1423
+ "step": 1005
1424
+ },
1425
+ {
1426
+ "epoch": 0.6840501185235354,
1427
+ "grad_norm": 1.1796875,
1428
+ "learning_rate": 5.4854070526528445e-06,
1429
+ "loss": 2.142,
1430
+ "step": 1010
1431
+ },
1432
+ {
1433
+ "epoch": 0.6874365052488994,
1434
+ "grad_norm": 1.171875,
1435
+ "learning_rate": 5.380182512197873e-06,
1436
+ "loss": 2.1231,
1437
+ "step": 1015
1438
+ },
1439
+ {
1440
+ "epoch": 0.6908228919742635,
1441
+ "grad_norm": 1.171875,
1442
+ "learning_rate": 5.2756043152032934e-06,
1443
+ "loss": 2.1146,
1444
+ "step": 1020
1445
+ },
1446
+ {
1447
+ "epoch": 0.6942092786996275,
1448
+ "grad_norm": 1.1875,
1449
+ "learning_rate": 5.171687092860464e-06,
1450
+ "loss": 2.1446,
1451
+ "step": 1025
1452
+ },
1453
+ {
1454
+ "epoch": 0.6975956654249915,
1455
+ "grad_norm": 1.171875,
1456
+ "learning_rate": 5.068445383885961e-06,
1457
+ "loss": 2.1376,
1458
+ "step": 1030
1459
+ },
1460
+ {
1461
+ "epoch": 0.7009820521503556,
1462
+ "grad_norm": 1.1796875,
1463
+ "learning_rate": 4.965893632487511e-06,
1464
+ "loss": 2.139,
1465
+ "step": 1035
1466
+ },
1467
+ {
1468
+ "epoch": 0.7043684388757196,
1469
+ "grad_norm": 1.1640625,
1470
+ "learning_rate": 4.86404618634314e-06,
1471
+ "loss": 2.1404,
1472
+ "step": 1040
1473
+ },
1474
+ {
1475
+ "epoch": 0.7077548256010836,
1476
+ "grad_norm": 1.15625,
1477
+ "learning_rate": 4.762917294593851e-06,
1478
+ "loss": 2.1275,
1479
+ "step": 1045
1480
+ },
1481
+ {
1482
+ "epoch": 0.7111412123264477,
1483
+ "grad_norm": 1.1796875,
1484
+ "learning_rate": 4.662521105850056e-06,
1485
+ "loss": 2.1326,
1486
+ "step": 1050
1487
+ },
1488
+ {
1489
+ "epoch": 0.7145275990518117,
1490
+ "grad_norm": 1.1640625,
1491
+ "learning_rate": 4.562871666212107e-06,
1492
+ "loss": 2.1461,
1493
+ "step": 1055
1494
+ },
1495
+ {
1496
+ "epoch": 0.7179139857771758,
1497
+ "grad_norm": 1.1484375,
1498
+ "learning_rate": 4.463982917305155e-06,
1499
+ "loss": 2.1244,
1500
+ "step": 1060
1501
+ },
1502
+ {
1503
+ "epoch": 0.7213003725025398,
1504
+ "grad_norm": 1.1796875,
1505
+ "learning_rate": 4.365868694328599e-06,
1506
+ "loss": 2.1384,
1507
+ "step": 1065
1508
+ },
1509
+ {
1510
+ "epoch": 0.7246867592279038,
1511
+ "grad_norm": 1.171875,
1512
+ "learning_rate": 4.2685427241204755e-06,
1513
+ "loss": 2.1444,
1514
+ "step": 1070
1515
+ },
1516
+ {
1517
+ "epoch": 0.7280731459532679,
1518
+ "grad_norm": 1.1875,
1519
+ "learning_rate": 4.1720186232369595e-06,
1520
+ "loss": 2.1365,
1521
+ "step": 1075
1522
+ },
1523
+ {
1524
+ "epoch": 0.731459532678632,
1525
+ "grad_norm": 1.2109375,
1526
+ "learning_rate": 4.076309896047337e-06,
1527
+ "loss": 2.1302,
1528
+ "step": 1080
1529
+ },
1530
+ {
1531
+ "epoch": 0.7348459194039959,
1532
+ "grad_norm": 1.171875,
1533
+ "learning_rate": 3.981429932844641e-06,
1534
+ "loss": 2.1665,
1535
+ "step": 1085
1536
+ },
1537
+ {
1538
+ "epoch": 0.73823230612936,
1539
+ "grad_norm": 1.2265625,
1540
+ "learning_rate": 3.887392007972265e-06,
1541
+ "loss": 2.1583,
1542
+ "step": 1090
1543
+ },
1544
+ {
1545
+ "epoch": 0.741618692854724,
1546
+ "grad_norm": 1.1796875,
1547
+ "learning_rate": 3.7942092779667904e-06,
1548
+ "loss": 2.1263,
1549
+ "step": 1095
1550
+ },
1551
+ {
1552
+ "epoch": 0.745005079580088,
1553
+ "grad_norm": 1.1953125,
1554
+ "learning_rate": 3.7018947797172864e-06,
1555
+ "loss": 2.1303,
1556
+ "step": 1100
1557
+ },
1558
+ {
1559
+ "epoch": 0.7483914663054521,
1560
+ "grad_norm": 1.15625,
1561
+ "learning_rate": 3.610461428641402e-06,
1562
+ "loss": 2.1155,
1563
+ "step": 1105
1564
+ },
1565
+ {
1566
+ "epoch": 0.7517778530308161,
1567
+ "grad_norm": 1.171875,
1568
+ "learning_rate": 3.5199220168783564e-06,
1569
+ "loss": 2.1145,
1570
+ "step": 1110
1571
+ },
1572
+ {
1573
+ "epoch": 0.7551642397561802,
1574
+ "grad_norm": 1.2109375,
1575
+ "learning_rate": 3.4302892114992814e-06,
1576
+ "loss": 2.1419,
1577
+ "step": 1115
1578
+ },
1579
+ {
1580
+ "epoch": 0.7585506264815441,
1581
+ "grad_norm": 1.1953125,
1582
+ "learning_rate": 3.341575552734978e-06,
1583
+ "loss": 2.1448,
1584
+ "step": 1120
1585
+ },
1586
+ {
1587
+ "epoch": 0.7619370132069082,
1588
+ "grad_norm": 1.1953125,
1589
+ "learning_rate": 3.253793452221485e-06,
1590
+ "loss": 2.1375,
1591
+ "step": 1125
1592
+ },
1593
+ {
1594
+ "epoch": 0.7653233999322723,
1595
+ "grad_norm": 1.203125,
1596
+ "learning_rate": 3.1669551912635874e-06,
1597
+ "loss": 2.1626,
1598
+ "step": 1130
1599
+ },
1600
+ {
1601
+ "epoch": 0.7687097866576363,
1602
+ "grad_norm": 1.140625,
1603
+ "learning_rate": 3.0810729191165722e-06,
1604
+ "loss": 2.1256,
1605
+ "step": 1135
1606
+ },
1607
+ {
1608
+ "epoch": 0.7720961733830003,
1609
+ "grad_norm": 1.171875,
1610
+ "learning_rate": 2.9961586512864947e-06,
1611
+ "loss": 2.1337,
1612
+ "step": 1140
1613
+ },
1614
+ {
1615
+ "epoch": 0.7754825601083644,
1616
+ "grad_norm": 1.2109375,
1617
+ "learning_rate": 2.9122242678490975e-06,
1618
+ "loss": 2.1191,
1619
+ "step": 1145
1620
+ },
1621
+ {
1622
+ "epoch": 0.7788689468337284,
1623
+ "grad_norm": 1.203125,
1624
+ "learning_rate": 2.829281511787739e-06,
1625
+ "loss": 2.1523,
1626
+ "step": 1150
1627
+ },
1628
+ {
1629
+ "epoch": 0.7822553335590925,
1630
+ "grad_norm": 1.171875,
1631
+ "learning_rate": 2.747341987350439e-06,
1632
+ "loss": 2.1494,
1633
+ "step": 1155
1634
+ },
1635
+ {
1636
+ "epoch": 0.7856417202844564,
1637
+ "grad_norm": 1.1640625,
1638
+ "learning_rate": 2.666417158426393e-06,
1639
+ "loss": 2.1255,
1640
+ "step": 1160
1641
+ },
1642
+ {
1643
+ "epoch": 0.7890281070098205,
1644
+ "grad_norm": 1.1953125,
1645
+ "learning_rate": 2.586518346942082e-06,
1646
+ "loss": 2.136,
1647
+ "step": 1165
1648
+ },
1649
+ {
1650
+ "epoch": 0.7924144937351846,
1651
+ "grad_norm": 1.2109375,
1652
+ "learning_rate": 2.5076567312772635e-06,
1653
+ "loss": 2.1349,
1654
+ "step": 1170
1655
+ },
1656
+ {
1657
+ "epoch": 0.7958008804605486,
1658
+ "grad_norm": 1.203125,
1659
+ "learning_rate": 2.4298433447010495e-06,
1660
+ "loss": 2.1331,
1661
+ "step": 1175
1662
+ },
1663
+ {
1664
+ "epoch": 0.7991872671859126,
1665
+ "grad_norm": 1.1953125,
1666
+ "learning_rate": 2.353089073828255e-06,
1667
+ "loss": 2.1608,
1668
+ "step": 1180
1669
+ },
1670
+ {
1671
+ "epoch": 0.8025736539112767,
1672
+ "grad_norm": 1.1640625,
1673
+ "learning_rate": 2.2774046570963147e-06,
1674
+ "loss": 2.1455,
1675
+ "step": 1185
1676
+ },
1677
+ {
1678
+ "epoch": 0.8059600406366407,
1679
+ "grad_norm": 1.1953125,
1680
+ "learning_rate": 2.202800683262888e-06,
1681
+ "loss": 2.1323,
1682
+ "step": 1190
1683
+ },
1684
+ {
1685
+ "epoch": 0.8093464273620048,
1686
+ "grad_norm": 1.203125,
1687
+ "learning_rate": 2.1292875899244237e-06,
1688
+ "loss": 2.1235,
1689
+ "step": 1195
1690
+ },
1691
+ {
1692
+ "epoch": 0.8127328140873687,
1693
+ "grad_norm": 1.15625,
1694
+ "learning_rate": 2.056875662055874e-06,
1695
+ "loss": 2.1479,
1696
+ "step": 1200
1697
+ },
1698
+ {
1699
+ "epoch": 0.8161192008127328,
1700
+ "grad_norm": 1.1484375,
1701
+ "learning_rate": 1.985575030571746e-06,
1702
+ "loss": 2.1424,
1703
+ "step": 1205
1704
+ },
1705
+ {
1706
+ "epoch": 0.8195055875380969,
1707
+ "grad_norm": 1.171875,
1708
+ "learning_rate": 1.915395670908734e-06,
1709
+ "loss": 2.147,
1710
+ "step": 1210
1711
+ },
1712
+ {
1713
+ "epoch": 0.8228919742634608,
1714
+ "grad_norm": 1.1875,
1715
+ "learning_rate": 1.8463474016300831e-06,
1716
+ "loss": 2.1461,
1717
+ "step": 1215
1718
+ },
1719
+ {
1720
+ "epoch": 0.8262783609888249,
1721
+ "grad_norm": 1.140625,
1722
+ "learning_rate": 1.7784398830519002e-06,
1723
+ "loss": 2.1333,
1724
+ "step": 1220
1725
+ },
1726
+ {
1727
+ "epoch": 0.829664747714189,
1728
+ "grad_norm": 1.171875,
1729
+ "learning_rate": 1.7116826158916212e-06,
1730
+ "loss": 2.1438,
1731
+ "step": 1225
1732
+ },
1733
+ {
1734
+ "epoch": 0.833051134439553,
1735
+ "grad_norm": 1.1640625,
1736
+ "learning_rate": 1.6460849399387845e-06,
1737
+ "loss": 2.1339,
1738
+ "step": 1230
1739
+ },
1740
+ {
1741
+ "epoch": 0.836437521164917,
1742
+ "grad_norm": 1.1640625,
1743
+ "learning_rate": 1.5816560327483466e-06,
1744
+ "loss": 2.1507,
1745
+ "step": 1235
1746
+ },
1747
+ {
1748
+ "epoch": 0.839823907890281,
1749
+ "grad_norm": 1.171875,
1750
+ "learning_rate": 1.5184049083566688e-06,
1751
+ "loss": 2.132,
1752
+ "step": 1240
1753
+ },
1754
+ {
1755
+ "epoch": 0.8432102946156451,
1756
+ "grad_norm": 1.15625,
1757
+ "learning_rate": 1.4563404160204008e-06,
1758
+ "loss": 2.1273,
1759
+ "step": 1245
1760
+ },
1761
+ {
1762
+ "epoch": 0.8465966813410092,
1763
+ "grad_norm": 1.15625,
1764
+ "learning_rate": 1.3954712389783998e-06,
1765
+ "loss": 2.1127,
1766
+ "step": 1250
1767
+ },
1768
+ {
1769
+ "epoch": 0.8499830680663731,
1770
+ "grad_norm": 1.1796875,
1771
+ "learning_rate": 1.3358058932369012e-06,
1772
+ "loss": 2.1482,
1773
+ "step": 1255
1774
+ },
1775
+ {
1776
+ "epoch": 0.8533694547917372,
1777
+ "grad_norm": 1.2109375,
1778
+ "learning_rate": 1.2773527263780626e-06,
1779
+ "loss": 2.1329,
1780
+ "step": 1260
1781
+ },
1782
+ {
1783
+ "epoch": 0.8567558415171013,
1784
+ "grad_norm": 1.15625,
1785
+ "learning_rate": 1.2201199163920829e-06,
1786
+ "loss": 2.1414,
1787
+ "step": 1265
1788
+ },
1789
+ {
1790
+ "epoch": 0.8601422282424653,
1791
+ "grad_norm": 1.203125,
1792
+ "learning_rate": 1.1641154705330503e-06,
1793
+ "loss": 2.141,
1794
+ "step": 1270
1795
+ },
1796
+ {
1797
+ "epoch": 0.8635286149678293,
1798
+ "grad_norm": 1.15625,
1799
+ "learning_rate": 1.1093472241986636e-06,
1800
+ "loss": 2.1145,
1801
+ "step": 1275
1802
+ },
1803
+ {
1804
+ "epoch": 0.8669150016931934,
1805
+ "grad_norm": 1.1875,
1806
+ "learning_rate": 1.0558228398340188e-06,
1807
+ "loss": 2.1221,
1808
+ "step": 1280
1809
+ },
1810
+ {
1811
+ "epoch": 0.8703013884185574,
1812
+ "grad_norm": 1.1640625,
1813
+ "learning_rate": 1.0035498058595838e-06,
1814
+ "loss": 2.1412,
1815
+ "step": 1285
1816
+ },
1817
+ {
1818
+ "epoch": 0.8736877751439215,
1819
+ "grad_norm": 1.1875,
1820
+ "learning_rate": 9.525354356235006e-07,
1821
+ "loss": 2.1334,
1822
+ "step": 1290
1823
+ },
1824
+ {
1825
+ "epoch": 0.8770741618692854,
1826
+ "grad_norm": 1.1875,
1827
+ "learning_rate": 9.027868663784189e-07,
1828
+ "loss": 2.1355,
1829
+ "step": 1295
1830
+ },
1831
+ {
1832
+ "epoch": 0.8804605485946495,
1833
+ "grad_norm": 1.15625,
1834
+ "learning_rate": 8.543110582829272e-07,
1835
+ "loss": 2.1282,
1836
+ "step": 1300
1837
+ },
1838
+ {
1839
+ "epoch": 0.8838469353200136,
1840
+ "grad_norm": 1.1875,
1841
+ "learning_rate": 8.071147934278045e-07,
1842
+ "loss": 2.138,
1843
+ "step": 1305
1844
+ },
1845
+ {
1846
+ "epoch": 0.8872333220453776,
1847
+ "grad_norm": 1.171875,
1848
+ "learning_rate": 7.612046748871327e-07,
1849
+ "loss": 2.1461,
1850
+ "step": 1310
1851
+ },
1852
+ {
1853
+ "epoch": 0.8906197087707416,
1854
+ "grad_norm": 1.1953125,
1855
+ "learning_rate": 7.165871257945046e-07,
1856
+ "loss": 2.1359,
1857
+ "step": 1315
1858
+ },
1859
+ {
1860
+ "epoch": 0.8940060954961057,
1861
+ "grad_norm": 1.15625,
1862
+ "learning_rate": 6.732683884443736e-07,
1863
+ "loss": 2.1379,
1864
+ "step": 1320
1865
+ },
1866
+ {
1867
+ "epoch": 0.8973924822214697,
1868
+ "grad_norm": 1.125,
1869
+ "learning_rate": 6.312545234187106e-07,
1870
+ "loss": 2.1351,
1871
+ "step": 1325
1872
+ },
1873
+ {
1874
+ "epoch": 0.9007788689468337,
1875
+ "grad_norm": 1.1875,
1876
+ "learning_rate": 5.905514087391051e-07,
1877
+ "loss": 2.15,
1878
+ "step": 1330
1879
+ },
1880
+ {
1881
+ "epoch": 0.9041652556721977,
1882
+ "grad_norm": 1.140625,
1883
+ "learning_rate": 5.511647390443697e-07,
1884
+ "loss": 2.1279,
1885
+ "step": 1335
1886
+ },
1887
+ {
1888
+ "epoch": 0.9075516423975618,
1889
+ "grad_norm": 1.203125,
1890
+ "learning_rate": 5.131000247938367e-07,
1891
+ "loss": 2.1455,
1892
+ "step": 1340
1893
+ },
1894
+ {
1895
+ "epoch": 0.9109380291229259,
1896
+ "grad_norm": 1.1640625,
1897
+ "learning_rate": 4.7636259149639983e-07,
1898
+ "loss": 2.1343,
1899
+ "step": 1345
1900
+ },
1901
+ {
1902
+ "epoch": 0.9143244158482898,
1903
+ "grad_norm": 1.171875,
1904
+ "learning_rate": 4.4095757896544747e-07,
1905
+ "loss": 2.1377,
1906
+ "step": 1350
1907
+ },
1908
+ {
1909
+ "epoch": 0.9177108025736539,
1910
+ "grad_norm": 1.15625,
1911
+ "learning_rate": 4.0688994059976175e-07,
1912
+ "loss": 2.1354,
1913
+ "step": 1355
1914
+ },
1915
+ {
1916
+ "epoch": 0.921097189299018,
1917
+ "grad_norm": 1.1796875,
1918
+ "learning_rate": 3.7416444269050335e-07,
1919
+ "loss": 2.1452,
1920
+ "step": 1360
1921
+ },
1922
+ {
1923
+ "epoch": 0.924483576024382,
1924
+ "grad_norm": 1.1875,
1925
+ "learning_rate": 3.427856637543825e-07,
1926
+ "loss": 2.1494,
1927
+ "step": 1365
1928
+ },
1929
+ {
1930
+ "epoch": 0.927869962749746,
1931
+ "grad_norm": 1.171875,
1932
+ "learning_rate": 3.127579938930891e-07,
1933
+ "loss": 2.1434,
1934
+ "step": 1370
1935
+ },
1936
+ {
1937
+ "epoch": 0.93125634947511,
1938
+ "grad_norm": 1.1875,
1939
+ "learning_rate": 2.8408563417908875e-07,
1940
+ "loss": 2.1375,
1941
+ "step": 1375
1942
+ },
1943
+ {
1944
+ "epoch": 0.9346427362004741,
1945
+ "grad_norm": 1.171875,
1946
+ "learning_rate": 2.5677259606786686e-07,
1947
+ "loss": 2.1249,
1948
+ "step": 1380
1949
+ },
1950
+ {
1951
+ "epoch": 0.9380291229258382,
1952
+ "grad_norm": 1.203125,
1953
+ "learning_rate": 2.3082270083669255e-07,
1954
+ "loss": 2.1593,
1955
+ "step": 1385
1956
+ },
1957
+ {
1958
+ "epoch": 0.9414155096512021,
1959
+ "grad_norm": 1.1875,
1960
+ "learning_rate": 2.0623957905000603e-07,
1961
+ "loss": 2.1368,
1962
+ "step": 1390
1963
+ },
1964
+ {
1965
+ "epoch": 0.9448018963765662,
1966
+ "grad_norm": 1.140625,
1967
+ "learning_rate": 1.8302667005146936e-07,
1968
+ "loss": 2.1378,
1969
+ "step": 1395
1970
+ },
1971
+ {
1972
+ "epoch": 0.9481882831019303,
1973
+ "grad_norm": 1.21875,
1974
+ "learning_rate": 1.6118722148278586e-07,
1975
+ "loss": 2.1432,
1976
+ "step": 1400
1977
+ },
1978
+ {
1979
+ "epoch": 0.9515746698272943,
1980
+ "grad_norm": 1.1875,
1981
+ "learning_rate": 1.40724288829327e-07,
1982
+ "loss": 2.1343,
1983
+ "step": 1405
1984
+ },
1985
+ {
1986
+ "epoch": 0.9549610565526583,
1987
+ "grad_norm": 1.1640625,
1988
+ "learning_rate": 1.2164073499265404e-07,
1989
+ "loss": 2.1412,
1990
+ "step": 1410
1991
+ },
1992
+ {
1993
+ "epoch": 0.9583474432780223,
1994
+ "grad_norm": 1.171875,
1995
+ "learning_rate": 1.0393922988997995e-07,
1996
+ "loss": 2.1336,
1997
+ "step": 1415
1998
+ },
1999
+ {
2000
+ "epoch": 0.9617338300033864,
2001
+ "grad_norm": 1.171875,
2002
+ "learning_rate": 8.762225008062675e-08,
2003
+ "loss": 2.1514,
2004
+ "step": 1420
2005
+ },
2006
+ {
2007
+ "epoch": 0.9651202167287504,
2008
+ "grad_norm": 1.15625,
2009
+ "learning_rate": 7.269207841953618e-08,
2010
+ "loss": 2.121,
2011
+ "step": 1425
2012
+ },
2013
+ {
2014
+ "epoch": 0.9685066034541144,
2015
+ "grad_norm": 1.1640625,
2016
+ "learning_rate": 5.915080373788962e-08,
2017
+ "loss": 2.1249,
2018
+ "step": 1430
2019
+ },
2020
+ {
2021
+ "epoch": 0.9718929901794785,
2022
+ "grad_norm": 1.3125,
2023
+ "learning_rate": 4.700032055086179e-08,
2024
+ "loss": 2.1442,
2025
+ "step": 1435
2026
+ },
2027
+ {
2028
+ "epoch": 0.9752793769048426,
2029
+ "grad_norm": 1.1796875,
2030
+ "learning_rate": 3.6242328792567286e-08,
2031
+ "loss": 2.1205,
2032
+ "step": 1440
2033
+ },
2034
+ {
2035
+ "epoch": 0.9786657636302065,
2036
+ "grad_norm": 1.15625,
2037
+ "learning_rate": 2.6878333578227446e-08,
2038
+ "loss": 2.1349,
2039
+ "step": 1445
2040
+ },
2041
+ {
2042
+ "epoch": 0.9820521503555706,
2043
+ "grad_norm": 1.1640625,
2044
+ "learning_rate": 1.890964499359327e-08,
2045
+ "loss": 2.1394,
2046
+ "step": 1450
2047
+ },
2048
+ {
2049
+ "epoch": 0.9854385370809347,
2050
+ "grad_norm": 1.171875,
2051
+ "learning_rate": 1.2337377911660897e-08,
2052
+ "loss": 2.1281,
2053
+ "step": 1455
2054
+ },
2055
+ {
2056
+ "epoch": 0.9888249238062987,
2057
+ "grad_norm": 1.171875,
2058
+ "learning_rate": 7.162451836685291e-09,
2059
+ "loss": 2.1459,
2060
+ "step": 1460
2061
+ },
2062
+ {
2063
+ "epoch": 0.9922113105316627,
2064
+ "grad_norm": 1.1484375,
2065
+ "learning_rate": 3.385590775544234e-09,
2066
+ "loss": 2.1328,
2067
+ "step": 1465
2068
+ },
2069
+ {
2070
+ "epoch": 0.9955976972570267,
2071
+ "grad_norm": 1.1875,
2072
+ "learning_rate": 1.007323136438254e-09,
2073
+ "loss": 2.1426,
2074
+ "step": 1470
2075
+ },
2076
+ {
2077
+ "epoch": 0.9989840839823908,
2078
+ "grad_norm": 1.1484375,
2079
+ "learning_rate": 2.7981654966424686e-11,
2080
+ "loss": 2.1307,
2081
+ "step": 1475
2082
+ },
2083
+ {
2084
+ "epoch": 0.9996613613274636,
2085
+ "eval_loss": 2.1349875926971436,
2086
+ "eval_runtime": 82.2944,
2087
+ "eval_samples_per_second": 15.858,
2088
+ "eval_steps_per_second": 1.993,
2089
+ "step": 1476
2090
+ },
2091
+ {
2092
+ "epoch": 0.9996613613274636,
2093
+ "step": 1476,
2094
+ "total_flos": 3.002086743311647e+17,
2095
+ "train_loss": 2.154247896135015,
2096
+ "train_runtime": 6599.6501,
2097
+ "train_samples_per_second": 3.579,
2098
+ "train_steps_per_second": 0.224
2099
+ }
2100
+ ],
2101
+ "logging_steps": 5,
2102
+ "max_steps": 1476,
2103
+ "num_input_tokens_seen": 0,
2104
+ "num_train_epochs": 1,
2105
+ "save_steps": 100,
2106
+ "stateful_callbacks": {
2107
+ "TrainerControl": {
2108
+ "args": {
2109
+ "should_epoch_stop": false,
2110
+ "should_evaluate": false,
2111
+ "should_log": false,
2112
+ "should_save": true,
2113
+ "should_training_stop": false
2114
+ },
2115
+ "attributes": {}
2116
+ }
2117
+ },
2118
+ "total_flos": 3.002086743311647e+17,
2119
+ "train_batch_size": 8,
2120
+ "trial_name": null,
2121
+ "trial_params": null
2122
+ }