Commit
•
740413a
1
Parent(s):
2a30159
Model save
Browse files- README.md +69 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +3300 -0
README.md
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
|
| 3 |
+
library_name: peft
|
| 4 |
+
license: llama3.1
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- sft
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: induction-20k-50seeds-barc-llama3.1-8b-instruct-lora64-induction-gpt4omini35k_lr2e-4_epoch3
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# induction-20k-50seeds-barc-llama3.1-8b-instruct-lora64-induction-gpt4omini35k_lr2e-4_epoch3
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) on an unknown dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.3344
|
| 22 |
+
|
| 23 |
+
## Model description
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Intended uses & limitations
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training and evaluation data
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Training procedure
|
| 36 |
+
|
| 37 |
+
### Training hyperparameters
|
| 38 |
+
|
| 39 |
+
The following hyperparameters were used during training:
|
| 40 |
+
- learning_rate: 0.0002
|
| 41 |
+
- train_batch_size: 8
|
| 42 |
+
- eval_batch_size: 4
|
| 43 |
+
- seed: 42
|
| 44 |
+
- distributed_type: multi-GPU
|
| 45 |
+
- num_devices: 8
|
| 46 |
+
- gradient_accumulation_steps: 2
|
| 47 |
+
- total_train_batch_size: 128
|
| 48 |
+
- total_eval_batch_size: 32
|
| 49 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 50 |
+
- lr_scheduler_type: cosine
|
| 51 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 52 |
+
- num_epochs: 3
|
| 53 |
+
|
| 54 |
+
### Training results
|
| 55 |
+
|
| 56 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 57 |
+
|:-------------:|:------:|:----:|:---------------:|
|
| 58 |
+
| 0.3653 | 0.9968 | 154 | 0.3545 |
|
| 59 |
+
| 0.3166 | 2.0 | 309 | 0.3375 |
|
| 60 |
+
| 0.2951 | 2.9903 | 462 | 0.3344 |
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
### Framework versions
|
| 64 |
+
|
| 65 |
+
- PEFT 0.13.0
|
| 66 |
+
- Transformers 4.45.0.dev0
|
| 67 |
+
- Pytorch 2.4.0+cu121
|
| 68 |
+
- Datasets 3.0.1
|
| 69 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.9902912621359223,
|
| 3 |
+
"total_flos": 1.8679451641472614e+17,
|
| 4 |
+
"train_loss": 0.36376328850205325,
|
| 5 |
+
"train_runtime": 5739.3163,
|
| 6 |
+
"train_samples": 19726,
|
| 7 |
+
"train_samples_per_second": 10.311,
|
| 8 |
+
"train_steps_per_second": 0.08
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.9902912621359223,
|
| 3 |
+
"total_flos": 1.8679451641472614e+17,
|
| 4 |
+
"train_loss": 0.36376328850205325,
|
| 5 |
+
"train_runtime": 5739.3163,
|
| 6 |
+
"train_samples": 19726,
|
| 7 |
+
"train_samples_per_second": 10.311,
|
| 8 |
+
"train_steps_per_second": 0.08
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3300 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.9902912621359223,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 462,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.006472491909385114,
|
| 13 |
+
"grad_norm": 0.8681851528027766,
|
| 14 |
+
"learning_rate": 4.255319148936171e-06,
|
| 15 |
+
"loss": 1.2864,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.012944983818770227,
|
| 20 |
+
"grad_norm": 0.8603766114064247,
|
| 21 |
+
"learning_rate": 8.510638297872341e-06,
|
| 22 |
+
"loss": 1.3029,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.019417475728155338,
|
| 27 |
+
"grad_norm": 0.8460211658943029,
|
| 28 |
+
"learning_rate": 1.2765957446808511e-05,
|
| 29 |
+
"loss": 1.2676,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.025889967637540454,
|
| 34 |
+
"grad_norm": 0.823879412927583,
|
| 35 |
+
"learning_rate": 1.7021276595744682e-05,
|
| 36 |
+
"loss": 1.2794,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.032362459546925564,
|
| 41 |
+
"grad_norm": 0.7999755689798974,
|
| 42 |
+
"learning_rate": 2.1276595744680852e-05,
|
| 43 |
+
"loss": 1.2632,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.038834951456310676,
|
| 48 |
+
"grad_norm": 0.779503875119536,
|
| 49 |
+
"learning_rate": 2.5531914893617022e-05,
|
| 50 |
+
"loss": 1.2356,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.045307443365695796,
|
| 55 |
+
"grad_norm": 0.7113356719326077,
|
| 56 |
+
"learning_rate": 2.9787234042553192e-05,
|
| 57 |
+
"loss": 1.2075,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.05177993527508091,
|
| 62 |
+
"grad_norm": 0.5733112055032882,
|
| 63 |
+
"learning_rate": 3.4042553191489365e-05,
|
| 64 |
+
"loss": 1.1078,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.05825242718446602,
|
| 69 |
+
"grad_norm": 0.4686147590427965,
|
| 70 |
+
"learning_rate": 3.829787234042553e-05,
|
| 71 |
+
"loss": 1.0108,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.06472491909385113,
|
| 76 |
+
"grad_norm": 0.4960364471309283,
|
| 77 |
+
"learning_rate": 4.2553191489361704e-05,
|
| 78 |
+
"loss": 0.9991,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.07119741100323625,
|
| 83 |
+
"grad_norm": 0.5462124995641008,
|
| 84 |
+
"learning_rate": 4.680851063829788e-05,
|
| 85 |
+
"loss": 0.9427,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.07766990291262135,
|
| 90 |
+
"grad_norm": 0.546681134551426,
|
| 91 |
+
"learning_rate": 5.1063829787234044e-05,
|
| 92 |
+
"loss": 0.8893,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.08414239482200647,
|
| 97 |
+
"grad_norm": 0.556093704754962,
|
| 98 |
+
"learning_rate": 5.531914893617022e-05,
|
| 99 |
+
"loss": 0.8444,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.09061488673139159,
|
| 104 |
+
"grad_norm": 0.4982225051601507,
|
| 105 |
+
"learning_rate": 5.9574468085106384e-05,
|
| 106 |
+
"loss": 0.7642,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.0970873786407767,
|
| 111 |
+
"grad_norm": 0.48253513519932356,
|
| 112 |
+
"learning_rate": 6.382978723404256e-05,
|
| 113 |
+
"loss": 0.7228,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.10355987055016182,
|
| 118 |
+
"grad_norm": 0.44162352908966596,
|
| 119 |
+
"learning_rate": 6.808510638297873e-05,
|
| 120 |
+
"loss": 0.6525,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.11003236245954692,
|
| 125 |
+
"grad_norm": 0.3918381633741909,
|
| 126 |
+
"learning_rate": 7.23404255319149e-05,
|
| 127 |
+
"loss": 0.604,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.11650485436893204,
|
| 132 |
+
"grad_norm": 0.2648043458799554,
|
| 133 |
+
"learning_rate": 7.659574468085106e-05,
|
| 134 |
+
"loss": 0.5693,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.12297734627831715,
|
| 139 |
+
"grad_norm": 0.243819913231238,
|
| 140 |
+
"learning_rate": 8.085106382978723e-05,
|
| 141 |
+
"loss": 0.5205,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.12944983818770225,
|
| 146 |
+
"grad_norm": 0.23168627433032857,
|
| 147 |
+
"learning_rate": 8.510638297872341e-05,
|
| 148 |
+
"loss": 0.5296,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.13592233009708737,
|
| 153 |
+
"grad_norm": 0.1994751857050266,
|
| 154 |
+
"learning_rate": 8.936170212765958e-05,
|
| 155 |
+
"loss": 0.5269,
|
| 156 |
+
"step": 21
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.1423948220064725,
|
| 160 |
+
"grad_norm": 0.2530061947493179,
|
| 161 |
+
"learning_rate": 9.361702127659576e-05,
|
| 162 |
+
"loss": 0.5091,
|
| 163 |
+
"step": 22
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.1488673139158576,
|
| 167 |
+
"grad_norm": 0.24049496280327698,
|
| 168 |
+
"learning_rate": 9.787234042553192e-05,
|
| 169 |
+
"loss": 0.5169,
|
| 170 |
+
"step": 23
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.1553398058252427,
|
| 174 |
+
"grad_norm": 0.22948907942121893,
|
| 175 |
+
"learning_rate": 0.00010212765957446809,
|
| 176 |
+
"loss": 0.4954,
|
| 177 |
+
"step": 24
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.16181229773462782,
|
| 181 |
+
"grad_norm": 0.17253285827605366,
|
| 182 |
+
"learning_rate": 0.00010638297872340425,
|
| 183 |
+
"loss": 0.4678,
|
| 184 |
+
"step": 25
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.16828478964401294,
|
| 188 |
+
"grad_norm": 0.1938226429644451,
|
| 189 |
+
"learning_rate": 0.00011063829787234043,
|
| 190 |
+
"loss": 0.4857,
|
| 191 |
+
"step": 26
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.17475728155339806,
|
| 195 |
+
"grad_norm": 0.13721502402963923,
|
| 196 |
+
"learning_rate": 0.00011489361702127661,
|
| 197 |
+
"loss": 0.466,
|
| 198 |
+
"step": 27
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.18122977346278318,
|
| 202 |
+
"grad_norm": 0.12808136935864992,
|
| 203 |
+
"learning_rate": 0.00011914893617021277,
|
| 204 |
+
"loss": 0.4652,
|
| 205 |
+
"step": 28
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.18770226537216828,
|
| 209 |
+
"grad_norm": 0.12041781390556437,
|
| 210 |
+
"learning_rate": 0.00012340425531914893,
|
| 211 |
+
"loss": 0.4527,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.1941747572815534,
|
| 216 |
+
"grad_norm": 0.11745727258563804,
|
| 217 |
+
"learning_rate": 0.00012765957446808513,
|
| 218 |
+
"loss": 0.4392,
|
| 219 |
+
"step": 30
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.20064724919093851,
|
| 223 |
+
"grad_norm": 0.11168371739950012,
|
| 224 |
+
"learning_rate": 0.00013191489361702127,
|
| 225 |
+
"loss": 0.4578,
|
| 226 |
+
"step": 31
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.20711974110032363,
|
| 230 |
+
"grad_norm": 0.11295507271229074,
|
| 231 |
+
"learning_rate": 0.00013617021276595746,
|
| 232 |
+
"loss": 0.4384,
|
| 233 |
+
"step": 32
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.21359223300970873,
|
| 237 |
+
"grad_norm": 0.11084156792638247,
|
| 238 |
+
"learning_rate": 0.00014042553191489363,
|
| 239 |
+
"loss": 0.4292,
|
| 240 |
+
"step": 33
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.22006472491909385,
|
| 244 |
+
"grad_norm": 0.10746985136176236,
|
| 245 |
+
"learning_rate": 0.0001446808510638298,
|
| 246 |
+
"loss": 0.4202,
|
| 247 |
+
"step": 34
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.22653721682847897,
|
| 251 |
+
"grad_norm": 0.10723808557482935,
|
| 252 |
+
"learning_rate": 0.00014893617021276596,
|
| 253 |
+
"loss": 0.4345,
|
| 254 |
+
"step": 35
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.23300970873786409,
|
| 258 |
+
"grad_norm": 0.10308761677232968,
|
| 259 |
+
"learning_rate": 0.00015319148936170213,
|
| 260 |
+
"loss": 0.4318,
|
| 261 |
+
"step": 36
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.23948220064724918,
|
| 265 |
+
"grad_norm": 0.1044215437881733,
|
| 266 |
+
"learning_rate": 0.00015744680851063832,
|
| 267 |
+
"loss": 0.4289,
|
| 268 |
+
"step": 37
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.2459546925566343,
|
| 272 |
+
"grad_norm": 0.10783517564137016,
|
| 273 |
+
"learning_rate": 0.00016170212765957446,
|
| 274 |
+
"loss": 0.4226,
|
| 275 |
+
"step": 38
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.2524271844660194,
|
| 279 |
+
"grad_norm": 0.11058134845234333,
|
| 280 |
+
"learning_rate": 0.00016595744680851065,
|
| 281 |
+
"loss": 0.4392,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.2588996763754045,
|
| 286 |
+
"grad_norm": 0.10721298987840806,
|
| 287 |
+
"learning_rate": 0.00017021276595744682,
|
| 288 |
+
"loss": 0.4097,
|
| 289 |
+
"step": 40
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.26537216828478966,
|
| 293 |
+
"grad_norm": 0.10203447619356035,
|
| 294 |
+
"learning_rate": 0.00017446808510638298,
|
| 295 |
+
"loss": 0.4371,
|
| 296 |
+
"step": 41
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.27184466019417475,
|
| 300 |
+
"grad_norm": 0.09611299924461707,
|
| 301 |
+
"learning_rate": 0.00017872340425531915,
|
| 302 |
+
"loss": 0.4264,
|
| 303 |
+
"step": 42
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.2783171521035599,
|
| 307 |
+
"grad_norm": 0.09911922773823158,
|
| 308 |
+
"learning_rate": 0.00018297872340425532,
|
| 309 |
+
"loss": 0.4177,
|
| 310 |
+
"step": 43
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.284789644012945,
|
| 314 |
+
"grad_norm": 0.10195593280531762,
|
| 315 |
+
"learning_rate": 0.0001872340425531915,
|
| 316 |
+
"loss": 0.4322,
|
| 317 |
+
"step": 44
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.2912621359223301,
|
| 321 |
+
"grad_norm": 0.0990363778451009,
|
| 322 |
+
"learning_rate": 0.00019148936170212768,
|
| 323 |
+
"loss": 0.4272,
|
| 324 |
+
"step": 45
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.2977346278317152,
|
| 328 |
+
"grad_norm": 0.10312296274780797,
|
| 329 |
+
"learning_rate": 0.00019574468085106384,
|
| 330 |
+
"loss": 0.3997,
|
| 331 |
+
"step": 46
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.3042071197411003,
|
| 335 |
+
"grad_norm": 0.09688182042517872,
|
| 336 |
+
"learning_rate": 0.0002,
|
| 337 |
+
"loss": 0.3955,
|
| 338 |
+
"step": 47
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3106796116504854,
|
| 342 |
+
"grad_norm": 0.09540799755823594,
|
| 343 |
+
"learning_rate": 0.00019999713469087867,
|
| 344 |
+
"loss": 0.4078,
|
| 345 |
+
"step": 48
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.31715210355987056,
|
| 349 |
+
"grad_norm": 0.10596966214446712,
|
| 350 |
+
"learning_rate": 0.00019998853892771453,
|
| 351 |
+
"loss": 0.4102,
|
| 352 |
+
"step": 49
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.32362459546925565,
|
| 356 |
+
"grad_norm": 0.10086883515116485,
|
| 357 |
+
"learning_rate": 0.00019997421320309795,
|
| 358 |
+
"loss": 0.4048,
|
| 359 |
+
"step": 50
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.3300970873786408,
|
| 363 |
+
"grad_norm": 0.10400887250356652,
|
| 364 |
+
"learning_rate": 0.00019995415833798158,
|
| 365 |
+
"loss": 0.403,
|
| 366 |
+
"step": 51
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.3365695792880259,
|
| 370 |
+
"grad_norm": 0.09067444994385492,
|
| 371 |
+
"learning_rate": 0.00019992837548163316,
|
| 372 |
+
"loss": 0.3867,
|
| 373 |
+
"step": 52
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.343042071197411,
|
| 377 |
+
"grad_norm": 0.09109660691369531,
|
| 378 |
+
"learning_rate": 0.00019989686611156972,
|
| 379 |
+
"loss": 0.3956,
|
| 380 |
+
"step": 53
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.34951456310679613,
|
| 384 |
+
"grad_norm": 0.09162094937144005,
|
| 385 |
+
"learning_rate": 0.000199859632033473,
|
| 386 |
+
"loss": 0.407,
|
| 387 |
+
"step": 54
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.3559870550161812,
|
| 391 |
+
"grad_norm": 0.09041901984453071,
|
| 392 |
+
"learning_rate": 0.00019981667538108587,
|
| 393 |
+
"loss": 0.3913,
|
| 394 |
+
"step": 55
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.36245954692556637,
|
| 398 |
+
"grad_norm": 0.0948520973493128,
|
| 399 |
+
"learning_rate": 0.00019976799861609008,
|
| 400 |
+
"loss": 0.3959,
|
| 401 |
+
"step": 56
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.36893203883495146,
|
| 405 |
+
"grad_norm": 0.09239535228342881,
|
| 406 |
+
"learning_rate": 0.00019971360452796522,
|
| 407 |
+
"loss": 0.3976,
|
| 408 |
+
"step": 57
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.37540453074433655,
|
| 412 |
+
"grad_norm": 0.09430664951378093,
|
| 413 |
+
"learning_rate": 0.0001996534962338288,
|
| 414 |
+
"loss": 0.3984,
|
| 415 |
+
"step": 58
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.3818770226537217,
|
| 419 |
+
"grad_norm": 0.08478493523992439,
|
| 420 |
+
"learning_rate": 0.0001995876771782577,
|
| 421 |
+
"loss": 0.3862,
|
| 422 |
+
"step": 59
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.3883495145631068,
|
| 426 |
+
"grad_norm": 0.08736157516174174,
|
| 427 |
+
"learning_rate": 0.00019951615113309075,
|
| 428 |
+
"loss": 0.3768,
|
| 429 |
+
"step": 60
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.3948220064724919,
|
| 433 |
+
"grad_norm": 0.08833359616935596,
|
| 434 |
+
"learning_rate": 0.00019943892219721253,
|
| 435 |
+
"loss": 0.3832,
|
| 436 |
+
"step": 61
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.40129449838187703,
|
| 440 |
+
"grad_norm": 0.09634861288945641,
|
| 441 |
+
"learning_rate": 0.0001993559947963185,
|
| 442 |
+
"loss": 0.3975,
|
| 443 |
+
"step": 62
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.4077669902912621,
|
| 447 |
+
"grad_norm": 0.08474514858686731,
|
| 448 |
+
"learning_rate": 0.00019926737368266144,
|
| 449 |
+
"loss": 0.3803,
|
| 450 |
+
"step": 63
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.41423948220064727,
|
| 454 |
+
"grad_norm": 0.0878556667338793,
|
| 455 |
+
"learning_rate": 0.00019917306393477907,
|
| 456 |
+
"loss": 0.3874,
|
| 457 |
+
"step": 64
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.42071197411003236,
|
| 461 |
+
"grad_norm": 0.08906183743795007,
|
| 462 |
+
"learning_rate": 0.00019907307095720303,
|
| 463 |
+
"loss": 0.384,
|
| 464 |
+
"step": 65
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.42718446601941745,
|
| 468 |
+
"grad_norm": 0.08738087160835142,
|
| 469 |
+
"learning_rate": 0.00019896740048014908,
|
| 470 |
+
"loss": 0.3883,
|
| 471 |
+
"step": 66
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.4336569579288026,
|
| 475 |
+
"grad_norm": 0.09070551920354154,
|
| 476 |
+
"learning_rate": 0.00019885605855918885,
|
| 477 |
+
"loss": 0.38,
|
| 478 |
+
"step": 67
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.4401294498381877,
|
| 482 |
+
"grad_norm": 0.08922144963205411,
|
| 483 |
+
"learning_rate": 0.00019873905157490285,
|
| 484 |
+
"loss": 0.3854,
|
| 485 |
+
"step": 68
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.44660194174757284,
|
| 489 |
+
"grad_norm": 0.09304718374139792,
|
| 490 |
+
"learning_rate": 0.0001986163862325146,
|
| 491 |
+
"loss": 0.3804,
|
| 492 |
+
"step": 69
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.45307443365695793,
|
| 496 |
+
"grad_norm": 0.09290101551147209,
|
| 497 |
+
"learning_rate": 0.0001984880695615066,
|
| 498 |
+
"loss": 0.3876,
|
| 499 |
+
"step": 70
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.459546925566343,
|
| 503 |
+
"grad_norm": 0.08862476310706492,
|
| 504 |
+
"learning_rate": 0.0001983541089152174,
|
| 505 |
+
"loss": 0.3701,
|
| 506 |
+
"step": 71
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.46601941747572817,
|
| 510 |
+
"grad_norm": 0.0864643410147614,
|
| 511 |
+
"learning_rate": 0.00019821451197042026,
|
| 512 |
+
"loss": 0.3625,
|
| 513 |
+
"step": 72
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.47249190938511326,
|
| 517 |
+
"grad_norm": 0.0889683558231721,
|
| 518 |
+
"learning_rate": 0.0001980692867268832,
|
| 519 |
+
"loss": 0.3635,
|
| 520 |
+
"step": 73
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.47896440129449835,
|
| 524 |
+
"grad_norm": 0.090198299336284,
|
| 525 |
+
"learning_rate": 0.0001979184415069104,
|
| 526 |
+
"loss": 0.3804,
|
| 527 |
+
"step": 74
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.4854368932038835,
|
| 531 |
+
"grad_norm": 0.09214744360572043,
|
| 532 |
+
"learning_rate": 0.00019776198495486565,
|
| 533 |
+
"loss": 0.367,
|
| 534 |
+
"step": 75
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.4919093851132686,
|
| 538 |
+
"grad_norm": 0.08545016444696638,
|
| 539 |
+
"learning_rate": 0.00019759992603667667,
|
| 540 |
+
"loss": 0.3671,
|
| 541 |
+
"step": 76
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.49838187702265374,
|
| 545 |
+
"grad_norm": 0.08836432043627582,
|
| 546 |
+
"learning_rate": 0.00019743227403932134,
|
| 547 |
+
"loss": 0.3759,
|
| 548 |
+
"step": 77
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.5048543689320388,
|
| 552 |
+
"grad_norm": 0.09220525179921107,
|
| 553 |
+
"learning_rate": 0.00019725903857029564,
|
| 554 |
+
"loss": 0.3759,
|
| 555 |
+
"step": 78
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.511326860841424,
|
| 559 |
+
"grad_norm": 0.09021919927906268,
|
| 560 |
+
"learning_rate": 0.00019708022955706292,
|
| 561 |
+
"loss": 0.3748,
|
| 562 |
+
"step": 79
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.517799352750809,
|
| 566 |
+
"grad_norm": 0.0857417469759507,
|
| 567 |
+
"learning_rate": 0.00019689585724648516,
|
| 568 |
+
"loss": 0.3778,
|
| 569 |
+
"step": 80
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.5242718446601942,
|
| 573 |
+
"grad_norm": 0.08899327468436542,
|
| 574 |
+
"learning_rate": 0.00019670593220423558,
|
| 575 |
+
"loss": 0.3711,
|
| 576 |
+
"step": 81
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.5307443365695793,
|
| 580 |
+
"grad_norm": 0.09420631205003693,
|
| 581 |
+
"learning_rate": 0.00019651046531419332,
|
| 582 |
+
"loss": 0.3678,
|
| 583 |
+
"step": 82
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.5372168284789643,
|
| 587 |
+
"grad_norm": 0.09175386946113716,
|
| 588 |
+
"learning_rate": 0.00019630946777781966,
|
| 589 |
+
"loss": 0.3669,
|
| 590 |
+
"step": 83
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.5436893203883495,
|
| 594 |
+
"grad_norm": 0.08372986124983309,
|
| 595 |
+
"learning_rate": 0.0001961029511135161,
|
| 596 |
+
"loss": 0.3593,
|
| 597 |
+
"step": 84
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.5501618122977346,
|
| 601 |
+
"grad_norm": 0.08923405383972152,
|
| 602 |
+
"learning_rate": 0.00019589092715596417,
|
| 603 |
+
"loss": 0.3729,
|
| 604 |
+
"step": 85
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.5566343042071198,
|
| 608 |
+
"grad_norm": 0.08853783615052746,
|
| 609 |
+
"learning_rate": 0.00019567340805544758,
|
| 610 |
+
"loss": 0.3512,
|
| 611 |
+
"step": 86
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.5631067961165048,
|
| 615 |
+
"grad_norm": 0.09158487904428567,
|
| 616 |
+
"learning_rate": 0.0001954504062771555,
|
| 617 |
+
"loss": 0.3673,
|
| 618 |
+
"step": 87
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.56957928802589,
|
| 622 |
+
"grad_norm": 0.09709095338695828,
|
| 623 |
+
"learning_rate": 0.00019522193460046864,
|
| 624 |
+
"loss": 0.3758,
|
| 625 |
+
"step": 88
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.5760517799352751,
|
| 629 |
+
"grad_norm": 0.09252319787935548,
|
| 630 |
+
"learning_rate": 0.00019498800611822645,
|
| 631 |
+
"loss": 0.3588,
|
| 632 |
+
"step": 89
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.5825242718446602,
|
| 636 |
+
"grad_norm": 0.10229380875651542,
|
| 637 |
+
"learning_rate": 0.00019474863423597728,
|
| 638 |
+
"loss": 0.3859,
|
| 639 |
+
"step": 90
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.5889967637540453,
|
| 643 |
+
"grad_norm": 0.09361866653135192,
|
| 644 |
+
"learning_rate": 0.00019450383267120982,
|
| 645 |
+
"loss": 0.3739,
|
| 646 |
+
"step": 91
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.5954692556634305,
|
| 650 |
+
"grad_norm": 0.0919566522897771,
|
| 651 |
+
"learning_rate": 0.00019425361545256727,
|
| 652 |
+
"loss": 0.3648,
|
| 653 |
+
"step": 92
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.6019417475728155,
|
| 657 |
+
"grad_norm": 0.10007063382515365,
|
| 658 |
+
"learning_rate": 0.0001939979969190432,
|
| 659 |
+
"loss": 0.3628,
|
| 660 |
+
"step": 93
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.6084142394822006,
|
| 664 |
+
"grad_norm": 0.08642815308708984,
|
| 665 |
+
"learning_rate": 0.00019373699171915988,
|
| 666 |
+
"loss": 0.3607,
|
| 667 |
+
"step": 94
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.6148867313915858,
|
| 671 |
+
"grad_norm": 0.08898100138256641,
|
| 672 |
+
"learning_rate": 0.00019347061481012894,
|
| 673 |
+
"loss": 0.3649,
|
| 674 |
+
"step": 95
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.6213592233009708,
|
| 678 |
+
"grad_norm": 0.09929344608921754,
|
| 679 |
+
"learning_rate": 0.00019319888145699415,
|
| 680 |
+
"loss": 0.3707,
|
| 681 |
+
"step": 96
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.627831715210356,
|
| 685 |
+
"grad_norm": 0.08479565009317201,
|
| 686 |
+
"learning_rate": 0.00019292180723175654,
|
| 687 |
+
"loss": 0.3596,
|
| 688 |
+
"step": 97
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.6343042071197411,
|
| 692 |
+
"grad_norm": 0.08536934965094412,
|
| 693 |
+
"learning_rate": 0.00019263940801248226,
|
| 694 |
+
"loss": 0.3693,
|
| 695 |
+
"step": 98
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.6407766990291263,
|
| 699 |
+
"grad_norm": 0.09264353563725661,
|
| 700 |
+
"learning_rate": 0.0001923516999823925,
|
| 701 |
+
"loss": 0.3679,
|
| 702 |
+
"step": 99
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.6472491909385113,
|
| 706 |
+
"grad_norm": 0.09075116775864867,
|
| 707 |
+
"learning_rate": 0.00019205869962893605,
|
| 708 |
+
"loss": 0.3594,
|
| 709 |
+
"step": 100
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.6537216828478964,
|
| 713 |
+
"grad_norm": 0.08696915654679535,
|
| 714 |
+
"learning_rate": 0.0001917604237428447,
|
| 715 |
+
"loss": 0.3699,
|
| 716 |
+
"step": 101
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.6601941747572816,
|
| 720 |
+
"grad_norm": 0.08387717136900205,
|
| 721 |
+
"learning_rate": 0.00019145688941717075,
|
| 722 |
+
"loss": 0.3569,
|
| 723 |
+
"step": 102
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.6666666666666666,
|
| 727 |
+
"grad_norm": 0.09727169698788522,
|
| 728 |
+
"learning_rate": 0.00019114811404630762,
|
| 729 |
+
"loss": 0.3725,
|
| 730 |
+
"step": 103
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.6731391585760518,
|
| 734 |
+
"grad_norm": 0.08785848039003973,
|
| 735 |
+
"learning_rate": 0.0001908341153249931,
|
| 736 |
+
"loss": 0.3625,
|
| 737 |
+
"step": 104
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.6796116504854369,
|
| 741 |
+
"grad_norm": 0.09111196159253629,
|
| 742 |
+
"learning_rate": 0.00019051491124729512,
|
| 743 |
+
"loss": 0.3658,
|
| 744 |
+
"step": 105
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.686084142394822,
|
| 748 |
+
"grad_norm": 0.09173527522438048,
|
| 749 |
+
"learning_rate": 0.00019019052010558088,
|
| 750 |
+
"loss": 0.3605,
|
| 751 |
+
"step": 106
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.6925566343042071,
|
| 755 |
+
"grad_norm": 0.08349190333555147,
|
| 756 |
+
"learning_rate": 0.00018986096048946824,
|
| 757 |
+
"loss": 0.3714,
|
| 758 |
+
"step": 107
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.6990291262135923,
|
| 762 |
+
"grad_norm": 0.08795475348576975,
|
| 763 |
+
"learning_rate": 0.0001895262512847607,
|
| 764 |
+
"loss": 0.3646,
|
| 765 |
+
"step": 108
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.7055016181229773,
|
| 769 |
+
"grad_norm": 0.0834930063587362,
|
| 770 |
+
"learning_rate": 0.00018918641167236505,
|
| 771 |
+
"loss": 0.3518,
|
| 772 |
+
"step": 109
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.7119741100323624,
|
| 776 |
+
"grad_norm": 0.08299225873111288,
|
| 777 |
+
"learning_rate": 0.00018884146112719207,
|
| 778 |
+
"loss": 0.3498,
|
| 779 |
+
"step": 110
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.7184466019417476,
|
| 783 |
+
"grad_norm": 0.08676290151306544,
|
| 784 |
+
"learning_rate": 0.00018849141941704067,
|
| 785 |
+
"loss": 0.3612,
|
| 786 |
+
"step": 111
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.7249190938511327,
|
| 790 |
+
"grad_norm": 0.09107846226354271,
|
| 791 |
+
"learning_rate": 0.00018813630660146488,
|
| 792 |
+
"loss": 0.3508,
|
| 793 |
+
"step": 112
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.7313915857605178,
|
| 797 |
+
"grad_norm": 0.09040771142681107,
|
| 798 |
+
"learning_rate": 0.00018777614303062457,
|
| 799 |
+
"loss": 0.358,
|
| 800 |
+
"step": 113
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.7378640776699029,
|
| 804 |
+
"grad_norm": 0.08473984809675678,
|
| 805 |
+
"learning_rate": 0.000187410949344119,
|
| 806 |
+
"loss": 0.3567,
|
| 807 |
+
"step": 114
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.7443365695792881,
|
| 811 |
+
"grad_norm": 0.09166724031633931,
|
| 812 |
+
"learning_rate": 0.00018704074646980415,
|
| 813 |
+
"loss": 0.3474,
|
| 814 |
+
"step": 115
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.7508090614886731,
|
| 818 |
+
"grad_norm": 0.09150633413591353,
|
| 819 |
+
"learning_rate": 0.00018666555562259356,
|
| 820 |
+
"loss": 0.3718,
|
| 821 |
+
"step": 116
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.7572815533980582,
|
| 825 |
+
"grad_norm": 0.09139304143928584,
|
| 826 |
+
"learning_rate": 0.00018628539830324229,
|
| 827 |
+
"loss": 0.3583,
|
| 828 |
+
"step": 117
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.7637540453074434,
|
| 832 |
+
"grad_norm": 0.08557332691225407,
|
| 833 |
+
"learning_rate": 0.00018590029629711506,
|
| 834 |
+
"loss": 0.3481,
|
| 835 |
+
"step": 118
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.7702265372168284,
|
| 839 |
+
"grad_norm": 0.08670915390405,
|
| 840 |
+
"learning_rate": 0.00018551027167293768,
|
| 841 |
+
"loss": 0.3486,
|
| 842 |
+
"step": 119
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.7766990291262136,
|
| 846 |
+
"grad_norm": 0.085015484261482,
|
| 847 |
+
"learning_rate": 0.00018511534678153244,
|
| 848 |
+
"loss": 0.367,
|
| 849 |
+
"step": 120
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.7831715210355987,
|
| 853 |
+
"grad_norm": 0.09010500519447748,
|
| 854 |
+
"learning_rate": 0.0001847155442545372,
|
| 855 |
+
"loss": 0.3673,
|
| 856 |
+
"step": 121
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.7896440129449838,
|
| 860 |
+
"grad_norm": 0.08749250339043911,
|
| 861 |
+
"learning_rate": 0.00018431088700310844,
|
| 862 |
+
"loss": 0.3581,
|
| 863 |
+
"step": 122
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.7961165048543689,
|
| 867 |
+
"grad_norm": 0.09294869349273092,
|
| 868 |
+
"learning_rate": 0.00018390139821660855,
|
| 869 |
+
"loss": 0.3502,
|
| 870 |
+
"step": 123
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.8025889967637541,
|
| 874 |
+
"grad_norm": 0.08999457767358941,
|
| 875 |
+
"learning_rate": 0.00018348710136127655,
|
| 876 |
+
"loss": 0.3604,
|
| 877 |
+
"step": 124
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.8090614886731392,
|
| 881 |
+
"grad_norm": 0.08450140095669934,
|
| 882 |
+
"learning_rate": 0.0001830680201788836,
|
| 883 |
+
"loss": 0.3585,
|
| 884 |
+
"step": 125
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.8155339805825242,
|
| 888 |
+
"grad_norm": 0.0885173201152873,
|
| 889 |
+
"learning_rate": 0.00018264417868537244,
|
| 890 |
+
"loss": 0.36,
|
| 891 |
+
"step": 126
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.8220064724919094,
|
| 895 |
+
"grad_norm": 0.08249252827967349,
|
| 896 |
+
"learning_rate": 0.00018221560116948103,
|
| 897 |
+
"loss": 0.3456,
|
| 898 |
+
"step": 127
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.8284789644012945,
|
| 902 |
+
"grad_norm": 0.09259010876584428,
|
| 903 |
+
"learning_rate": 0.0001817823121913506,
|
| 904 |
+
"loss": 0.3523,
|
| 905 |
+
"step": 128
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.8349514563106796,
|
| 909 |
+
"grad_norm": 0.0829493165024249,
|
| 910 |
+
"learning_rate": 0.00018134433658111845,
|
| 911 |
+
"loss": 0.3502,
|
| 912 |
+
"step": 129
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.8414239482200647,
|
| 916 |
+
"grad_norm": 0.0870233230162085,
|
| 917 |
+
"learning_rate": 0.00018090169943749476,
|
| 918 |
+
"loss": 0.3503,
|
| 919 |
+
"step": 130
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.8478964401294499,
|
| 923 |
+
"grad_norm": 0.09013336960665569,
|
| 924 |
+
"learning_rate": 0.00018045442612632444,
|
| 925 |
+
"loss": 0.355,
|
| 926 |
+
"step": 131
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.8543689320388349,
|
| 930 |
+
"grad_norm": 0.0887494126249745,
|
| 931 |
+
"learning_rate": 0.00018000254227913348,
|
| 932 |
+
"loss": 0.3492,
|
| 933 |
+
"step": 132
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.86084142394822,
|
| 937 |
+
"grad_norm": 0.08245510607718151,
|
| 938 |
+
"learning_rate": 0.00017954607379166,
|
| 939 |
+
"loss": 0.3509,
|
| 940 |
+
"step": 133
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.8673139158576052,
|
| 944 |
+
"grad_norm": 0.08652334474586301,
|
| 945 |
+
"learning_rate": 0.00017908504682237047,
|
| 946 |
+
"loss": 0.3464,
|
| 947 |
+
"step": 134
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.8737864077669902,
|
| 951 |
+
"grad_norm": 0.0832923145834622,
|
| 952 |
+
"learning_rate": 0.00017861948779096046,
|
| 953 |
+
"loss": 0.3392,
|
| 954 |
+
"step": 135
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.8802588996763754,
|
| 958 |
+
"grad_norm": 0.0812465660556935,
|
| 959 |
+
"learning_rate": 0.0001781494233768408,
|
| 960 |
+
"loss": 0.333,
|
| 961 |
+
"step": 136
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.8867313915857605,
|
| 965 |
+
"grad_norm": 0.088984541060325,
|
| 966 |
+
"learning_rate": 0.00017767488051760857,
|
| 967 |
+
"loss": 0.3615,
|
| 968 |
+
"step": 137
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.8932038834951457,
|
| 972 |
+
"grad_norm": 0.0867087441619348,
|
| 973 |
+
"learning_rate": 0.00017719588640750336,
|
| 974 |
+
"loss": 0.3553,
|
| 975 |
+
"step": 138
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.8996763754045307,
|
| 979 |
+
"grad_norm": 0.08382301700407316,
|
| 980 |
+
"learning_rate": 0.00017671246849584903,
|
| 981 |
+
"loss": 0.3592,
|
| 982 |
+
"step": 139
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.9061488673139159,
|
| 986 |
+
"grad_norm": 0.0865944397225724,
|
| 987 |
+
"learning_rate": 0.0001762246544854807,
|
| 988 |
+
"loss": 0.3605,
|
| 989 |
+
"step": 140
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.912621359223301,
|
| 993 |
+
"grad_norm": 0.08628029323884819,
|
| 994 |
+
"learning_rate": 0.00017573247233115694,
|
| 995 |
+
"loss": 0.371,
|
| 996 |
+
"step": 141
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.919093851132686,
|
| 1000 |
+
"grad_norm": 0.08863123995414322,
|
| 1001 |
+
"learning_rate": 0.00017523595023795813,
|
| 1002 |
+
"loss": 0.3504,
|
| 1003 |
+
"step": 142
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.9255663430420712,
|
| 1007 |
+
"grad_norm": 0.08627162896846262,
|
| 1008 |
+
"learning_rate": 0.00017473511665966993,
|
| 1009 |
+
"loss": 0.3545,
|
| 1010 |
+
"step": 143
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.9320388349514563,
|
| 1014 |
+
"grad_norm": 0.08839402497347855,
|
| 1015 |
+
"learning_rate": 0.00017423000029715267,
|
| 1016 |
+
"loss": 0.3641,
|
| 1017 |
+
"step": 144
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.9385113268608414,
|
| 1021 |
+
"grad_norm": 0.08651737654873981,
|
| 1022 |
+
"learning_rate": 0.00017372063009669686,
|
| 1023 |
+
"loss": 0.3431,
|
| 1024 |
+
"step": 145
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.9449838187702265,
|
| 1028 |
+
"grad_norm": 0.07938521602972776,
|
| 1029 |
+
"learning_rate": 0.00017320703524836405,
|
| 1030 |
+
"loss": 0.3541,
|
| 1031 |
+
"step": 146
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.9514563106796117,
|
| 1035 |
+
"grad_norm": 0.08679861689760528,
|
| 1036 |
+
"learning_rate": 0.00017268924518431438,
|
| 1037 |
+
"loss": 0.3492,
|
| 1038 |
+
"step": 147
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.9579288025889967,
|
| 1042 |
+
"grad_norm": 0.08562044946225537,
|
| 1043 |
+
"learning_rate": 0.00017216728957711967,
|
| 1044 |
+
"loss": 0.3497,
|
| 1045 |
+
"step": 148
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.9644012944983819,
|
| 1049 |
+
"grad_norm": 0.08302055814489633,
|
| 1050 |
+
"learning_rate": 0.0001716411983380632,
|
| 1051 |
+
"loss": 0.3472,
|
| 1052 |
+
"step": 149
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.970873786407767,
|
| 1056 |
+
"grad_norm": 0.08275430206926206,
|
| 1057 |
+
"learning_rate": 0.00017111100161542545,
|
| 1058 |
+
"loss": 0.3383,
|
| 1059 |
+
"step": 150
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.9773462783171522,
|
| 1063 |
+
"grad_norm": 0.08114036438803068,
|
| 1064 |
+
"learning_rate": 0.00017057672979275656,
|
| 1065 |
+
"loss": 0.36,
|
| 1066 |
+
"step": 151
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.9838187702265372,
|
| 1070 |
+
"grad_norm": 0.0875761383748598,
|
| 1071 |
+
"learning_rate": 0.0001700384134871351,
|
| 1072 |
+
"loss": 0.3727,
|
| 1073 |
+
"step": 152
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.9902912621359223,
|
| 1077 |
+
"grad_norm": 0.0826665996105745,
|
| 1078 |
+
"learning_rate": 0.0001694960835474134,
|
| 1079 |
+
"loss": 0.3443,
|
| 1080 |
+
"step": 153
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.9967637540453075,
|
| 1084 |
+
"grad_norm": 0.08226814015705675,
|
| 1085 |
+
"learning_rate": 0.00016894977105244997,
|
| 1086 |
+
"loss": 0.3653,
|
| 1087 |
+
"step": 154
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.9967637540453075,
|
| 1091 |
+
"eval_loss": 0.3545331358909607,
|
| 1092 |
+
"eval_runtime": 38.077,
|
| 1093 |
+
"eval_samples_per_second": 27.287,
|
| 1094 |
+
"eval_steps_per_second": 0.867,
|
| 1095 |
+
"step": 154
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 1.0032362459546926,
|
| 1099 |
+
"grad_norm": 0.08335200234700724,
|
| 1100 |
+
"learning_rate": 0.0001683995073093283,
|
| 1101 |
+
"loss": 0.3492,
|
| 1102 |
+
"step": 155
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 1.0097087378640777,
|
| 1106 |
+
"grad_norm": 0.0865080297629732,
|
| 1107 |
+
"learning_rate": 0.00016784532385156285,
|
| 1108 |
+
"loss": 0.3377,
|
| 1109 |
+
"step": 156
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.0161812297734627,
|
| 1113 |
+
"grad_norm": 0.08441339355225498,
|
| 1114 |
+
"learning_rate": 0.0001672872524372919,
|
| 1115 |
+
"loss": 0.34,
|
| 1116 |
+
"step": 157
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 1.022653721682848,
|
| 1120 |
+
"grad_norm": 0.0801171782758195,
|
| 1121 |
+
"learning_rate": 0.00016672532504745778,
|
| 1122 |
+
"loss": 0.3354,
|
| 1123 |
+
"step": 158
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.029126213592233,
|
| 1127 |
+
"grad_norm": 0.08282538959029859,
|
| 1128 |
+
"learning_rate": 0.00016615957388397399,
|
| 1129 |
+
"loss": 0.3497,
|
| 1130 |
+
"step": 159
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 1.035598705501618,
|
| 1134 |
+
"grad_norm": 0.0883355169428613,
|
| 1135 |
+
"learning_rate": 0.00016559003136787988,
|
| 1136 |
+
"loss": 0.338,
|
| 1137 |
+
"step": 160
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 1.0420711974110033,
|
| 1141 |
+
"grad_norm": 0.08616766891600182,
|
| 1142 |
+
"learning_rate": 0.00016501673013748284,
|
| 1143 |
+
"loss": 0.3491,
|
| 1144 |
+
"step": 161
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 1.0485436893203883,
|
| 1148 |
+
"grad_norm": 0.08188626612719885,
|
| 1149 |
+
"learning_rate": 0.0001644397030464877,
|
| 1150 |
+
"loss": 0.3359,
|
| 1151 |
+
"step": 162
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 1.0550161812297734,
|
| 1155 |
+
"grad_norm": 0.08630887969447615,
|
| 1156 |
+
"learning_rate": 0.00016385898316211426,
|
| 1157 |
+
"loss": 0.3268,
|
| 1158 |
+
"step": 163
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 1.0614886731391586,
|
| 1162 |
+
"grad_norm": 0.09109351353175275,
|
| 1163 |
+
"learning_rate": 0.0001632746037632021,
|
| 1164 |
+
"loss": 0.3356,
|
| 1165 |
+
"step": 164
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.0679611650485437,
|
| 1169 |
+
"grad_norm": 0.08843718177405194,
|
| 1170 |
+
"learning_rate": 0.00016268659833830367,
|
| 1171 |
+
"loss": 0.3205,
|
| 1172 |
+
"step": 165
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 1.074433656957929,
|
| 1176 |
+
"grad_norm": 0.09214475868602304,
|
| 1177 |
+
"learning_rate": 0.00016209500058376515,
|
| 1178 |
+
"loss": 0.3433,
|
| 1179 |
+
"step": 166
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.080906148867314,
|
| 1183 |
+
"grad_norm": 0.08469945389831732,
|
| 1184 |
+
"learning_rate": 0.00016149984440179537,
|
| 1185 |
+
"loss": 0.3252,
|
| 1186 |
+
"step": 167
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.087378640776699,
|
| 1190 |
+
"grad_norm": 0.09167891300803974,
|
| 1191 |
+
"learning_rate": 0.00016090116389852306,
|
| 1192 |
+
"loss": 0.3372,
|
| 1193 |
+
"step": 168
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.0938511326860842,
|
| 1197 |
+
"grad_norm": 0.08439372223738191,
|
| 1198 |
+
"learning_rate": 0.00016029899338204233,
|
| 1199 |
+
"loss": 0.3339,
|
| 1200 |
+
"step": 169
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.1003236245954693,
|
| 1204 |
+
"grad_norm": 0.09671189992628862,
|
| 1205 |
+
"learning_rate": 0.0001596933673604467,
|
| 1206 |
+
"loss": 0.331,
|
| 1207 |
+
"step": 170
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.1067961165048543,
|
| 1211 |
+
"grad_norm": 0.08926808998386844,
|
| 1212 |
+
"learning_rate": 0.00015908432053985143,
|
| 1213 |
+
"loss": 0.3438,
|
| 1214 |
+
"step": 171
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 1.1132686084142396,
|
| 1218 |
+
"grad_norm": 0.08651511861187883,
|
| 1219 |
+
"learning_rate": 0.0001584718878224047,
|
| 1220 |
+
"loss": 0.3199,
|
| 1221 |
+
"step": 172
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.1197411003236246,
|
| 1225 |
+
"grad_norm": 0.09087672552517187,
|
| 1226 |
+
"learning_rate": 0.00015785610430428762,
|
| 1227 |
+
"loss": 0.3312,
|
| 1228 |
+
"step": 173
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.1262135922330097,
|
| 1232 |
+
"grad_norm": 0.08473819565220535,
|
| 1233 |
+
"learning_rate": 0.00015723700527370268,
|
| 1234 |
+
"loss": 0.3265,
|
| 1235 |
+
"step": 174
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.132686084142395,
|
| 1239 |
+
"grad_norm": 0.08959082008669542,
|
| 1240 |
+
"learning_rate": 0.00015661462620885199,
|
| 1241 |
+
"loss": 0.3361,
|
| 1242 |
+
"step": 175
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.13915857605178,
|
| 1246 |
+
"grad_norm": 0.08820985748174735,
|
| 1247 |
+
"learning_rate": 0.0001559890027759037,
|
| 1248 |
+
"loss": 0.3352,
|
| 1249 |
+
"step": 176
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.145631067961165,
|
| 1253 |
+
"grad_norm": 0.08959538258978733,
|
| 1254 |
+
"learning_rate": 0.00015536017082694846,
|
| 1255 |
+
"loss": 0.353,
|
| 1256 |
+
"step": 177
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.1521035598705502,
|
| 1260 |
+
"grad_norm": 0.08556029404315754,
|
| 1261 |
+
"learning_rate": 0.0001547281663979446,
|
| 1262 |
+
"loss": 0.338,
|
| 1263 |
+
"step": 178
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.1585760517799353,
|
| 1267 |
+
"grad_norm": 0.08651722009131164,
|
| 1268 |
+
"learning_rate": 0.00015409302570665325,
|
| 1269 |
+
"loss": 0.3428,
|
| 1270 |
+
"step": 179
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 1.1650485436893203,
|
| 1274 |
+
"grad_norm": 0.09127001503693886,
|
| 1275 |
+
"learning_rate": 0.00015345478515056267,
|
| 1276 |
+
"loss": 0.3196,
|
| 1277 |
+
"step": 180
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.1715210355987056,
|
| 1281 |
+
"grad_norm": 0.08856796253981307,
|
| 1282 |
+
"learning_rate": 0.00015281348130480272,
|
| 1283 |
+
"loss": 0.3309,
|
| 1284 |
+
"step": 181
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.1779935275080906,
|
| 1288 |
+
"grad_norm": 0.08742855227811815,
|
| 1289 |
+
"learning_rate": 0.00015216915092004847,
|
| 1290 |
+
"loss": 0.3273,
|
| 1291 |
+
"step": 182
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.1844660194174756,
|
| 1295 |
+
"grad_norm": 0.09076657129408205,
|
| 1296 |
+
"learning_rate": 0.0001515218309204145,
|
| 1297 |
+
"loss": 0.3324,
|
| 1298 |
+
"step": 183
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.190938511326861,
|
| 1302 |
+
"grad_norm": 0.09296734178135176,
|
| 1303 |
+
"learning_rate": 0.00015087155840133888,
|
| 1304 |
+
"loss": 0.3443,
|
| 1305 |
+
"step": 184
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.197411003236246,
|
| 1309 |
+
"grad_norm": 0.08431288630485151,
|
| 1310 |
+
"learning_rate": 0.00015021837062745714,
|
| 1311 |
+
"loss": 0.3383,
|
| 1312 |
+
"step": 185
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 1.203883495145631,
|
| 1316 |
+
"grad_norm": 0.08374691953765048,
|
| 1317 |
+
"learning_rate": 0.00014956230503046703,
|
| 1318 |
+
"loss": 0.3368,
|
| 1319 |
+
"step": 186
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.2103559870550162,
|
| 1323 |
+
"grad_norm": 0.08899017726880568,
|
| 1324 |
+
"learning_rate": 0.00014890339920698334,
|
| 1325 |
+
"loss": 0.3368,
|
| 1326 |
+
"step": 187
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 1.2168284789644013,
|
| 1330 |
+
"grad_norm": 0.08562973948918146,
|
| 1331 |
+
"learning_rate": 0.00014824169091638337,
|
| 1332 |
+
"loss": 0.3397,
|
| 1333 |
+
"step": 188
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.2233009708737863,
|
| 1337 |
+
"grad_norm": 0.08827756751106554,
|
| 1338 |
+
"learning_rate": 0.00014757721807864317,
|
| 1339 |
+
"loss": 0.3405,
|
| 1340 |
+
"step": 189
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.2297734627831716,
|
| 1344 |
+
"grad_norm": 0.08736187981084419,
|
| 1345 |
+
"learning_rate": 0.0001469100187721644,
|
| 1346 |
+
"loss": 0.3262,
|
| 1347 |
+
"step": 190
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 1.2362459546925566,
|
| 1351 |
+
"grad_norm": 0.09050203121679472,
|
| 1352 |
+
"learning_rate": 0.0001462401312315922,
|
| 1353 |
+
"loss": 0.3345,
|
| 1354 |
+
"step": 191
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 1.2427184466019416,
|
| 1358 |
+
"grad_norm": 0.08534856926861663,
|
| 1359 |
+
"learning_rate": 0.00014556759384562416,
|
| 1360 |
+
"loss": 0.3277,
|
| 1361 |
+
"step": 192
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 1.249190938511327,
|
| 1365 |
+
"grad_norm": 0.08611809442550349,
|
| 1366 |
+
"learning_rate": 0.00014489244515481046,
|
| 1367 |
+
"loss": 0.3253,
|
| 1368 |
+
"step": 193
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 1.255663430420712,
|
| 1372 |
+
"grad_norm": 0.08688819258876039,
|
| 1373 |
+
"learning_rate": 0.0001442147238493451,
|
| 1374 |
+
"loss": 0.3349,
|
| 1375 |
+
"step": 194
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.262135922330097,
|
| 1379 |
+
"grad_norm": 0.08416654255653569,
|
| 1380 |
+
"learning_rate": 0.00014353446876684892,
|
| 1381 |
+
"loss": 0.3308,
|
| 1382 |
+
"step": 195
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.2686084142394822,
|
| 1386 |
+
"grad_norm": 0.09183514403063862,
|
| 1387 |
+
"learning_rate": 0.0001428517188901437,
|
| 1388 |
+
"loss": 0.3383,
|
| 1389 |
+
"step": 196
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.2750809061488673,
|
| 1393 |
+
"grad_norm": 0.08749863979244098,
|
| 1394 |
+
"learning_rate": 0.0001421665133450184,
|
| 1395 |
+
"loss": 0.3294,
|
| 1396 |
+
"step": 197
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 1.2815533980582523,
|
| 1400 |
+
"grad_norm": 0.08786155414487208,
|
| 1401 |
+
"learning_rate": 0.00014147889139798708,
|
| 1402 |
+
"loss": 0.332,
|
| 1403 |
+
"step": 198
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.2880258899676376,
|
| 1407 |
+
"grad_norm": 0.08754391459720788,
|
| 1408 |
+
"learning_rate": 0.00014078889245403844,
|
| 1409 |
+
"loss": 0.3268,
|
| 1410 |
+
"step": 199
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 1.2944983818770226,
|
| 1414 |
+
"grad_norm": 0.09026975504709552,
|
| 1415 |
+
"learning_rate": 0.0001400965560543778,
|
| 1416 |
+
"loss": 0.3344,
|
| 1417 |
+
"step": 200
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.3009708737864076,
|
| 1421 |
+
"grad_norm": 0.08953266507881985,
|
| 1422 |
+
"learning_rate": 0.0001394019218741612,
|
| 1423 |
+
"loss": 0.3355,
|
| 1424 |
+
"step": 201
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.307443365695793,
|
| 1428 |
+
"grad_norm": 0.08868590943641125,
|
| 1429 |
+
"learning_rate": 0.00013870502972022173,
|
| 1430 |
+
"loss": 0.3282,
|
| 1431 |
+
"step": 202
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 1.313915857605178,
|
| 1435 |
+
"grad_norm": 0.09182148747745249,
|
| 1436 |
+
"learning_rate": 0.00013800591952878825,
|
| 1437 |
+
"loss": 0.3394,
|
| 1438 |
+
"step": 203
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.3203883495145632,
|
| 1442 |
+
"grad_norm": 0.08835010573043636,
|
| 1443 |
+
"learning_rate": 0.00013730463136319692,
|
| 1444 |
+
"loss": 0.3316,
|
| 1445 |
+
"step": 204
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 1.3268608414239482,
|
| 1449 |
+
"grad_norm": 0.08573944970518392,
|
| 1450 |
+
"learning_rate": 0.00013660120541159537,
|
| 1451 |
+
"loss": 0.337,
|
| 1452 |
+
"step": 205
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 1.3333333333333333,
|
| 1456 |
+
"grad_norm": 0.08411384694221385,
|
| 1457 |
+
"learning_rate": 0.00013589568198463944,
|
| 1458 |
+
"loss": 0.3356,
|
| 1459 |
+
"step": 206
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.3398058252427185,
|
| 1463 |
+
"grad_norm": 0.08762468166968049,
|
| 1464 |
+
"learning_rate": 0.0001351881015131833,
|
| 1465 |
+
"loss": 0.3388,
|
| 1466 |
+
"step": 207
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.3462783171521036,
|
| 1470 |
+
"grad_norm": 0.08837885567018512,
|
| 1471 |
+
"learning_rate": 0.00013447850454596265,
|
| 1472 |
+
"loss": 0.3354,
|
| 1473 |
+
"step": 208
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 1.3527508090614886,
|
| 1477 |
+
"grad_norm": 0.08823334999060412,
|
| 1478 |
+
"learning_rate": 0.00013376693174727065,
|
| 1479 |
+
"loss": 0.3341,
|
| 1480 |
+
"step": 209
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.3592233009708738,
|
| 1484 |
+
"grad_norm": 0.08665353483827595,
|
| 1485 |
+
"learning_rate": 0.00013305342389462792,
|
| 1486 |
+
"loss": 0.3282,
|
| 1487 |
+
"step": 210
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 1.3656957928802589,
|
| 1491 |
+
"grad_norm": 0.08780215390647159,
|
| 1492 |
+
"learning_rate": 0.00013233802187644566,
|
| 1493 |
+
"loss": 0.3245,
|
| 1494 |
+
"step": 211
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.3721682847896441,
|
| 1498 |
+
"grad_norm": 0.0880005148626301,
|
| 1499 |
+
"learning_rate": 0.0001316207666896824,
|
| 1500 |
+
"loss": 0.3264,
|
| 1501 |
+
"step": 212
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.3786407766990292,
|
| 1505 |
+
"grad_norm": 0.09350061568656749,
|
| 1506 |
+
"learning_rate": 0.00013090169943749476,
|
| 1507 |
+
"loss": 0.33,
|
| 1508 |
+
"step": 213
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.3851132686084142,
|
| 1512 |
+
"grad_norm": 0.0869291546207145,
|
| 1513 |
+
"learning_rate": 0.00013018086132688184,
|
| 1514 |
+
"loss": 0.3219,
|
| 1515 |
+
"step": 214
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 1.3915857605177995,
|
| 1519 |
+
"grad_norm": 0.0926767409167509,
|
| 1520 |
+
"learning_rate": 0.0001294582936663239,
|
| 1521 |
+
"loss": 0.3294,
|
| 1522 |
+
"step": 215
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.3980582524271845,
|
| 1526 |
+
"grad_norm": 0.08850918948495869,
|
| 1527 |
+
"learning_rate": 0.00012873403786341513,
|
| 1528 |
+
"loss": 0.3362,
|
| 1529 |
+
"step": 216
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 1.4045307443365695,
|
| 1533 |
+
"grad_norm": 0.08797553458327365,
|
| 1534 |
+
"learning_rate": 0.00012800813542249072,
|
| 1535 |
+
"loss": 0.3293,
|
| 1536 |
+
"step": 217
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 1.4110032362459548,
|
| 1540 |
+
"grad_norm": 0.08816787591032364,
|
| 1541 |
+
"learning_rate": 0.00012728062794224832,
|
| 1542 |
+
"loss": 0.3297,
|
| 1543 |
+
"step": 218
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.4174757281553398,
|
| 1547 |
+
"grad_norm": 0.08844464048062073,
|
| 1548 |
+
"learning_rate": 0.0001265515571133643,
|
| 1549 |
+
"loss": 0.3405,
|
| 1550 |
+
"step": 219
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 1.4239482200647249,
|
| 1554 |
+
"grad_norm": 0.08777743550690727,
|
| 1555 |
+
"learning_rate": 0.00012582096471610467,
|
| 1556 |
+
"loss": 0.33,
|
| 1557 |
+
"step": 220
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.4304207119741101,
|
| 1561 |
+
"grad_norm": 0.0841701096943823,
|
| 1562 |
+
"learning_rate": 0.00012508889261793059,
|
| 1563 |
+
"loss": 0.3243,
|
| 1564 |
+
"step": 221
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.4368932038834952,
|
| 1568 |
+
"grad_norm": 0.08211677315785172,
|
| 1569 |
+
"learning_rate": 0.0001243553827710992,
|
| 1570 |
+
"loss": 0.3162,
|
| 1571 |
+
"step": 222
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.4433656957928802,
|
| 1575 |
+
"grad_norm": 0.09229614871738043,
|
| 1576 |
+
"learning_rate": 0.00012362047721025968,
|
| 1577 |
+
"loss": 0.3214,
|
| 1578 |
+
"step": 223
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 1.4498381877022655,
|
| 1582 |
+
"grad_norm": 0.08827293639891444,
|
| 1583 |
+
"learning_rate": 0.00012288421805004414,
|
| 1584 |
+
"loss": 0.3388,
|
| 1585 |
+
"step": 224
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.4563106796116505,
|
| 1589 |
+
"grad_norm": 0.08650939834799051,
|
| 1590 |
+
"learning_rate": 0.0001221466474826543,
|
| 1591 |
+
"loss": 0.3336,
|
| 1592 |
+
"step": 225
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 1.4627831715210355,
|
| 1596 |
+
"grad_norm": 0.08745414194563839,
|
| 1597 |
+
"learning_rate": 0.00012140780777544367,
|
| 1598 |
+
"loss": 0.3249,
|
| 1599 |
+
"step": 226
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.4692556634304208,
|
| 1603 |
+
"grad_norm": 0.08966472430098649,
|
| 1604 |
+
"learning_rate": 0.00012066774126849529,
|
| 1605 |
+
"loss": 0.3358,
|
| 1606 |
+
"step": 227
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.4757281553398058,
|
| 1610 |
+
"grad_norm": 0.08547479634425736,
|
| 1611 |
+
"learning_rate": 0.00011992649037219545,
|
| 1612 |
+
"loss": 0.3415,
|
| 1613 |
+
"step": 228
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.4822006472491909,
|
| 1617 |
+
"grad_norm": 0.08630753448523301,
|
| 1618 |
+
"learning_rate": 0.0001191840975648032,
|
| 1619 |
+
"loss": 0.3369,
|
| 1620 |
+
"step": 229
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 1.4886731391585761,
|
| 1624 |
+
"grad_norm": 0.08171192229188709,
|
| 1625 |
+
"learning_rate": 0.00011844060539001618,
|
| 1626 |
+
"loss": 0.3207,
|
| 1627 |
+
"step": 230
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.4951456310679612,
|
| 1631 |
+
"grad_norm": 0.08413278558608812,
|
| 1632 |
+
"learning_rate": 0.00011769605645453265,
|
| 1633 |
+
"loss": 0.3219,
|
| 1634 |
+
"step": 231
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 1.5016181229773462,
|
| 1638 |
+
"grad_norm": 0.08498283806676993,
|
| 1639 |
+
"learning_rate": 0.00011695049342560968,
|
| 1640 |
+
"loss": 0.3339,
|
| 1641 |
+
"step": 232
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.5080906148867315,
|
| 1645 |
+
"grad_norm": 0.08365977359578973,
|
| 1646 |
+
"learning_rate": 0.00011620395902861822,
|
| 1647 |
+
"loss": 0.3297,
|
| 1648 |
+
"step": 233
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.5145631067961165,
|
| 1652 |
+
"grad_norm": 0.0858081054599164,
|
| 1653 |
+
"learning_rate": 0.00011545649604459466,
|
| 1654 |
+
"loss": 0.3305,
|
| 1655 |
+
"step": 234
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.5210355987055015,
|
| 1659 |
+
"grad_norm": 0.08916196136213667,
|
| 1660 |
+
"learning_rate": 0.00011470814730778905,
|
| 1661 |
+
"loss": 0.3352,
|
| 1662 |
+
"step": 235
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.5275080906148868,
|
| 1666 |
+
"grad_norm": 0.08650692110960809,
|
| 1667 |
+
"learning_rate": 0.00011395895570321064,
|
| 1668 |
+
"loss": 0.3347,
|
| 1669 |
+
"step": 236
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.5339805825242718,
|
| 1673 |
+
"grad_norm": 0.08839636357345756,
|
| 1674 |
+
"learning_rate": 0.00011320896416417026,
|
| 1675 |
+
"loss": 0.3284,
|
| 1676 |
+
"step": 237
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 1.5404530744336569,
|
| 1680 |
+
"grad_norm": 0.08477612347310413,
|
| 1681 |
+
"learning_rate": 0.00011245821566981976,
|
| 1682 |
+
"loss": 0.3331,
|
| 1683 |
+
"step": 238
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 1.5469255663430421,
|
| 1687 |
+
"grad_norm": 0.08158083954977735,
|
| 1688 |
+
"learning_rate": 0.00011170675324268942,
|
| 1689 |
+
"loss": 0.3284,
|
| 1690 |
+
"step": 239
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.5533980582524272,
|
| 1694 |
+
"grad_norm": 0.08210674030802527,
|
| 1695 |
+
"learning_rate": 0.00011095461994622209,
|
| 1696 |
+
"loss": 0.3152,
|
| 1697 |
+
"step": 240
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.5598705501618122,
|
| 1701 |
+
"grad_norm": 0.08164700760175722,
|
| 1702 |
+
"learning_rate": 0.00011020185888230571,
|
| 1703 |
+
"loss": 0.3212,
|
| 1704 |
+
"step": 241
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 1.5663430420711975,
|
| 1708 |
+
"grad_norm": 0.08252073365071766,
|
| 1709 |
+
"learning_rate": 0.00010944851318880314,
|
| 1710 |
+
"loss": 0.3349,
|
| 1711 |
+
"step": 242
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 1.5728155339805825,
|
| 1715 |
+
"grad_norm": 0.0847504452125806,
|
| 1716 |
+
"learning_rate": 0.00010869462603708011,
|
| 1717 |
+
"loss": 0.3355,
|
| 1718 |
+
"step": 243
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 1.5792880258899675,
|
| 1722 |
+
"grad_norm": 0.08651133299846145,
|
| 1723 |
+
"learning_rate": 0.00010794024062953123,
|
| 1724 |
+
"loss": 0.3295,
|
| 1725 |
+
"step": 244
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 1.5857605177993528,
|
| 1729 |
+
"grad_norm": 0.08848888104315868,
|
| 1730 |
+
"learning_rate": 0.00010718540019710432,
|
| 1731 |
+
"loss": 0.334,
|
| 1732 |
+
"step": 245
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 1.5922330097087378,
|
| 1736 |
+
"grad_norm": 0.0838746079221116,
|
| 1737 |
+
"learning_rate": 0.00010643014799682296,
|
| 1738 |
+
"loss": 0.3236,
|
| 1739 |
+
"step": 246
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.5987055016181229,
|
| 1743 |
+
"grad_norm": 0.0834887968620315,
|
| 1744 |
+
"learning_rate": 0.00010567452730930743,
|
| 1745 |
+
"loss": 0.3231,
|
| 1746 |
+
"step": 247
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 1.6051779935275081,
|
| 1750 |
+
"grad_norm": 0.08903656754039865,
|
| 1751 |
+
"learning_rate": 0.00010491858143629469,
|
| 1752 |
+
"loss": 0.3303,
|
| 1753 |
+
"step": 248
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.6116504854368932,
|
| 1757 |
+
"grad_norm": 0.08833888604572854,
|
| 1758 |
+
"learning_rate": 0.00010416235369815693,
|
| 1759 |
+
"loss": 0.3417,
|
| 1760 |
+
"step": 249
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 1.6181229773462782,
|
| 1764 |
+
"grad_norm": 0.08946556956131413,
|
| 1765 |
+
"learning_rate": 0.00010340588743141879,
|
| 1766 |
+
"loss": 0.3341,
|
| 1767 |
+
"step": 250
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 1.6245954692556634,
|
| 1771 |
+
"grad_norm": 0.09195837741425467,
|
| 1772 |
+
"learning_rate": 0.00010264922598627418,
|
| 1773 |
+
"loss": 0.3292,
|
| 1774 |
+
"step": 251
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 1.6310679611650487,
|
| 1778 |
+
"grad_norm": 0.08669211913864286,
|
| 1779 |
+
"learning_rate": 0.0001018924127241019,
|
| 1780 |
+
"loss": 0.329,
|
| 1781 |
+
"step": 252
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.6375404530744335,
|
| 1785 |
+
"grad_norm": 0.08493981781947878,
|
| 1786 |
+
"learning_rate": 0.00010113549101498086,
|
| 1787 |
+
"loss": 0.317,
|
| 1788 |
+
"step": 253
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 1.6440129449838188,
|
| 1792 |
+
"grad_norm": 0.08859940503358235,
|
| 1793 |
+
"learning_rate": 0.00010037850423520454,
|
| 1794 |
+
"loss": 0.3239,
|
| 1795 |
+
"step": 254
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 1.650485436893204,
|
| 1799 |
+
"grad_norm": 0.08691997093628756,
|
| 1800 |
+
"learning_rate": 9.962149576479545e-05,
|
| 1801 |
+
"loss": 0.3276,
|
| 1802 |
+
"step": 255
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 1.6569579288025889,
|
| 1806 |
+
"grad_norm": 0.08182220008549855,
|
| 1807 |
+
"learning_rate": 9.886450898501917e-05,
|
| 1808 |
+
"loss": 0.322,
|
| 1809 |
+
"step": 256
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 1.6634304207119741,
|
| 1813 |
+
"grad_norm": 0.08983867569693547,
|
| 1814 |
+
"learning_rate": 9.810758727589813e-05,
|
| 1815 |
+
"loss": 0.3289,
|
| 1816 |
+
"step": 257
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 1.6699029126213594,
|
| 1820 |
+
"grad_norm": 0.08440426328661797,
|
| 1821 |
+
"learning_rate": 9.735077401372583e-05,
|
| 1822 |
+
"loss": 0.335,
|
| 1823 |
+
"step": 258
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 1.6763754045307442,
|
| 1827 |
+
"grad_norm": 0.09081741368893546,
|
| 1828 |
+
"learning_rate": 9.659411256858122e-05,
|
| 1829 |
+
"loss": 0.3285,
|
| 1830 |
+
"step": 259
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 1.6828478964401294,
|
| 1834 |
+
"grad_norm": 0.08638692257401422,
|
| 1835 |
+
"learning_rate": 9.583764630184311e-05,
|
| 1836 |
+
"loss": 0.3283,
|
| 1837 |
+
"step": 260
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 1.6893203883495147,
|
| 1841 |
+
"grad_norm": 0.08661785904833329,
|
| 1842 |
+
"learning_rate": 9.508141856370532e-05,
|
| 1843 |
+
"loss": 0.3315,
|
| 1844 |
+
"step": 261
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 1.6957928802588995,
|
| 1848 |
+
"grad_norm": 0.08867942067172753,
|
| 1849 |
+
"learning_rate": 9.432547269069261e-05,
|
| 1850 |
+
"loss": 0.3298,
|
| 1851 |
+
"step": 262
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 1.7022653721682848,
|
| 1855 |
+
"grad_norm": 0.08121634636493791,
|
| 1856 |
+
"learning_rate": 9.356985200317709e-05,
|
| 1857 |
+
"loss": 0.3264,
|
| 1858 |
+
"step": 263
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 1.70873786407767,
|
| 1862 |
+
"grad_norm": 0.0857656466474692,
|
| 1863 |
+
"learning_rate": 9.281459980289567e-05,
|
| 1864 |
+
"loss": 0.3265,
|
| 1865 |
+
"step": 264
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 1.715210355987055,
|
| 1869 |
+
"grad_norm": 0.08325063625580009,
|
| 1870 |
+
"learning_rate": 9.205975937046879e-05,
|
| 1871 |
+
"loss": 0.3248,
|
| 1872 |
+
"step": 265
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 1.72168284789644,
|
| 1876 |
+
"grad_norm": 0.09125541887572731,
|
| 1877 |
+
"learning_rate": 9.130537396291994e-05,
|
| 1878 |
+
"loss": 0.3336,
|
| 1879 |
+
"step": 266
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 1.7281553398058254,
|
| 1883 |
+
"grad_norm": 0.0858478153556312,
|
| 1884 |
+
"learning_rate": 9.055148681119688e-05,
|
| 1885 |
+
"loss": 0.3335,
|
| 1886 |
+
"step": 267
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 1.7346278317152104,
|
| 1890 |
+
"grad_norm": 0.08412291629199992,
|
| 1891 |
+
"learning_rate": 8.979814111769431e-05,
|
| 1892 |
+
"loss": 0.3247,
|
| 1893 |
+
"step": 268
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 1.7411003236245954,
|
| 1897 |
+
"grad_norm": 0.08342610941802514,
|
| 1898 |
+
"learning_rate": 8.904538005377794e-05,
|
| 1899 |
+
"loss": 0.3271,
|
| 1900 |
+
"step": 269
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 1.7475728155339807,
|
| 1904 |
+
"grad_norm": 0.08566344893624273,
|
| 1905 |
+
"learning_rate": 8.829324675731059e-05,
|
| 1906 |
+
"loss": 0.3342,
|
| 1907 |
+
"step": 270
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 1.7540453074433657,
|
| 1911 |
+
"grad_norm": 0.08349328064040033,
|
| 1912 |
+
"learning_rate": 8.754178433018025e-05,
|
| 1913 |
+
"loss": 0.3169,
|
| 1914 |
+
"step": 271
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 1.7605177993527508,
|
| 1918 |
+
"grad_norm": 0.08290650287411376,
|
| 1919 |
+
"learning_rate": 8.679103583582979e-05,
|
| 1920 |
+
"loss": 0.32,
|
| 1921 |
+
"step": 272
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 1.766990291262136,
|
| 1925 |
+
"grad_norm": 0.08355664601674428,
|
| 1926 |
+
"learning_rate": 8.604104429678935e-05,
|
| 1927 |
+
"loss": 0.3284,
|
| 1928 |
+
"step": 273
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 1.773462783171521,
|
| 1932 |
+
"grad_norm": 0.08088172666023426,
|
| 1933 |
+
"learning_rate": 8.529185269221097e-05,
|
| 1934 |
+
"loss": 0.3189,
|
| 1935 |
+
"step": 274
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.779935275080906,
|
| 1939 |
+
"grad_norm": 0.08494057369308265,
|
| 1940 |
+
"learning_rate": 8.45435039554054e-05,
|
| 1941 |
+
"loss": 0.3175,
|
| 1942 |
+
"step": 275
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 1.7864077669902914,
|
| 1946 |
+
"grad_norm": 0.08435898747241219,
|
| 1947 |
+
"learning_rate": 8.379604097138179e-05,
|
| 1948 |
+
"loss": 0.3359,
|
| 1949 |
+
"step": 276
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 1.7928802588996764,
|
| 1953 |
+
"grad_norm": 0.08369793345019513,
|
| 1954 |
+
"learning_rate": 8.304950657439033e-05,
|
| 1955 |
+
"loss": 0.3277,
|
| 1956 |
+
"step": 277
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 1.7993527508090614,
|
| 1960 |
+
"grad_norm": 0.08234674894169963,
|
| 1961 |
+
"learning_rate": 8.230394354546737e-05,
|
| 1962 |
+
"loss": 0.3106,
|
| 1963 |
+
"step": 278
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 1.8058252427184467,
|
| 1967 |
+
"grad_norm": 0.08480659421208028,
|
| 1968 |
+
"learning_rate": 8.15593946099838e-05,
|
| 1969 |
+
"loss": 0.3219,
|
| 1970 |
+
"step": 279
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 1.8122977346278317,
|
| 1974 |
+
"grad_norm": 0.087986920120639,
|
| 1975 |
+
"learning_rate": 8.08159024351968e-05,
|
| 1976 |
+
"loss": 0.3442,
|
| 1977 |
+
"step": 280
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 1.8187702265372168,
|
| 1981 |
+
"grad_norm": 0.0839196545152189,
|
| 1982 |
+
"learning_rate": 8.007350962780456e-05,
|
| 1983 |
+
"loss": 0.3216,
|
| 1984 |
+
"step": 281
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 1.825242718446602,
|
| 1988 |
+
"grad_norm": 0.08396185408224062,
|
| 1989 |
+
"learning_rate": 7.93322587315047e-05,
|
| 1990 |
+
"loss": 0.3249,
|
| 1991 |
+
"step": 282
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 1.831715210355987,
|
| 1995 |
+
"grad_norm": 0.08739956374343912,
|
| 1996 |
+
"learning_rate": 7.859219222455634e-05,
|
| 1997 |
+
"loss": 0.3366,
|
| 1998 |
+
"step": 283
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 1.838187702265372,
|
| 2002 |
+
"grad_norm": 0.08539428449034953,
|
| 2003 |
+
"learning_rate": 7.785335251734573e-05,
|
| 2004 |
+
"loss": 0.3399,
|
| 2005 |
+
"step": 284
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 1.8446601941747574,
|
| 2009 |
+
"grad_norm": 0.08443259129123379,
|
| 2010 |
+
"learning_rate": 7.711578194995589e-05,
|
| 2011 |
+
"loss": 0.326,
|
| 2012 |
+
"step": 285
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 1.8511326860841424,
|
| 2016 |
+
"grad_norm": 0.08226727888900415,
|
| 2017 |
+
"learning_rate": 7.637952278974034e-05,
|
| 2018 |
+
"loss": 0.3122,
|
| 2019 |
+
"step": 286
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 1.8576051779935274,
|
| 2023 |
+
"grad_norm": 0.08664703990759018,
|
| 2024 |
+
"learning_rate": 7.564461722890081e-05,
|
| 2025 |
+
"loss": 0.3276,
|
| 2026 |
+
"step": 287
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.8640776699029127,
|
| 2030 |
+
"grad_norm": 0.08284483357408984,
|
| 2031 |
+
"learning_rate": 7.491110738206942e-05,
|
| 2032 |
+
"loss": 0.3331,
|
| 2033 |
+
"step": 288
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 1.8705501618122977,
|
| 2037 |
+
"grad_norm": 0.08547813142706377,
|
| 2038 |
+
"learning_rate": 7.417903528389534e-05,
|
| 2039 |
+
"loss": 0.3108,
|
| 2040 |
+
"step": 289
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 1.8770226537216828,
|
| 2044 |
+
"grad_norm": 0.08375066019242061,
|
| 2045 |
+
"learning_rate": 7.344844288663571e-05,
|
| 2046 |
+
"loss": 0.3356,
|
| 2047 |
+
"step": 290
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 1.883495145631068,
|
| 2051 |
+
"grad_norm": 0.08564167180521602,
|
| 2052 |
+
"learning_rate": 7.27193720577517e-05,
|
| 2053 |
+
"loss": 0.3264,
|
| 2054 |
+
"step": 291
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 1.889967637540453,
|
| 2058 |
+
"grad_norm": 0.08560457179852256,
|
| 2059 |
+
"learning_rate": 7.19918645775093e-05,
|
| 2060 |
+
"loss": 0.3274,
|
| 2061 |
+
"step": 292
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 1.896440129449838,
|
| 2065 |
+
"grad_norm": 0.08594335919666654,
|
| 2066 |
+
"learning_rate": 7.126596213658488e-05,
|
| 2067 |
+
"loss": 0.3268,
|
| 2068 |
+
"step": 293
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 1.9029126213592233,
|
| 2072 |
+
"grad_norm": 0.086743773673858,
|
| 2073 |
+
"learning_rate": 7.05417063336761e-05,
|
| 2074 |
+
"loss": 0.3291,
|
| 2075 |
+
"step": 294
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 1.9093851132686084,
|
| 2079 |
+
"grad_norm": 0.08512277813785797,
|
| 2080 |
+
"learning_rate": 6.981913867311819e-05,
|
| 2081 |
+
"loss": 0.3199,
|
| 2082 |
+
"step": 295
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 1.9158576051779934,
|
| 2086 |
+
"grad_norm": 0.08482302988433711,
|
| 2087 |
+
"learning_rate": 6.909830056250527e-05,
|
| 2088 |
+
"loss": 0.3308,
|
| 2089 |
+
"step": 296
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 1.9223300970873787,
|
| 2093 |
+
"grad_norm": 0.08336748744395403,
|
| 2094 |
+
"learning_rate": 6.83792333103176e-05,
|
| 2095 |
+
"loss": 0.3303,
|
| 2096 |
+
"step": 297
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 1.9288025889967637,
|
| 2100 |
+
"grad_norm": 0.08685899308663286,
|
| 2101 |
+
"learning_rate": 6.766197812355438e-05,
|
| 2102 |
+
"loss": 0.3156,
|
| 2103 |
+
"step": 298
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 1.9352750809061487,
|
| 2107 |
+
"grad_norm": 0.08670025171649215,
|
| 2108 |
+
"learning_rate": 6.69465761053721e-05,
|
| 2109 |
+
"loss": 0.3263,
|
| 2110 |
+
"step": 299
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 1.941747572815534,
|
| 2114 |
+
"grad_norm": 0.08367058440741144,
|
| 2115 |
+
"learning_rate": 6.623306825272937e-05,
|
| 2116 |
+
"loss": 0.3184,
|
| 2117 |
+
"step": 300
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.948220064724919,
|
| 2121 |
+
"grad_norm": 0.08265849338783704,
|
| 2122 |
+
"learning_rate": 6.552149545403739e-05,
|
| 2123 |
+
"loss": 0.3141,
|
| 2124 |
+
"step": 301
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 1.954692556634304,
|
| 2128 |
+
"grad_norm": 0.08667064648503052,
|
| 2129 |
+
"learning_rate": 6.48118984868167e-05,
|
| 2130 |
+
"loss": 0.3176,
|
| 2131 |
+
"step": 302
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 1.9611650485436893,
|
| 2135 |
+
"grad_norm": 0.08624477625370294,
|
| 2136 |
+
"learning_rate": 6.410431801536058e-05,
|
| 2137 |
+
"loss": 0.3245,
|
| 2138 |
+
"step": 303
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 1.9676375404530746,
|
| 2142 |
+
"grad_norm": 0.0863914502801744,
|
| 2143 |
+
"learning_rate": 6.339879458840465e-05,
|
| 2144 |
+
"loss": 0.3345,
|
| 2145 |
+
"step": 304
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 1.9741100323624594,
|
| 2149 |
+
"grad_norm": 0.0841627898701783,
|
| 2150 |
+
"learning_rate": 6.269536863680307e-05,
|
| 2151 |
+
"loss": 0.3194,
|
| 2152 |
+
"step": 305
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 1.9805825242718447,
|
| 2156 |
+
"grad_norm": 0.0854832505200848,
|
| 2157 |
+
"learning_rate": 6.199408047121174e-05,
|
| 2158 |
+
"loss": 0.3396,
|
| 2159 |
+
"step": 306
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 1.98705501618123,
|
| 2163 |
+
"grad_norm": 0.07970724191486213,
|
| 2164 |
+
"learning_rate": 6.129497027977829e-05,
|
| 2165 |
+
"loss": 0.3228,
|
| 2166 |
+
"step": 307
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 1.9935275080906147,
|
| 2170 |
+
"grad_norm": 0.08216914347596244,
|
| 2171 |
+
"learning_rate": 6.059807812583883e-05,
|
| 2172 |
+
"loss": 0.3119,
|
| 2173 |
+
"step": 308
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 2.0,
|
| 2177 |
+
"grad_norm": 0.08150138639289872,
|
| 2178 |
+
"learning_rate": 5.990344394562226e-05,
|
| 2179 |
+
"loss": 0.3166,
|
| 2180 |
+
"step": 309
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 2.0,
|
| 2184 |
+
"eval_loss": 0.3374881148338318,
|
| 2185 |
+
"eval_runtime": 35.4418,
|
| 2186 |
+
"eval_samples_per_second": 29.316,
|
| 2187 |
+
"eval_steps_per_second": 0.931,
|
| 2188 |
+
"step": 309
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 2.0064724919093853,
|
| 2192 |
+
"grad_norm": 0.08131373994961688,
|
| 2193 |
+
"learning_rate": 5.92111075459616e-05,
|
| 2194 |
+
"loss": 0.3175,
|
| 2195 |
+
"step": 310
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 2.01294498381877,
|
| 2199 |
+
"grad_norm": 0.08591014053929732,
|
| 2200 |
+
"learning_rate": 5.852110860201294e-05,
|
| 2201 |
+
"loss": 0.3109,
|
| 2202 |
+
"step": 311
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 2.0194174757281553,
|
| 2206 |
+
"grad_norm": 0.08486358862414822,
|
| 2207 |
+
"learning_rate": 5.7833486654981606e-05,
|
| 2208 |
+
"loss": 0.3056,
|
| 2209 |
+
"step": 312
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 2.0258899676375406,
|
| 2213 |
+
"grad_norm": 0.08216589197196723,
|
| 2214 |
+
"learning_rate": 5.714828110985635e-05,
|
| 2215 |
+
"loss": 0.308,
|
| 2216 |
+
"step": 313
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 2.0323624595469254,
|
| 2220 |
+
"grad_norm": 0.08472956112082734,
|
| 2221 |
+
"learning_rate": 5.6465531233151126e-05,
|
| 2222 |
+
"loss": 0.3063,
|
| 2223 |
+
"step": 314
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 2.0388349514563107,
|
| 2227 |
+
"grad_norm": 0.08715839296976118,
|
| 2228 |
+
"learning_rate": 5.578527615065492e-05,
|
| 2229 |
+
"loss": 0.3146,
|
| 2230 |
+
"step": 315
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 2.045307443365696,
|
| 2234 |
+
"grad_norm": 0.08426241524579575,
|
| 2235 |
+
"learning_rate": 5.510755484518955e-05,
|
| 2236 |
+
"loss": 0.305,
|
| 2237 |
+
"step": 316
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 2.0517799352750807,
|
| 2241 |
+
"grad_norm": 0.08998957963984325,
|
| 2242 |
+
"learning_rate": 5.443240615437586e-05,
|
| 2243 |
+
"loss": 0.3082,
|
| 2244 |
+
"step": 317
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 2.058252427184466,
|
| 2248 |
+
"grad_norm": 0.09157974577803293,
|
| 2249 |
+
"learning_rate": 5.375986876840784e-05,
|
| 2250 |
+
"loss": 0.3187,
|
| 2251 |
+
"step": 318
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 2.0647249190938513,
|
| 2255 |
+
"grad_norm": 0.09110362940023115,
|
| 2256 |
+
"learning_rate": 5.30899812278356e-05,
|
| 2257 |
+
"loss": 0.3089,
|
| 2258 |
+
"step": 319
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 2.071197411003236,
|
| 2262 |
+
"grad_norm": 0.08924632615443999,
|
| 2263 |
+
"learning_rate": 5.2422781921356826e-05,
|
| 2264 |
+
"loss": 0.3108,
|
| 2265 |
+
"step": 320
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 2.0776699029126213,
|
| 2269 |
+
"grad_norm": 0.08653114959635026,
|
| 2270 |
+
"learning_rate": 5.1758309083616673e-05,
|
| 2271 |
+
"loss": 0.3033,
|
| 2272 |
+
"step": 321
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 2.0841423948220066,
|
| 2276 |
+
"grad_norm": 0.0897639664776455,
|
| 2277 |
+
"learning_rate": 5.109660079301668e-05,
|
| 2278 |
+
"loss": 0.3044,
|
| 2279 |
+
"step": 322
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 2.0906148867313914,
|
| 2283 |
+
"grad_norm": 0.08654171610929062,
|
| 2284 |
+
"learning_rate": 5.043769496953299e-05,
|
| 2285 |
+
"loss": 0.3033,
|
| 2286 |
+
"step": 323
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 2.0970873786407767,
|
| 2290 |
+
"grad_norm": 0.0883439805695338,
|
| 2291 |
+
"learning_rate": 4.9781629372542895e-05,
|
| 2292 |
+
"loss": 0.3081,
|
| 2293 |
+
"step": 324
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 2.103559870550162,
|
| 2297 |
+
"grad_norm": 0.08874070518690325,
|
| 2298 |
+
"learning_rate": 4.912844159866112e-05,
|
| 2299 |
+
"loss": 0.3029,
|
| 2300 |
+
"step": 325
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 2.1100323624595467,
|
| 2304 |
+
"grad_norm": 0.08970836183938685,
|
| 2305 |
+
"learning_rate": 4.847816907958549e-05,
|
| 2306 |
+
"loss": 0.3051,
|
| 2307 |
+
"step": 326
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 2.116504854368932,
|
| 2311 |
+
"grad_norm": 0.08702359256692323,
|
| 2312 |
+
"learning_rate": 4.783084907995156e-05,
|
| 2313 |
+
"loss": 0.3083,
|
| 2314 |
+
"step": 327
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 2.1229773462783172,
|
| 2318 |
+
"grad_norm": 0.08620252985570646,
|
| 2319 |
+
"learning_rate": 4.718651869519731e-05,
|
| 2320 |
+
"loss": 0.3009,
|
| 2321 |
+
"step": 328
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 2.129449838187702,
|
| 2325 |
+
"grad_norm": 0.09026535273366211,
|
| 2326 |
+
"learning_rate": 4.654521484943735e-05,
|
| 2327 |
+
"loss": 0.3095,
|
| 2328 |
+
"step": 329
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 2.1359223300970873,
|
| 2332 |
+
"grad_norm": 0.08797218673185163,
|
| 2333 |
+
"learning_rate": 4.59069742933468e-05,
|
| 2334 |
+
"loss": 0.3037,
|
| 2335 |
+
"step": 330
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 2.1423948220064726,
|
| 2339 |
+
"grad_norm": 0.08992206831884317,
|
| 2340 |
+
"learning_rate": 4.527183360205541e-05,
|
| 2341 |
+
"loss": 0.3058,
|
| 2342 |
+
"step": 331
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 2.148867313915858,
|
| 2346 |
+
"grad_norm": 0.09019684013585973,
|
| 2347 |
+
"learning_rate": 4.4639829173051554e-05,
|
| 2348 |
+
"loss": 0.3091,
|
| 2349 |
+
"step": 332
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 2.1553398058252426,
|
| 2353 |
+
"grad_norm": 0.08846989758042821,
|
| 2354 |
+
"learning_rate": 4.401099722409631e-05,
|
| 2355 |
+
"loss": 0.2968,
|
| 2356 |
+
"step": 333
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 2.161812297734628,
|
| 2360 |
+
"grad_norm": 0.09000600047149172,
|
| 2361 |
+
"learning_rate": 4.338537379114801e-05,
|
| 2362 |
+
"loss": 0.322,
|
| 2363 |
+
"step": 334
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 2.168284789644013,
|
| 2367 |
+
"grad_norm": 0.09084909315613375,
|
| 2368 |
+
"learning_rate": 4.2762994726297346e-05,
|
| 2369 |
+
"loss": 0.3104,
|
| 2370 |
+
"step": 335
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 2.174757281553398,
|
| 2374 |
+
"grad_norm": 0.08897581163092526,
|
| 2375 |
+
"learning_rate": 4.2143895695712444e-05,
|
| 2376 |
+
"loss": 0.3029,
|
| 2377 |
+
"step": 336
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 2.1812297734627832,
|
| 2381 |
+
"grad_norm": 0.08806995729841001,
|
| 2382 |
+
"learning_rate": 4.152811217759529e-05,
|
| 2383 |
+
"loss": 0.3127,
|
| 2384 |
+
"step": 337
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 2.1877022653721685,
|
| 2388 |
+
"grad_norm": 0.0909392143685305,
|
| 2389 |
+
"learning_rate": 4.091567946014858e-05,
|
| 2390 |
+
"loss": 0.3077,
|
| 2391 |
+
"step": 338
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 2.1941747572815533,
|
| 2395 |
+
"grad_norm": 0.08526830124486821,
|
| 2396 |
+
"learning_rate": 4.0306632639553323e-05,
|
| 2397 |
+
"loss": 0.3092,
|
| 2398 |
+
"step": 339
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 2.2006472491909386,
|
| 2402 |
+
"grad_norm": 0.0889661858323667,
|
| 2403 |
+
"learning_rate": 3.970100661795766e-05,
|
| 2404 |
+
"loss": 0.306,
|
| 2405 |
+
"step": 340
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.207119741100324,
|
| 2409 |
+
"grad_norm": 0.08742864887103546,
|
| 2410 |
+
"learning_rate": 3.909883610147696e-05,
|
| 2411 |
+
"loss": 0.3021,
|
| 2412 |
+
"step": 341
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 2.2135922330097086,
|
| 2416 |
+
"grad_norm": 0.08736016824506748,
|
| 2417 |
+
"learning_rate": 3.8500155598204644e-05,
|
| 2418 |
+
"loss": 0.2974,
|
| 2419 |
+
"step": 342
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 2.220064724919094,
|
| 2423 |
+
"grad_norm": 0.0895276470969271,
|
| 2424 |
+
"learning_rate": 3.7904999416234864e-05,
|
| 2425 |
+
"loss": 0.3121,
|
| 2426 |
+
"step": 343
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 2.226537216828479,
|
| 2430 |
+
"grad_norm": 0.08707410830498415,
|
| 2431 |
+
"learning_rate": 3.731340166169635e-05,
|
| 2432 |
+
"loss": 0.304,
|
| 2433 |
+
"step": 344
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 2.233009708737864,
|
| 2437 |
+
"grad_norm": 0.09084514125957084,
|
| 2438 |
+
"learning_rate": 3.6725396236797935e-05,
|
| 2439 |
+
"loss": 0.3139,
|
| 2440 |
+
"step": 345
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 2.2394822006472492,
|
| 2444 |
+
"grad_norm": 0.08595616841814938,
|
| 2445 |
+
"learning_rate": 3.614101683788575e-05,
|
| 2446 |
+
"loss": 0.3034,
|
| 2447 |
+
"step": 346
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.2459546925566345,
|
| 2451 |
+
"grad_norm": 0.08918897946715092,
|
| 2452 |
+
"learning_rate": 3.5560296953512295e-05,
|
| 2453 |
+
"loss": 0.2871,
|
| 2454 |
+
"step": 347
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 2.2524271844660193,
|
| 2458 |
+
"grad_norm": 0.08951525733240658,
|
| 2459 |
+
"learning_rate": 3.498326986251717e-05,
|
| 2460 |
+
"loss": 0.3064,
|
| 2461 |
+
"step": 348
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 2.2588996763754046,
|
| 2465 |
+
"grad_norm": 0.08959019506927517,
|
| 2466 |
+
"learning_rate": 3.4409968632120126e-05,
|
| 2467 |
+
"loss": 0.3062,
|
| 2468 |
+
"step": 349
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 2.26537216828479,
|
| 2472 |
+
"grad_norm": 0.09072525944669337,
|
| 2473 |
+
"learning_rate": 3.3840426116026044e-05,
|
| 2474 |
+
"loss": 0.3113,
|
| 2475 |
+
"step": 350
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 2.2718446601941746,
|
| 2479 |
+
"grad_norm": 0.09370535186596111,
|
| 2480 |
+
"learning_rate": 3.327467495254225e-05,
|
| 2481 |
+
"loss": 0.3195,
|
| 2482 |
+
"step": 351
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 2.27831715210356,
|
| 2486 |
+
"grad_norm": 0.08681410960013289,
|
| 2487 |
+
"learning_rate": 3.2712747562708115e-05,
|
| 2488 |
+
"loss": 0.3039,
|
| 2489 |
+
"step": 352
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 2.284789644012945,
|
| 2493 |
+
"grad_norm": 0.0884544764695178,
|
| 2494 |
+
"learning_rate": 3.215467614843719e-05,
|
| 2495 |
+
"loss": 0.3028,
|
| 2496 |
+
"step": 353
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 2.29126213592233,
|
| 2500 |
+
"grad_norm": 0.08974826283064652,
|
| 2501 |
+
"learning_rate": 3.160049269067174e-05,
|
| 2502 |
+
"loss": 0.3073,
|
| 2503 |
+
"step": 354
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 2.2977346278317152,
|
| 2507 |
+
"grad_norm": 0.08791055069767624,
|
| 2508 |
+
"learning_rate": 3.105022894755003e-05,
|
| 2509 |
+
"loss": 0.3011,
|
| 2510 |
+
"step": 355
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 2.3042071197411005,
|
| 2514 |
+
"grad_norm": 0.08951734113600135,
|
| 2515 |
+
"learning_rate": 3.0503916452586612e-05,
|
| 2516 |
+
"loss": 0.3056,
|
| 2517 |
+
"step": 356
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 2.3106796116504853,
|
| 2521 |
+
"grad_norm": 0.08762327614071365,
|
| 2522 |
+
"learning_rate": 2.9961586512864947e-05,
|
| 2523 |
+
"loss": 0.3138,
|
| 2524 |
+
"step": 357
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 2.3171521035598706,
|
| 2528 |
+
"grad_norm": 0.09032122428603973,
|
| 2529 |
+
"learning_rate": 2.9423270207243437e-05,
|
| 2530 |
+
"loss": 0.2978,
|
| 2531 |
+
"step": 358
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 2.323624595469256,
|
| 2535 |
+
"grad_norm": 0.0917821868399838,
|
| 2536 |
+
"learning_rate": 2.888899838457455e-05,
|
| 2537 |
+
"loss": 0.3084,
|
| 2538 |
+
"step": 359
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 2.3300970873786406,
|
| 2542 |
+
"grad_norm": 0.09193242261010433,
|
| 2543 |
+
"learning_rate": 2.835880166193683e-05,
|
| 2544 |
+
"loss": 0.3064,
|
| 2545 |
+
"step": 360
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 2.336569579288026,
|
| 2549 |
+
"grad_norm": 0.08883763145021951,
|
| 2550 |
+
"learning_rate": 2.7832710422880328e-05,
|
| 2551 |
+
"loss": 0.2993,
|
| 2552 |
+
"step": 361
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 2.343042071197411,
|
| 2556 |
+
"grad_norm": 0.08991962234888566,
|
| 2557 |
+
"learning_rate": 2.7310754815685624e-05,
|
| 2558 |
+
"loss": 0.3034,
|
| 2559 |
+
"step": 362
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 2.349514563106796,
|
| 2563 |
+
"grad_norm": 0.08994982579652418,
|
| 2564 |
+
"learning_rate": 2.679296475163595e-05,
|
| 2565 |
+
"loss": 0.3072,
|
| 2566 |
+
"step": 363
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 2.355987055016181,
|
| 2570 |
+
"grad_norm": 0.08970808859923911,
|
| 2571 |
+
"learning_rate": 2.6279369903303175e-05,
|
| 2572 |
+
"loss": 0.3023,
|
| 2573 |
+
"step": 364
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 2.3624595469255665,
|
| 2577 |
+
"grad_norm": 0.08588546055096836,
|
| 2578 |
+
"learning_rate": 2.5769999702847346e-05,
|
| 2579 |
+
"loss": 0.296,
|
| 2580 |
+
"step": 365
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 2.3689320388349513,
|
| 2584 |
+
"grad_norm": 0.08990846870006015,
|
| 2585 |
+
"learning_rate": 2.5264883340330113e-05,
|
| 2586 |
+
"loss": 0.3022,
|
| 2587 |
+
"step": 366
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 2.3754045307443366,
|
| 2591 |
+
"grad_norm": 0.09151580193749502,
|
| 2592 |
+
"learning_rate": 2.4764049762041874e-05,
|
| 2593 |
+
"loss": 0.2978,
|
| 2594 |
+
"step": 367
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 2.381877022653722,
|
| 2598 |
+
"grad_norm": 0.08959471455140089,
|
| 2599 |
+
"learning_rate": 2.426752766884306e-05,
|
| 2600 |
+
"loss": 0.3099,
|
| 2601 |
+
"step": 368
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 2.3883495145631066,
|
| 2605 |
+
"grad_norm": 0.0887364245299765,
|
| 2606 |
+
"learning_rate": 2.377534551451932e-05,
|
| 2607 |
+
"loss": 0.3016,
|
| 2608 |
+
"step": 369
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 2.394822006472492,
|
| 2612 |
+
"grad_norm": 0.08882940905923181,
|
| 2613 |
+
"learning_rate": 2.328753150415094e-05,
|
| 2614 |
+
"loss": 0.3004,
|
| 2615 |
+
"step": 370
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 2.401294498381877,
|
| 2619 |
+
"grad_norm": 0.08843646634108779,
|
| 2620 |
+
"learning_rate": 2.280411359249668e-05,
|
| 2621 |
+
"loss": 0.2933,
|
| 2622 |
+
"step": 371
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 2.407766990291262,
|
| 2626 |
+
"grad_norm": 0.09301769304474337,
|
| 2627 |
+
"learning_rate": 2.2325119482391467e-05,
|
| 2628 |
+
"loss": 0.3134,
|
| 2629 |
+
"step": 372
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 2.414239482200647,
|
| 2633 |
+
"grad_norm": 0.08707266390112682,
|
| 2634 |
+
"learning_rate": 2.185057662315918e-05,
|
| 2635 |
+
"loss": 0.3003,
|
| 2636 |
+
"step": 373
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 2.4207119741100325,
|
| 2640 |
+
"grad_norm": 0.0868807473282313,
|
| 2641 |
+
"learning_rate": 2.1380512209039528e-05,
|
| 2642 |
+
"loss": 0.302,
|
| 2643 |
+
"step": 374
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 2.4271844660194173,
|
| 2647 |
+
"grad_norm": 0.09363452018708772,
|
| 2648 |
+
"learning_rate": 2.0914953177629548e-05,
|
| 2649 |
+
"loss": 0.3148,
|
| 2650 |
+
"step": 375
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 2.4336569579288025,
|
| 2654 |
+
"grad_norm": 0.08938831583372829,
|
| 2655 |
+
"learning_rate": 2.0453926208340003e-05,
|
| 2656 |
+
"loss": 0.3055,
|
| 2657 |
+
"step": 376
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 2.440129449838188,
|
| 2661 |
+
"grad_norm": 0.09230125425366141,
|
| 2662 |
+
"learning_rate": 1.999745772086655e-05,
|
| 2663 |
+
"loss": 0.3041,
|
| 2664 |
+
"step": 377
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 2.4466019417475726,
|
| 2668 |
+
"grad_norm": 0.08932842130151852,
|
| 2669 |
+
"learning_rate": 1.954557387367557e-05,
|
| 2670 |
+
"loss": 0.3037,
|
| 2671 |
+
"step": 378
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 2.453074433656958,
|
| 2675 |
+
"grad_norm": 0.08923122998718774,
|
| 2676 |
+
"learning_rate": 1.9098300562505266e-05,
|
| 2677 |
+
"loss": 0.3055,
|
| 2678 |
+
"step": 379
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 2.459546925566343,
|
| 2682 |
+
"grad_norm": 0.0917592218140953,
|
| 2683 |
+
"learning_rate": 1.8655663418881584e-05,
|
| 2684 |
+
"loss": 0.3006,
|
| 2685 |
+
"step": 380
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 2.466019417475728,
|
| 2689 |
+
"grad_norm": 0.08856315870273591,
|
| 2690 |
+
"learning_rate": 1.821768780864943e-05,
|
| 2691 |
+
"loss": 0.3072,
|
| 2692 |
+
"step": 381
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 2.472491909385113,
|
| 2696 |
+
"grad_norm": 0.08867770851651259,
|
| 2697 |
+
"learning_rate": 1.7784398830519e-05,
|
| 2698 |
+
"loss": 0.298,
|
| 2699 |
+
"step": 382
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 2.4789644012944985,
|
| 2703 |
+
"grad_norm": 0.08895980547301872,
|
| 2704 |
+
"learning_rate": 1.7355821314627564e-05,
|
| 2705 |
+
"loss": 0.3087,
|
| 2706 |
+
"step": 383
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 2.4854368932038833,
|
| 2710 |
+
"grad_norm": 0.09119431484259279,
|
| 2711 |
+
"learning_rate": 1.6931979821116418e-05,
|
| 2712 |
+
"loss": 0.305,
|
| 2713 |
+
"step": 384
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 2.4919093851132685,
|
| 2717 |
+
"grad_norm": 0.09318125516492613,
|
| 2718 |
+
"learning_rate": 1.6512898638723497e-05,
|
| 2719 |
+
"loss": 0.3095,
|
| 2720 |
+
"step": 385
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 2.498381877022654,
|
| 2724 |
+
"grad_norm": 0.08770991666106868,
|
| 2725 |
+
"learning_rate": 1.6098601783391487e-05,
|
| 2726 |
+
"loss": 0.292,
|
| 2727 |
+
"step": 386
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 2.5048543689320386,
|
| 2731 |
+
"grad_norm": 0.08730384586077941,
|
| 2732 |
+
"learning_rate": 1.5689112996891576e-05,
|
| 2733 |
+
"loss": 0.2982,
|
| 2734 |
+
"step": 387
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 2.511326860841424,
|
| 2738 |
+
"grad_norm": 0.09025336858704619,
|
| 2739 |
+
"learning_rate": 1.5284455745462834e-05,
|
| 2740 |
+
"loss": 0.3175,
|
| 2741 |
+
"step": 388
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 2.517799352750809,
|
| 2745 |
+
"grad_norm": 0.09170439944127616,
|
| 2746 |
+
"learning_rate": 1.4884653218467571e-05,
|
| 2747 |
+
"loss": 0.3131,
|
| 2748 |
+
"step": 389
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 2.524271844660194,
|
| 2752 |
+
"grad_norm": 0.08975472589683904,
|
| 2753 |
+
"learning_rate": 1.4489728327062324e-05,
|
| 2754 |
+
"loss": 0.3077,
|
| 2755 |
+
"step": 390
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 2.530744336569579,
|
| 2759 |
+
"grad_norm": 0.09151247517360593,
|
| 2760 |
+
"learning_rate": 1.4099703702884936e-05,
|
| 2761 |
+
"loss": 0.2987,
|
| 2762 |
+
"step": 391
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 2.5372168284789645,
|
| 2766 |
+
"grad_norm": 0.08707773026948659,
|
| 2767 |
+
"learning_rate": 1.3714601696757712e-05,
|
| 2768 |
+
"loss": 0.301,
|
| 2769 |
+
"step": 392
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 2.5436893203883493,
|
| 2773 |
+
"grad_norm": 0.09236291075818845,
|
| 2774 |
+
"learning_rate": 1.3334444377406452e-05,
|
| 2775 |
+
"loss": 0.3095,
|
| 2776 |
+
"step": 393
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 2.5501618122977345,
|
| 2780 |
+
"grad_norm": 0.09003080825117314,
|
| 2781 |
+
"learning_rate": 1.2959253530195836e-05,
|
| 2782 |
+
"loss": 0.301,
|
| 2783 |
+
"step": 394
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 2.55663430420712,
|
| 2787 |
+
"grad_norm": 0.08899294818724815,
|
| 2788 |
+
"learning_rate": 1.258905065588103e-05,
|
| 2789 |
+
"loss": 0.2996,
|
| 2790 |
+
"step": 395
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 2.5631067961165046,
|
| 2794 |
+
"grad_norm": 0.08642480774869472,
|
| 2795 |
+
"learning_rate": 1.2223856969375447e-05,
|
| 2796 |
+
"loss": 0.2972,
|
| 2797 |
+
"step": 396
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 2.56957928802589,
|
| 2801 |
+
"grad_norm": 0.09069465990264014,
|
| 2802 |
+
"learning_rate": 1.1863693398535114e-05,
|
| 2803 |
+
"loss": 0.3028,
|
| 2804 |
+
"step": 397
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 2.576051779935275,
|
| 2808 |
+
"grad_norm": 0.09007456341294559,
|
| 2809 |
+
"learning_rate": 1.1508580582959349e-05,
|
| 2810 |
+
"loss": 0.3026,
|
| 2811 |
+
"step": 398
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 2.58252427184466,
|
| 2815 |
+
"grad_norm": 0.0874176261621062,
|
| 2816 |
+
"learning_rate": 1.1158538872807933e-05,
|
| 2817 |
+
"loss": 0.302,
|
| 2818 |
+
"step": 399
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 2.588996763754045,
|
| 2822 |
+
"grad_norm": 0.08762089070699175,
|
| 2823 |
+
"learning_rate": 1.0813588327634961e-05,
|
| 2824 |
+
"loss": 0.2845,
|
| 2825 |
+
"step": 400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 2.5954692556634305,
|
| 2829 |
+
"grad_norm": 0.08875402737182131,
|
| 2830 |
+
"learning_rate": 1.0473748715239307e-05,
|
| 2831 |
+
"loss": 0.3016,
|
| 2832 |
+
"step": 401
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 2.6019417475728153,
|
| 2836 |
+
"grad_norm": 0.09129343356719948,
|
| 2837 |
+
"learning_rate": 1.01390395105318e-05,
|
| 2838 |
+
"loss": 0.3154,
|
| 2839 |
+
"step": 402
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 2.6084142394822005,
|
| 2843 |
+
"grad_norm": 0.08810297072389008,
|
| 2844 |
+
"learning_rate": 9.809479894419149e-06,
|
| 2845 |
+
"loss": 0.3033,
|
| 2846 |
+
"step": 403
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 2.614886731391586,
|
| 2850 |
+
"grad_norm": 0.09018340343849121,
|
| 2851 |
+
"learning_rate": 9.485088752704885e-06,
|
| 2852 |
+
"loss": 0.3088,
|
| 2853 |
+
"step": 404
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 2.6213592233009706,
|
| 2857 |
+
"grad_norm": 0.08968883260752107,
|
| 2858 |
+
"learning_rate": 9.16588467500693e-06,
|
| 2859 |
+
"loss": 0.2982,
|
| 2860 |
+
"step": 405
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"epoch": 2.627831715210356,
|
| 2864 |
+
"grad_norm": 0.08957666144874589,
|
| 2865 |
+
"learning_rate": 8.851885953692374e-06,
|
| 2866 |
+
"loss": 0.3082,
|
| 2867 |
+
"step": 406
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 2.634304207119741,
|
| 2871 |
+
"grad_norm": 0.09198641057091544,
|
| 2872 |
+
"learning_rate": 8.543110582829272e-06,
|
| 2873 |
+
"loss": 0.3015,
|
| 2874 |
+
"step": 407
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 2.6407766990291264,
|
| 2878 |
+
"grad_norm": 0.08915392372194866,
|
| 2879 |
+
"learning_rate": 8.239576257155334e-06,
|
| 2880 |
+
"loss": 0.3068,
|
| 2881 |
+
"step": 408
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 2.647249190938511,
|
| 2885 |
+
"grad_norm": 0.09135989217201736,
|
| 2886 |
+
"learning_rate": 7.941300371063953e-06,
|
| 2887 |
+
"loss": 0.3118,
|
| 2888 |
+
"step": 409
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 2.6537216828478964,
|
| 2892 |
+
"grad_norm": 0.09150707922550998,
|
| 2893 |
+
"learning_rate": 7.648300017607534e-06,
|
| 2894 |
+
"loss": 0.3049,
|
| 2895 |
+
"step": 410
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 2.6601941747572817,
|
| 2899 |
+
"grad_norm": 0.08679197622255307,
|
| 2900 |
+
"learning_rate": 7.360591987517762e-06,
|
| 2901 |
+
"loss": 0.2973,
|
| 2902 |
+
"step": 411
|
| 2903 |
+
},
|
| 2904 |
+
{
|
| 2905 |
+
"epoch": 2.6666666666666665,
|
| 2906 |
+
"grad_norm": 0.09043247718192007,
|
| 2907 |
+
"learning_rate": 7.078192768243486e-06,
|
| 2908 |
+
"loss": 0.3016,
|
| 2909 |
+
"step": 412
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 2.6731391585760518,
|
| 2913 |
+
"grad_norm": 0.09227814811133576,
|
| 2914 |
+
"learning_rate": 6.80111854300588e-06,
|
| 2915 |
+
"loss": 0.3109,
|
| 2916 |
+
"step": 413
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 2.679611650485437,
|
| 2920 |
+
"grad_norm": 0.08752395500418549,
|
| 2921 |
+
"learning_rate": 6.5293851898710625e-06,
|
| 2922 |
+
"loss": 0.301,
|
| 2923 |
+
"step": 414
|
| 2924 |
+
},
|
| 2925 |
+
{
|
| 2926 |
+
"epoch": 2.686084142394822,
|
| 2927 |
+
"grad_norm": 0.08959180387680037,
|
| 2928 |
+
"learning_rate": 6.2630082808401326e-06,
|
| 2929 |
+
"loss": 0.2992,
|
| 2930 |
+
"step": 415
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 2.692556634304207,
|
| 2934 |
+
"grad_norm": 0.08743171313620991,
|
| 2935 |
+
"learning_rate": 6.00200308095682e-06,
|
| 2936 |
+
"loss": 0.301,
|
| 2937 |
+
"step": 416
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 2.6990291262135924,
|
| 2941 |
+
"grad_norm": 0.08980865803761802,
|
| 2942 |
+
"learning_rate": 5.746384547432737e-06,
|
| 2943 |
+
"loss": 0.3174,
|
| 2944 |
+
"step": 417
|
| 2945 |
+
},
|
| 2946 |
+
{
|
| 2947 |
+
"epoch": 2.705501618122977,
|
| 2948 |
+
"grad_norm": 0.08661713786701486,
|
| 2949 |
+
"learning_rate": 5.496167328790191e-06,
|
| 2950 |
+
"loss": 0.2939,
|
| 2951 |
+
"step": 418
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 2.7119741100323624,
|
| 2955 |
+
"grad_norm": 0.08786309161255973,
|
| 2956 |
+
"learning_rate": 5.251365764022753e-06,
|
| 2957 |
+
"loss": 0.2978,
|
| 2958 |
+
"step": 419
|
| 2959 |
+
},
|
| 2960 |
+
{
|
| 2961 |
+
"epoch": 2.7184466019417477,
|
| 2962 |
+
"grad_norm": 0.08733419882477451,
|
| 2963 |
+
"learning_rate": 5.011993881773569e-06,
|
| 2964 |
+
"loss": 0.3002,
|
| 2965 |
+
"step": 420
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 2.724919093851133,
|
| 2969 |
+
"grad_norm": 0.08926182705644836,
|
| 2970 |
+
"learning_rate": 4.778065399531395e-06,
|
| 2971 |
+
"loss": 0.2975,
|
| 2972 |
+
"step": 421
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 2.7313915857605178,
|
| 2976 |
+
"grad_norm": 0.0894176370961602,
|
| 2977 |
+
"learning_rate": 4.549593722844492e-06,
|
| 2978 |
+
"loss": 0.3067,
|
| 2979 |
+
"step": 422
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 2.737864077669903,
|
| 2983 |
+
"grad_norm": 0.0876923455017712,
|
| 2984 |
+
"learning_rate": 4.326591944552438e-06,
|
| 2985 |
+
"loss": 0.3097,
|
| 2986 |
+
"step": 423
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 2.7443365695792883,
|
| 2990 |
+
"grad_norm": 0.08572253083705766,
|
| 2991 |
+
"learning_rate": 4.109072844035844e-06,
|
| 2992 |
+
"loss": 0.2911,
|
| 2993 |
+
"step": 424
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 2.750809061488673,
|
| 2997 |
+
"grad_norm": 0.08883901045408762,
|
| 2998 |
+
"learning_rate": 3.8970488864839334e-06,
|
| 2999 |
+
"loss": 0.2963,
|
| 3000 |
+
"step": 425
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 2.7572815533980584,
|
| 3004 |
+
"grad_norm": 0.09240186844770727,
|
| 3005 |
+
"learning_rate": 3.690532222180343e-06,
|
| 3006 |
+
"loss": 0.3024,
|
| 3007 |
+
"step": 426
|
| 3008 |
+
},
|
| 3009 |
+
{
|
| 3010 |
+
"epoch": 2.7637540453074436,
|
| 3011 |
+
"grad_norm": 0.09051220479105677,
|
| 3012 |
+
"learning_rate": 3.4895346858066724e-06,
|
| 3013 |
+
"loss": 0.3089,
|
| 3014 |
+
"step": 427
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 2.7702265372168284,
|
| 3018 |
+
"grad_norm": 0.0891289977534799,
|
| 3019 |
+
"learning_rate": 3.2940677957644215e-06,
|
| 3020 |
+
"loss": 0.3047,
|
| 3021 |
+
"step": 428
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 2.7766990291262137,
|
| 3025 |
+
"grad_norm": 0.08973918683676917,
|
| 3026 |
+
"learning_rate": 3.104142753514849e-06,
|
| 3027 |
+
"loss": 0.3151,
|
| 3028 |
+
"step": 429
|
| 3029 |
+
},
|
| 3030 |
+
{
|
| 3031 |
+
"epoch": 2.783171521035599,
|
| 3032 |
+
"grad_norm": 0.08839195406678332,
|
| 3033 |
+
"learning_rate": 2.9197704429370977e-06,
|
| 3034 |
+
"loss": 0.2995,
|
| 3035 |
+
"step": 430
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 2.7896440129449838,
|
| 3039 |
+
"grad_norm": 0.08711643702101148,
|
| 3040 |
+
"learning_rate": 2.7409614297043806e-06,
|
| 3041 |
+
"loss": 0.2983,
|
| 3042 |
+
"step": 431
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 2.796116504854369,
|
| 3046 |
+
"grad_norm": 0.08827751306740735,
|
| 3047 |
+
"learning_rate": 2.5677259606786684e-06,
|
| 3048 |
+
"loss": 0.3071,
|
| 3049 |
+
"step": 432
|
| 3050 |
+
},
|
| 3051 |
+
{
|
| 3052 |
+
"epoch": 2.8025889967637543,
|
| 3053 |
+
"grad_norm": 0.08989812034868257,
|
| 3054 |
+
"learning_rate": 2.4000739633233347e-06,
|
| 3055 |
+
"loss": 0.3185,
|
| 3056 |
+
"step": 433
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 2.809061488673139,
|
| 3060 |
+
"grad_norm": 0.08814993018203572,
|
| 3061 |
+
"learning_rate": 2.238015045134334e-06,
|
| 3062 |
+
"loss": 0.303,
|
| 3063 |
+
"step": 434
|
| 3064 |
+
},
|
| 3065 |
+
{
|
| 3066 |
+
"epoch": 2.8155339805825244,
|
| 3067 |
+
"grad_norm": 0.08731684328369535,
|
| 3068 |
+
"learning_rate": 2.0815584930895972e-06,
|
| 3069 |
+
"loss": 0.3053,
|
| 3070 |
+
"step": 435
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"epoch": 2.8220064724919096,
|
| 3074 |
+
"grad_norm": 0.09248953078205228,
|
| 3075 |
+
"learning_rate": 1.9307132731168352e-06,
|
| 3076 |
+
"loss": 0.3048,
|
| 3077 |
+
"step": 436
|
| 3078 |
+
},
|
| 3079 |
+
{
|
| 3080 |
+
"epoch": 2.8284789644012944,
|
| 3081 |
+
"grad_norm": 0.08979246033823442,
|
| 3082 |
+
"learning_rate": 1.7854880295797405e-06,
|
| 3083 |
+
"loss": 0.3046,
|
| 3084 |
+
"step": 437
|
| 3085 |
+
},
|
| 3086 |
+
{
|
| 3087 |
+
"epoch": 2.8349514563106797,
|
| 3088 |
+
"grad_norm": 0.08615293987508164,
|
| 3089 |
+
"learning_rate": 1.6458910847826026e-06,
|
| 3090 |
+
"loss": 0.2991,
|
| 3091 |
+
"step": 438
|
| 3092 |
+
},
|
| 3093 |
+
{
|
| 3094 |
+
"epoch": 2.841423948220065,
|
| 3095 |
+
"grad_norm": 0.08795133012526277,
|
| 3096 |
+
"learning_rate": 1.5119304384934252e-06,
|
| 3097 |
+
"loss": 0.2922,
|
| 3098 |
+
"step": 439
|
| 3099 |
+
},
|
| 3100 |
+
{
|
| 3101 |
+
"epoch": 2.8478964401294498,
|
| 3102 |
+
"grad_norm": 0.0862166031371393,
|
| 3103 |
+
"learning_rate": 1.3836137674854255e-06,
|
| 3104 |
+
"loss": 0.2921,
|
| 3105 |
+
"step": 440
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 2.854368932038835,
|
| 3109 |
+
"grad_norm": 0.08955236492173343,
|
| 3110 |
+
"learning_rate": 1.2609484250971749e-06,
|
| 3111 |
+
"loss": 0.2997,
|
| 3112 |
+
"step": 441
|
| 3113 |
+
},
|
| 3114 |
+
{
|
| 3115 |
+
"epoch": 2.8608414239482203,
|
| 3116 |
+
"grad_norm": 0.08843363251406085,
|
| 3117 |
+
"learning_rate": 1.143941440811147e-06,
|
| 3118 |
+
"loss": 0.3088,
|
| 3119 |
+
"step": 442
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 2.867313915857605,
|
| 3123 |
+
"grad_norm": 0.09009839385408548,
|
| 3124 |
+
"learning_rate": 1.0325995198509409e-06,
|
| 3125 |
+
"loss": 0.3109,
|
| 3126 |
+
"step": 443
|
| 3127 |
+
},
|
| 3128 |
+
{
|
| 3129 |
+
"epoch": 2.8737864077669903,
|
| 3130 |
+
"grad_norm": 0.08785361574469933,
|
| 3131 |
+
"learning_rate": 9.269290427969868e-07,
|
| 3132 |
+
"loss": 0.2971,
|
| 3133 |
+
"step": 444
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 2.8802588996763756,
|
| 3137 |
+
"grad_norm": 0.08841485850159143,
|
| 3138 |
+
"learning_rate": 8.26936065220929e-07,
|
| 3139 |
+
"loss": 0.3035,
|
| 3140 |
+
"step": 445
|
| 3141 |
+
},
|
| 3142 |
+
{
|
| 3143 |
+
"epoch": 2.8867313915857604,
|
| 3144 |
+
"grad_norm": 0.08779257157327024,
|
| 3145 |
+
"learning_rate": 7.326263173385584e-07,
|
| 3146 |
+
"loss": 0.3013,
|
| 3147 |
+
"step": 446
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 2.8932038834951457,
|
| 3151 |
+
"grad_norm": 0.0883658852563661,
|
| 3152 |
+
"learning_rate": 6.440052036815081e-07,
|
| 3153 |
+
"loss": 0.305,
|
| 3154 |
+
"step": 447
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 2.899676375404531,
|
| 3158 |
+
"grad_norm": 0.08862179969635353,
|
| 3159 |
+
"learning_rate": 5.610778027874908e-07,
|
| 3160 |
+
"loss": 0.2952,
|
| 3161 |
+
"step": 448
|
| 3162 |
+
},
|
| 3163 |
+
{
|
| 3164 |
+
"epoch": 2.9061488673139158,
|
| 3165 |
+
"grad_norm": 0.08863424887187872,
|
| 3166 |
+
"learning_rate": 4.838488669092534e-07,
|
| 3167 |
+
"loss": 0.2956,
|
| 3168 |
+
"step": 449
|
| 3169 |
+
},
|
| 3170 |
+
{
|
| 3171 |
+
"epoch": 2.912621359223301,
|
| 3172 |
+
"grad_norm": 0.08998809785580231,
|
| 3173 |
+
"learning_rate": 4.123228217422948e-07,
|
| 3174 |
+
"loss": 0.3015,
|
| 3175 |
+
"step": 450
|
| 3176 |
+
},
|
| 3177 |
+
{
|
| 3178 |
+
"epoch": 2.9190938511326863,
|
| 3179 |
+
"grad_norm": 0.08707387881008244,
|
| 3180 |
+
"learning_rate": 3.465037661712134e-07,
|
| 3181 |
+
"loss": 0.3095,
|
| 3182 |
+
"step": 451
|
| 3183 |
+
},
|
| 3184 |
+
{
|
| 3185 |
+
"epoch": 2.925566343042071,
|
| 3186 |
+
"grad_norm": 0.08740571796909039,
|
| 3187 |
+
"learning_rate": 2.86395472034795e-07,
|
| 3188 |
+
"loss": 0.3033,
|
| 3189 |
+
"step": 452
|
| 3190 |
+
},
|
| 3191 |
+
{
|
| 3192 |
+
"epoch": 2.9320388349514563,
|
| 3193 |
+
"grad_norm": 0.08599743041374118,
|
| 3194 |
+
"learning_rate": 2.3200138390993e-07,
|
| 3195 |
+
"loss": 0.2904,
|
| 3196 |
+
"step": 453
|
| 3197 |
+
},
|
| 3198 |
+
{
|
| 3199 |
+
"epoch": 2.9385113268608416,
|
| 3200 |
+
"grad_norm": 0.08726631646259621,
|
| 3201 |
+
"learning_rate": 1.83324618914138e-07,
|
| 3202 |
+
"loss": 0.3007,
|
| 3203 |
+
"step": 454
|
| 3204 |
+
},
|
| 3205 |
+
{
|
| 3206 |
+
"epoch": 2.9449838187702264,
|
| 3207 |
+
"grad_norm": 0.08898358262047706,
|
| 3208 |
+
"learning_rate": 1.4036796652701078e-07,
|
| 3209 |
+
"loss": 0.3017,
|
| 3210 |
+
"step": 455
|
| 3211 |
+
},
|
| 3212 |
+
{
|
| 3213 |
+
"epoch": 2.9514563106796117,
|
| 3214 |
+
"grad_norm": 0.08765396985812648,
|
| 3215 |
+
"learning_rate": 1.031338884302846e-07,
|
| 3216 |
+
"loss": 0.2887,
|
| 3217 |
+
"step": 456
|
| 3218 |
+
},
|
| 3219 |
+
{
|
| 3220 |
+
"epoch": 2.957928802588997,
|
| 3221 |
+
"grad_norm": 0.09001756499046232,
|
| 3222 |
+
"learning_rate": 7.162451836685291e-08,
|
| 3223 |
+
"loss": 0.309,
|
| 3224 |
+
"step": 457
|
| 3225 |
+
},
|
| 3226 |
+
{
|
| 3227 |
+
"epoch": 2.9644012944983817,
|
| 3228 |
+
"grad_norm": 0.09045952221990206,
|
| 3229 |
+
"learning_rate": 4.584166201841988e-08,
|
| 3230 |
+
"loss": 0.2996,
|
| 3231 |
+
"step": 458
|
| 3232 |
+
},
|
| 3233 |
+
{
|
| 3234 |
+
"epoch": 2.970873786407767,
|
| 3235 |
+
"grad_norm": 0.08709990251871128,
|
| 3236 |
+
"learning_rate": 2.578679690204977e-08,
|
| 3237 |
+
"loss": 0.2982,
|
| 3238 |
+
"step": 459
|
| 3239 |
+
},
|
| 3240 |
+
{
|
| 3241 |
+
"epoch": 2.9773462783171523,
|
| 3242 |
+
"grad_norm": 0.08844965304987025,
|
| 3243 |
+
"learning_rate": 1.1461072285490204e-08,
|
| 3244 |
+
"loss": 0.3136,
|
| 3245 |
+
"step": 460
|
| 3246 |
+
},
|
| 3247 |
+
{
|
| 3248 |
+
"epoch": 2.983818770226537,
|
| 3249 |
+
"grad_norm": 0.0892220185111147,
|
| 3250 |
+
"learning_rate": 2.865309121358184e-09,
|
| 3251 |
+
"loss": 0.3011,
|
| 3252 |
+
"step": 461
|
| 3253 |
+
},
|
| 3254 |
+
{
|
| 3255 |
+
"epoch": 2.9902912621359223,
|
| 3256 |
+
"grad_norm": 0.08562579597167659,
|
| 3257 |
+
"learning_rate": 0.0,
|
| 3258 |
+
"loss": 0.2951,
|
| 3259 |
+
"step": 462
|
| 3260 |
+
},
|
| 3261 |
+
{
|
| 3262 |
+
"epoch": 2.9902912621359223,
|
| 3263 |
+
"eval_loss": 0.3344117999076843,
|
| 3264 |
+
"eval_runtime": 35.3894,
|
| 3265 |
+
"eval_samples_per_second": 29.359,
|
| 3266 |
+
"eval_steps_per_second": 0.932,
|
| 3267 |
+
"step": 462
|
| 3268 |
+
},
|
| 3269 |
+
{
|
| 3270 |
+
"epoch": 2.9902912621359223,
|
| 3271 |
+
"step": 462,
|
| 3272 |
+
"total_flos": 1.8679451641472614e+17,
|
| 3273 |
+
"train_loss": 0.36376328850205325,
|
| 3274 |
+
"train_runtime": 5739.3163,
|
| 3275 |
+
"train_samples_per_second": 10.311,
|
| 3276 |
+
"train_steps_per_second": 0.08
|
| 3277 |
+
}
|
| 3278 |
+
],
|
| 3279 |
+
"logging_steps": 1,
|
| 3280 |
+
"max_steps": 462,
|
| 3281 |
+
"num_input_tokens_seen": 0,
|
| 3282 |
+
"num_train_epochs": 3,
|
| 3283 |
+
"save_steps": 500,
|
| 3284 |
+
"stateful_callbacks": {
|
| 3285 |
+
"TrainerControl": {
|
| 3286 |
+
"args": {
|
| 3287 |
+
"should_epoch_stop": false,
|
| 3288 |
+
"should_evaluate": false,
|
| 3289 |
+
"should_log": false,
|
| 3290 |
+
"should_save": true,
|
| 3291 |
+
"should_training_stop": true
|
| 3292 |
+
},
|
| 3293 |
+
"attributes": {}
|
| 3294 |
+
}
|
| 3295 |
+
},
|
| 3296 |
+
"total_flos": 1.8679451641472614e+17,
|
| 3297 |
+
"train_batch_size": 8,
|
| 3298 |
+
"trial_name": null,
|
| 3299 |
+
"trial_params": null
|
| 3300 |
+
}
|