Model save
Browse files- README.md +60 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- train_results.json +9 -0
- trainer_state.json +3430 -0
README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: mistralai/Mistral-7B-Instruct-v0.2
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- sft
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: preference_p0.1_seed42_level2_rarecleanbatch16
|
| 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 |
+
# preference_p0.1_seed42_level2_rarecleanbatch16
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) on the None dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 1e-05
|
| 39 |
+
- train_batch_size: 4
|
| 40 |
+
- eval_batch_size: 1
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 4
|
| 44 |
+
- total_train_batch_size: 16
|
| 45 |
+
- total_eval_batch_size: 4
|
| 46 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
+
- lr_scheduler_type: cosine
|
| 48 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
+
- num_epochs: 1
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
### Framework versions
|
| 56 |
+
|
| 57 |
+
- Transformers 4.44.2
|
| 58 |
+
- Pytorch 2.3.1+cu121
|
| 59 |
+
- Datasets 2.19.1
|
| 60 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 252721244405760.0,
|
| 4 |
+
"train_loss": 0.5161569981705382,
|
| 5 |
+
"train_runtime": 26044.7738,
|
| 6 |
+
"train_samples": 98881,
|
| 7 |
+
"train_samples_per_second": 1.483,
|
| 8 |
+
"train_steps_per_second": 0.093
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.44.2"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 252721244405760.0,
|
| 4 |
+
"train_loss": 0.5161569981705382,
|
| 5 |
+
"train_runtime": 26044.7738,
|
| 6 |
+
"train_samples": 98881,
|
| 7 |
+
"train_samples_per_second": 1.483,
|
| 8 |
+
"train_steps_per_second": 0.093
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3430 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2414,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.00041425020712510354,
|
| 13 |
+
"grad_norm": 24.495001410392245,
|
| 14 |
+
"learning_rate": 4.132231404958678e-08,
|
| 15 |
+
"loss": 1.4066,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.002071251035625518,
|
| 20 |
+
"grad_norm": 24.278159346480546,
|
| 21 |
+
"learning_rate": 2.066115702479339e-07,
|
| 22 |
+
"loss": 1.4104,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.004142502071251036,
|
| 27 |
+
"grad_norm": 15.207256152484726,
|
| 28 |
+
"learning_rate": 4.132231404958678e-07,
|
| 29 |
+
"loss": 1.3724,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.006213753106876553,
|
| 34 |
+
"grad_norm": 8.665623744138836,
|
| 35 |
+
"learning_rate": 6.198347107438018e-07,
|
| 36 |
+
"loss": 1.2508,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.008285004142502071,
|
| 41 |
+
"grad_norm": 11.05134062750248,
|
| 42 |
+
"learning_rate": 8.264462809917356e-07,
|
| 43 |
+
"loss": 1.1327,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.010356255178127589,
|
| 48 |
+
"grad_norm": 4.379402402170414,
|
| 49 |
+
"learning_rate": 1.0330578512396695e-06,
|
| 50 |
+
"loss": 1.0227,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.012427506213753107,
|
| 55 |
+
"grad_norm": 3.3489843705809355,
|
| 56 |
+
"learning_rate": 1.2396694214876035e-06,
|
| 57 |
+
"loss": 0.9669,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.014498757249378625,
|
| 62 |
+
"grad_norm": 3.3543061730911634,
|
| 63 |
+
"learning_rate": 1.4462809917355372e-06,
|
| 64 |
+
"loss": 0.9568,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.016570008285004142,
|
| 69 |
+
"grad_norm": 3.0246929252712045,
|
| 70 |
+
"learning_rate": 1.6528925619834712e-06,
|
| 71 |
+
"loss": 0.9285,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.018641259320629662,
|
| 76 |
+
"grad_norm": 3.0205344615509735,
|
| 77 |
+
"learning_rate": 1.859504132231405e-06,
|
| 78 |
+
"loss": 0.9156,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.020712510356255178,
|
| 83 |
+
"grad_norm": 3.2104230970271694,
|
| 84 |
+
"learning_rate": 2.066115702479339e-06,
|
| 85 |
+
"loss": 0.9066,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.022783761391880698,
|
| 90 |
+
"grad_norm": 3.0769403333765366,
|
| 91 |
+
"learning_rate": 2.2727272727272728e-06,
|
| 92 |
+
"loss": 0.9123,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.024855012427506214,
|
| 97 |
+
"grad_norm": 3.1192139331466073,
|
| 98 |
+
"learning_rate": 2.479338842975207e-06,
|
| 99 |
+
"loss": 0.8824,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.026926263463131733,
|
| 104 |
+
"grad_norm": 3.174458267006907,
|
| 105 |
+
"learning_rate": 2.6859504132231405e-06,
|
| 106 |
+
"loss": 0.8872,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.02899751449875725,
|
| 111 |
+
"grad_norm": 3.3704437475031614,
|
| 112 |
+
"learning_rate": 2.8925619834710743e-06,
|
| 113 |
+
"loss": 0.876,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.03106876553438277,
|
| 118 |
+
"grad_norm": 3.2746704156283384,
|
| 119 |
+
"learning_rate": 3.0991735537190086e-06,
|
| 120 |
+
"loss": 0.8687,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.033140016570008285,
|
| 125 |
+
"grad_norm": 3.324353070402648,
|
| 126 |
+
"learning_rate": 3.3057851239669424e-06,
|
| 127 |
+
"loss": 0.8715,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.035211267605633804,
|
| 132 |
+
"grad_norm": 2.9659428539001875,
|
| 133 |
+
"learning_rate": 3.5123966942148763e-06,
|
| 134 |
+
"loss": 0.8686,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.037282518641259324,
|
| 139 |
+
"grad_norm": 3.1234867961603885,
|
| 140 |
+
"learning_rate": 3.71900826446281e-06,
|
| 141 |
+
"loss": 0.8637,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.03935376967688484,
|
| 146 |
+
"grad_norm": 3.1123775472434705,
|
| 147 |
+
"learning_rate": 3.925619834710744e-06,
|
| 148 |
+
"loss": 0.8553,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.041425020712510356,
|
| 153 |
+
"grad_norm": 3.115177689829066,
|
| 154 |
+
"learning_rate": 4.132231404958678e-06,
|
| 155 |
+
"loss": 0.8695,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.043496271748135876,
|
| 160 |
+
"grad_norm": 3.0340671696442714,
|
| 161 |
+
"learning_rate": 4.338842975206612e-06,
|
| 162 |
+
"loss": 0.8712,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.045567522783761395,
|
| 167 |
+
"grad_norm": 3.073424358865733,
|
| 168 |
+
"learning_rate": 4.5454545454545455e-06,
|
| 169 |
+
"loss": 0.8353,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.04763877381938691,
|
| 174 |
+
"grad_norm": 2.9777104137179227,
|
| 175 |
+
"learning_rate": 4.75206611570248e-06,
|
| 176 |
+
"loss": 0.8473,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.04971002485501243,
|
| 181 |
+
"grad_norm": 3.034579402136387,
|
| 182 |
+
"learning_rate": 4.958677685950414e-06,
|
| 183 |
+
"loss": 0.8475,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.05178127589063795,
|
| 188 |
+
"grad_norm": 3.0813316554860704,
|
| 189 |
+
"learning_rate": 5.165289256198347e-06,
|
| 190 |
+
"loss": 0.8507,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.053852526926263466,
|
| 195 |
+
"grad_norm": 3.06671775904819,
|
| 196 |
+
"learning_rate": 5.371900826446281e-06,
|
| 197 |
+
"loss": 0.8463,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.05592377796188898,
|
| 202 |
+
"grad_norm": 3.2475108548122957,
|
| 203 |
+
"learning_rate": 5.578512396694216e-06,
|
| 204 |
+
"loss": 0.8335,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.0579950289975145,
|
| 209 |
+
"grad_norm": 3.05517684817537,
|
| 210 |
+
"learning_rate": 5.785123966942149e-06,
|
| 211 |
+
"loss": 0.8197,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.06006628003314002,
|
| 216 |
+
"grad_norm": 3.209092202417741,
|
| 217 |
+
"learning_rate": 5.991735537190083e-06,
|
| 218 |
+
"loss": 0.838,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.06213753106876554,
|
| 223 |
+
"grad_norm": 3.317353575296279,
|
| 224 |
+
"learning_rate": 6.198347107438017e-06,
|
| 225 |
+
"loss": 0.8424,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.06420878210439106,
|
| 230 |
+
"grad_norm": 3.50463295880707,
|
| 231 |
+
"learning_rate": 6.404958677685951e-06,
|
| 232 |
+
"loss": 0.833,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.06628003314001657,
|
| 237 |
+
"grad_norm": 2.966129268982567,
|
| 238 |
+
"learning_rate": 6.611570247933885e-06,
|
| 239 |
+
"loss": 0.8198,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.06835128417564208,
|
| 244 |
+
"grad_norm": 3.060604077046953,
|
| 245 |
+
"learning_rate": 6.818181818181818e-06,
|
| 246 |
+
"loss": 0.8123,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.07042253521126761,
|
| 251 |
+
"grad_norm": 3.106442872246759,
|
| 252 |
+
"learning_rate": 7.0247933884297525e-06,
|
| 253 |
+
"loss": 0.8256,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.07249378624689312,
|
| 258 |
+
"grad_norm": 3.2061341459353603,
|
| 259 |
+
"learning_rate": 7.231404958677687e-06,
|
| 260 |
+
"loss": 0.8272,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.07456503728251865,
|
| 265 |
+
"grad_norm": 3.261013201810741,
|
| 266 |
+
"learning_rate": 7.43801652892562e-06,
|
| 267 |
+
"loss": 0.8226,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.07663628831814416,
|
| 272 |
+
"grad_norm": 3.0418461202096387,
|
| 273 |
+
"learning_rate": 7.644628099173555e-06,
|
| 274 |
+
"loss": 0.8134,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.07870753935376967,
|
| 279 |
+
"grad_norm": 2.925954602214559,
|
| 280 |
+
"learning_rate": 7.851239669421489e-06,
|
| 281 |
+
"loss": 0.8106,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.0807787903893952,
|
| 286 |
+
"grad_norm": 3.096505735396263,
|
| 287 |
+
"learning_rate": 8.057851239669421e-06,
|
| 288 |
+
"loss": 0.8219,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.08285004142502071,
|
| 293 |
+
"grad_norm": 3.0178444567465883,
|
| 294 |
+
"learning_rate": 8.264462809917356e-06,
|
| 295 |
+
"loss": 0.8139,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.08492129246064622,
|
| 300 |
+
"grad_norm": 3.085074860786418,
|
| 301 |
+
"learning_rate": 8.47107438016529e-06,
|
| 302 |
+
"loss": 0.8079,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.08699254349627175,
|
| 307 |
+
"grad_norm": 2.9513280908493438,
|
| 308 |
+
"learning_rate": 8.677685950413224e-06,
|
| 309 |
+
"loss": 0.791,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.08906379453189726,
|
| 314 |
+
"grad_norm": 3.0685126689734843,
|
| 315 |
+
"learning_rate": 8.884297520661158e-06,
|
| 316 |
+
"loss": 0.8065,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.09113504556752279,
|
| 321 |
+
"grad_norm": 3.3564757893140067,
|
| 322 |
+
"learning_rate": 9.090909090909091e-06,
|
| 323 |
+
"loss": 0.817,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.0932062966031483,
|
| 328 |
+
"grad_norm": 3.144009095079128,
|
| 329 |
+
"learning_rate": 9.297520661157025e-06,
|
| 330 |
+
"loss": 0.799,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.09527754763877382,
|
| 335 |
+
"grad_norm": 2.9152439896159703,
|
| 336 |
+
"learning_rate": 9.50413223140496e-06,
|
| 337 |
+
"loss": 0.8123,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.09734879867439934,
|
| 342 |
+
"grad_norm": 2.959984059385148,
|
| 343 |
+
"learning_rate": 9.710743801652894e-06,
|
| 344 |
+
"loss": 0.7906,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.09942004971002485,
|
| 349 |
+
"grad_norm": 3.0028265600360067,
|
| 350 |
+
"learning_rate": 9.917355371900828e-06,
|
| 351 |
+
"loss": 0.8028,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.10149130074565037,
|
| 356 |
+
"grad_norm": 2.8289351291922693,
|
| 357 |
+
"learning_rate": 9.999952928077044e-06,
|
| 358 |
+
"loss": 0.7937,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.1035625517812759,
|
| 363 |
+
"grad_norm": 2.8859057720780013,
|
| 364 |
+
"learning_rate": 9.999665269535307e-06,
|
| 365 |
+
"loss": 0.7891,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.1056338028169014,
|
| 370 |
+
"grad_norm": 2.8213578917004445,
|
| 371 |
+
"learning_rate": 9.99911611854702e-06,
|
| 372 |
+
"loss": 0.7941,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.10770505385252693,
|
| 377 |
+
"grad_norm": 2.9934763093271486,
|
| 378 |
+
"learning_rate": 9.998305503833872e-06,
|
| 379 |
+
"loss": 0.8038,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.10977630488815245,
|
| 384 |
+
"grad_norm": 3.1184221834210617,
|
| 385 |
+
"learning_rate": 9.997233467792626e-06,
|
| 386 |
+
"loss": 0.7883,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.11184755592377796,
|
| 391 |
+
"grad_norm": 3.080255277284623,
|
| 392 |
+
"learning_rate": 9.995900066492902e-06,
|
| 393 |
+
"loss": 0.7908,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.11391880695940348,
|
| 398 |
+
"grad_norm": 2.961023177669383,
|
| 399 |
+
"learning_rate": 9.994305369674242e-06,
|
| 400 |
+
"loss": 0.8018,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.115990057995029,
|
| 405 |
+
"grad_norm": 3.2994763750087572,
|
| 406 |
+
"learning_rate": 9.992449460742464e-06,
|
| 407 |
+
"loss": 0.7929,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.11806130903065451,
|
| 412 |
+
"grad_norm": 3.118171539850542,
|
| 413 |
+
"learning_rate": 9.9903324367653e-06,
|
| 414 |
+
"loss": 0.7756,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.12013256006628004,
|
| 419 |
+
"grad_norm": 2.93509763542972,
|
| 420 |
+
"learning_rate": 9.98795440846732e-06,
|
| 421 |
+
"loss": 0.7893,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.12220381110190555,
|
| 426 |
+
"grad_norm": 3.005099209957883,
|
| 427 |
+
"learning_rate": 9.985315500224135e-06,
|
| 428 |
+
"loss": 0.7837,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.12427506213753108,
|
| 433 |
+
"grad_norm": 3.0065053233637262,
|
| 434 |
+
"learning_rate": 9.982415850055902e-06,
|
| 435 |
+
"loss": 0.775,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.1263463131731566,
|
| 440 |
+
"grad_norm": 2.8875884081237726,
|
| 441 |
+
"learning_rate": 9.979255609620095e-06,
|
| 442 |
+
"loss": 0.773,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.12841756420878211,
|
| 447 |
+
"grad_norm": 2.953711663180962,
|
| 448 |
+
"learning_rate": 9.975834944203581e-06,
|
| 449 |
+
"loss": 0.7833,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.1304888152444076,
|
| 454 |
+
"grad_norm": 3.0971637074544875,
|
| 455 |
+
"learning_rate": 9.972154032713973e-06,
|
| 456 |
+
"loss": 0.778,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.13256006628003314,
|
| 461 |
+
"grad_norm": 2.9856834222834334,
|
| 462 |
+
"learning_rate": 9.968213067670265e-06,
|
| 463 |
+
"loss": 0.7632,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.13463131731565867,
|
| 468 |
+
"grad_norm": 3.026025328201696,
|
| 469 |
+
"learning_rate": 9.964012255192776e-06,
|
| 470 |
+
"loss": 0.76,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.13670256835128416,
|
| 475 |
+
"grad_norm": 3.1660979738179336,
|
| 476 |
+
"learning_rate": 9.959551814992364e-06,
|
| 477 |
+
"loss": 0.7553,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.1387738193869097,
|
| 482 |
+
"grad_norm": 3.082731646367652,
|
| 483 |
+
"learning_rate": 9.954831980358928e-06,
|
| 484 |
+
"loss": 0.7724,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.14084507042253522,
|
| 489 |
+
"grad_norm": 2.808393207259557,
|
| 490 |
+
"learning_rate": 9.949852998149217e-06,
|
| 491 |
+
"loss": 0.7741,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.14291632145816072,
|
| 496 |
+
"grad_norm": 3.1777761122613013,
|
| 497 |
+
"learning_rate": 9.944615128773911e-06,
|
| 498 |
+
"loss": 0.7634,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.14498757249378624,
|
| 503 |
+
"grad_norm": 2.740280537246316,
|
| 504 |
+
"learning_rate": 9.939118646184007e-06,
|
| 505 |
+
"loss": 0.7518,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.14705882352941177,
|
| 510 |
+
"grad_norm": 2.796661144343301,
|
| 511 |
+
"learning_rate": 9.933363837856485e-06,
|
| 512 |
+
"loss": 0.7508,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.1491300745650373,
|
| 517 |
+
"grad_norm": 2.739682595325342,
|
| 518 |
+
"learning_rate": 9.927351004779275e-06,
|
| 519 |
+
"loss": 0.74,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.1512013256006628,
|
| 524 |
+
"grad_norm": 2.859085581091372,
|
| 525 |
+
"learning_rate": 9.921080461435522e-06,
|
| 526 |
+
"loss": 0.7528,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.15327257663628832,
|
| 531 |
+
"grad_norm": 2.7680953782694586,
|
| 532 |
+
"learning_rate": 9.914552535787122e-06,
|
| 533 |
+
"loss": 0.7421,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.15534382767191385,
|
| 538 |
+
"grad_norm": 2.8377090221372274,
|
| 539 |
+
"learning_rate": 9.90776756925758e-06,
|
| 540 |
+
"loss": 0.7455,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.15741507870753935,
|
| 545 |
+
"grad_norm": 2.932524071556518,
|
| 546 |
+
"learning_rate": 9.900725916714157e-06,
|
| 547 |
+
"loss": 0.7335,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.15948632974316487,
|
| 552 |
+
"grad_norm": 2.995608607614311,
|
| 553 |
+
"learning_rate": 9.893427946449297e-06,
|
| 554 |
+
"loss": 0.7451,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.1615575807787904,
|
| 559 |
+
"grad_norm": 3.4887684282046254,
|
| 560 |
+
"learning_rate": 9.885874040161373e-06,
|
| 561 |
+
"loss": 0.7333,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.1636288318144159,
|
| 566 |
+
"grad_norm": 2.685483854880427,
|
| 567 |
+
"learning_rate": 9.878064592934723e-06,
|
| 568 |
+
"loss": 0.7387,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.16570008285004142,
|
| 573 |
+
"grad_norm": 2.6709003339090436,
|
| 574 |
+
"learning_rate": 9.87000001321898e-06,
|
| 575 |
+
"loss": 0.7332,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.16777133388566695,
|
| 580 |
+
"grad_norm": 2.8316780066782177,
|
| 581 |
+
"learning_rate": 9.86168072280772e-06,
|
| 582 |
+
"loss": 0.7306,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.16984258492129245,
|
| 587 |
+
"grad_norm": 2.800680145525587,
|
| 588 |
+
"learning_rate": 9.853107156816393e-06,
|
| 589 |
+
"loss": 0.7258,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.17191383595691798,
|
| 594 |
+
"grad_norm": 3.015838299479087,
|
| 595 |
+
"learning_rate": 9.844279763659566e-06,
|
| 596 |
+
"loss": 0.716,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.1739850869925435,
|
| 601 |
+
"grad_norm": 2.861365629083953,
|
| 602 |
+
"learning_rate": 9.835199005027477e-06,
|
| 603 |
+
"loss": 0.7256,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.176056338028169,
|
| 608 |
+
"grad_norm": 2.8083228586917532,
|
| 609 |
+
"learning_rate": 9.825865355861878e-06,
|
| 610 |
+
"loss": 0.7249,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.17812758906379453,
|
| 615 |
+
"grad_norm": 2.9188232580232696,
|
| 616 |
+
"learning_rate": 9.816279304331202e-06,
|
| 617 |
+
"loss": 0.728,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.18019884009942005,
|
| 622 |
+
"grad_norm": 3.047823960774399,
|
| 623 |
+
"learning_rate": 9.806441351805025e-06,
|
| 624 |
+
"loss": 0.7158,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.18227009113504558,
|
| 629 |
+
"grad_norm": 26.196165622685598,
|
| 630 |
+
"learning_rate": 9.79635201282785e-06,
|
| 631 |
+
"loss": 0.73,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.18434134217067108,
|
| 636 |
+
"grad_norm": 2.941695755078049,
|
| 637 |
+
"learning_rate": 9.786011815092193e-06,
|
| 638 |
+
"loss": 0.709,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.1864125932062966,
|
| 643 |
+
"grad_norm": 2.983234429221489,
|
| 644 |
+
"learning_rate": 9.775421299410977e-06,
|
| 645 |
+
"loss": 0.7113,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.18848384424192213,
|
| 650 |
+
"grad_norm": 2.7152686163875512,
|
| 651 |
+
"learning_rate": 9.764581019689255e-06,
|
| 652 |
+
"loss": 0.7245,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.19055509527754763,
|
| 657 |
+
"grad_norm": 3.140548600333453,
|
| 658 |
+
"learning_rate": 9.753491542895237e-06,
|
| 659 |
+
"loss": 0.6944,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.19262634631317316,
|
| 664 |
+
"grad_norm": 2.7171623737634953,
|
| 665 |
+
"learning_rate": 9.742153449030639e-06,
|
| 666 |
+
"loss": 0.6997,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.19469759734879868,
|
| 671 |
+
"grad_norm": 2.8513078067731144,
|
| 672 |
+
"learning_rate": 9.730567331100333e-06,
|
| 673 |
+
"loss": 0.6851,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.19676884838442418,
|
| 678 |
+
"grad_norm": 2.7624587568061973,
|
| 679 |
+
"learning_rate": 9.71873379508136e-06,
|
| 680 |
+
"loss": 0.6879,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.1988400994200497,
|
| 685 |
+
"grad_norm": 2.704414137904599,
|
| 686 |
+
"learning_rate": 9.706653459891207e-06,
|
| 687 |
+
"loss": 0.6869,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.20091135045567524,
|
| 692 |
+
"grad_norm": 2.692949982244341,
|
| 693 |
+
"learning_rate": 9.694326957355452e-06,
|
| 694 |
+
"loss": 0.6939,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.20298260149130073,
|
| 699 |
+
"grad_norm": 3.036163976932751,
|
| 700 |
+
"learning_rate": 9.681754932174719e-06,
|
| 701 |
+
"loss": 0.6873,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.20505385252692626,
|
| 706 |
+
"grad_norm": 2.7169963354471163,
|
| 707 |
+
"learning_rate": 9.668938041890952e-06,
|
| 708 |
+
"loss": 0.6777,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.2071251035625518,
|
| 713 |
+
"grad_norm": 2.938914484551871,
|
| 714 |
+
"learning_rate": 9.655876956853025e-06,
|
| 715 |
+
"loss": 0.6922,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.2091963545981773,
|
| 720 |
+
"grad_norm": 2.933588065770229,
|
| 721 |
+
"learning_rate": 9.64257236018169e-06,
|
| 722 |
+
"loss": 0.6983,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.2112676056338028,
|
| 727 |
+
"grad_norm": 2.6620735297970284,
|
| 728 |
+
"learning_rate": 9.629024947733836e-06,
|
| 729 |
+
"loss": 0.6801,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.21333885666942834,
|
| 734 |
+
"grad_norm": 2.9246922115868244,
|
| 735 |
+
"learning_rate": 9.615235428066106e-06,
|
| 736 |
+
"loss": 0.6745,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.21541010770505387,
|
| 741 |
+
"grad_norm": 2.7814126484359605,
|
| 742 |
+
"learning_rate": 9.601204522397826e-06,
|
| 743 |
+
"loss": 0.6902,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.21748135874067936,
|
| 748 |
+
"grad_norm": 2.8666993290962917,
|
| 749 |
+
"learning_rate": 9.586932964573298e-06,
|
| 750 |
+
"loss": 0.6736,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.2195526097763049,
|
| 755 |
+
"grad_norm": 2.8502393291510058,
|
| 756 |
+
"learning_rate": 9.572421501023403e-06,
|
| 757 |
+
"loss": 0.6726,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.22162386081193042,
|
| 762 |
+
"grad_norm": 2.7285891092196493,
|
| 763 |
+
"learning_rate": 9.557670890726576e-06,
|
| 764 |
+
"loss": 0.7026,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.22369511184755592,
|
| 769 |
+
"grad_norm": 2.7516541498998053,
|
| 770 |
+
"learning_rate": 9.5426819051691e-06,
|
| 771 |
+
"loss": 0.6498,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.22576636288318144,
|
| 776 |
+
"grad_norm": 2.7688759610024336,
|
| 777 |
+
"learning_rate": 9.527455328304756e-06,
|
| 778 |
+
"loss": 0.6617,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.22783761391880697,
|
| 783 |
+
"grad_norm": 2.7974559352399284,
|
| 784 |
+
"learning_rate": 9.511991956513828e-06,
|
| 785 |
+
"loss": 0.6577,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.22990886495443247,
|
| 790 |
+
"grad_norm": 2.913490071476846,
|
| 791 |
+
"learning_rate": 9.496292598561445e-06,
|
| 792 |
+
"loss": 0.6574,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.231980115990058,
|
| 797 |
+
"grad_norm": 2.800168111857422,
|
| 798 |
+
"learning_rate": 9.480358075555278e-06,
|
| 799 |
+
"loss": 0.6615,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.23405136702568352,
|
| 804 |
+
"grad_norm": 2.8532660153153073,
|
| 805 |
+
"learning_rate": 9.464189220902603e-06,
|
| 806 |
+
"loss": 0.6401,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.23612261806130902,
|
| 811 |
+
"grad_norm": 2.731343089226633,
|
| 812 |
+
"learning_rate": 9.447786880266706e-06,
|
| 813 |
+
"loss": 0.6765,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.23819386909693455,
|
| 818 |
+
"grad_norm": 2.7368891548658745,
|
| 819 |
+
"learning_rate": 9.431151911522656e-06,
|
| 820 |
+
"loss": 0.6559,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.24026512013256007,
|
| 825 |
+
"grad_norm": 2.8586019443103443,
|
| 826 |
+
"learning_rate": 9.414285184712432e-06,
|
| 827 |
+
"loss": 0.6549,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.24233637116818557,
|
| 832 |
+
"grad_norm": 2.6893316866041292,
|
| 833 |
+
"learning_rate": 9.397187581999424e-06,
|
| 834 |
+
"loss": 0.6548,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.2444076222038111,
|
| 839 |
+
"grad_norm": 2.960263188185504,
|
| 840 |
+
"learning_rate": 9.37985999762229e-06,
|
| 841 |
+
"loss": 0.6629,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.24647887323943662,
|
| 846 |
+
"grad_norm": 2.794468346887519,
|
| 847 |
+
"learning_rate": 9.362303337848188e-06,
|
| 848 |
+
"loss": 0.6528,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.24855012427506215,
|
| 853 |
+
"grad_norm": 2.7296543899015404,
|
| 854 |
+
"learning_rate": 9.344518520925377e-06,
|
| 855 |
+
"loss": 0.6539,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.2506213753106877,
|
| 860 |
+
"grad_norm": 2.7814326323481793,
|
| 861 |
+
"learning_rate": 9.326506477035179e-06,
|
| 862 |
+
"loss": 0.6509,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.2526926263463132,
|
| 867 |
+
"grad_norm": 2.7657465860310664,
|
| 868 |
+
"learning_rate": 9.308268148243355e-06,
|
| 869 |
+
"loss": 0.657,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.2547638773819387,
|
| 874 |
+
"grad_norm": 2.8791689822910884,
|
| 875 |
+
"learning_rate": 9.289804488450805e-06,
|
| 876 |
+
"loss": 0.6363,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.25683512841756423,
|
| 881 |
+
"grad_norm": 2.7171465550163045,
|
| 882 |
+
"learning_rate": 9.271116463343692e-06,
|
| 883 |
+
"loss": 0.6455,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.25890637945318973,
|
| 888 |
+
"grad_norm": 2.73393442140209,
|
| 889 |
+
"learning_rate": 9.25220505034293e-06,
|
| 890 |
+
"loss": 0.6513,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.2609776304888152,
|
| 895 |
+
"grad_norm": 2.632241124925423,
|
| 896 |
+
"learning_rate": 9.23307123855307e-06,
|
| 897 |
+
"loss": 0.6346,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.2630488815244408,
|
| 902 |
+
"grad_norm": 2.993601522963101,
|
| 903 |
+
"learning_rate": 9.213716028710558e-06,
|
| 904 |
+
"loss": 0.6316,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.2651201325600663,
|
| 909 |
+
"grad_norm": 2.76014350673819,
|
| 910 |
+
"learning_rate": 9.194140433131397e-06,
|
| 911 |
+
"loss": 0.6219,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.2671913835956918,
|
| 916 |
+
"grad_norm": 2.759230217705747,
|
| 917 |
+
"learning_rate": 9.174345475658208e-06,
|
| 918 |
+
"loss": 0.6193,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.26926263463131733,
|
| 923 |
+
"grad_norm": 2.740361207957431,
|
| 924 |
+
"learning_rate": 9.154332191606671e-06,
|
| 925 |
+
"loss": 0.6146,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.27133388566694283,
|
| 930 |
+
"grad_norm": 2.7156945902866,
|
| 931 |
+
"learning_rate": 9.134101627711384e-06,
|
| 932 |
+
"loss": 0.6372,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.27340513670256833,
|
| 937 |
+
"grad_norm": 2.580365026128783,
|
| 938 |
+
"learning_rate": 9.113654842071114e-06,
|
| 939 |
+
"loss": 0.6167,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.2754763877381939,
|
| 944 |
+
"grad_norm": 2.863886132177369,
|
| 945 |
+
"learning_rate": 9.092992904093451e-06,
|
| 946 |
+
"loss": 0.6059,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.2775476387738194,
|
| 951 |
+
"grad_norm": 2.6578998294664733,
|
| 952 |
+
"learning_rate": 9.072116894438885e-06,
|
| 953 |
+
"loss": 0.6168,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.2796188898094449,
|
| 958 |
+
"grad_norm": 2.631660528391497,
|
| 959 |
+
"learning_rate": 9.051027904964279e-06,
|
| 960 |
+
"loss": 0.6146,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.28169014084507044,
|
| 965 |
+
"grad_norm": 2.7672365380596164,
|
| 966 |
+
"learning_rate": 9.029727038665765e-06,
|
| 967 |
+
"loss": 0.6123,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.28376139188069593,
|
| 972 |
+
"grad_norm": 3.0968185232658088,
|
| 973 |
+
"learning_rate": 9.008215409621053e-06,
|
| 974 |
+
"loss": 0.6142,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.28583264291632143,
|
| 979 |
+
"grad_norm": 2.8569762938850514,
|
| 980 |
+
"learning_rate": 8.986494142931168e-06,
|
| 981 |
+
"loss": 0.602,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.287903893951947,
|
| 986 |
+
"grad_norm": 2.7348379248853356,
|
| 987 |
+
"learning_rate": 8.964564374661597e-06,
|
| 988 |
+
"loss": 0.62,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.2899751449875725,
|
| 993 |
+
"grad_norm": 2.6233125816324137,
|
| 994 |
+
"learning_rate": 8.94242725178288e-06,
|
| 995 |
+
"loss": 0.6135,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.292046396023198,
|
| 1000 |
+
"grad_norm": 2.929517704380803,
|
| 1001 |
+
"learning_rate": 8.920083932110608e-06,
|
| 1002 |
+
"loss": 0.6036,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.29411764705882354,
|
| 1007 |
+
"grad_norm": 2.6098554882359553,
|
| 1008 |
+
"learning_rate": 8.89753558424488e-06,
|
| 1009 |
+
"loss": 0.5931,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.29618889809444904,
|
| 1014 |
+
"grad_norm": 2.8924998708965397,
|
| 1015 |
+
"learning_rate": 8.874783387509181e-06,
|
| 1016 |
+
"loss": 0.5934,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.2982601491300746,
|
| 1021 |
+
"grad_norm": 2.596801412566951,
|
| 1022 |
+
"learning_rate": 8.851828531888692e-06,
|
| 1023 |
+
"loss": 0.6051,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.3003314001657001,
|
| 1028 |
+
"grad_norm": 2.619656208203822,
|
| 1029 |
+
"learning_rate": 8.828672217968055e-06,
|
| 1030 |
+
"loss": 0.5834,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.3024026512013256,
|
| 1035 |
+
"grad_norm": 2.791932552629316,
|
| 1036 |
+
"learning_rate": 8.805315656868587e-06,
|
| 1037 |
+
"loss": 0.5977,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.30447390223695114,
|
| 1042 |
+
"grad_norm": 2.6328719857151817,
|
| 1043 |
+
"learning_rate": 8.781760070184933e-06,
|
| 1044 |
+
"loss": 0.5786,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.30654515327257664,
|
| 1049 |
+
"grad_norm": 2.7134170911670306,
|
| 1050 |
+
"learning_rate": 8.75800668992117e-06,
|
| 1051 |
+
"loss": 0.6108,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.30861640430820214,
|
| 1056 |
+
"grad_norm": 2.60002217254436,
|
| 1057 |
+
"learning_rate": 8.734056758426367e-06,
|
| 1058 |
+
"loss": 0.6052,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.3106876553438277,
|
| 1063 |
+
"grad_norm": 2.714229727153767,
|
| 1064 |
+
"learning_rate": 8.709911528329623e-06,
|
| 1065 |
+
"loss": 0.5938,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.3127589063794532,
|
| 1070 |
+
"grad_norm": 2.7258279509398426,
|
| 1071 |
+
"learning_rate": 8.685572262474538e-06,
|
| 1072 |
+
"loss": 0.5964,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.3148301574150787,
|
| 1077 |
+
"grad_norm": 2.7585133476772636,
|
| 1078 |
+
"learning_rate": 8.661040233853166e-06,
|
| 1079 |
+
"loss": 0.5775,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.31690140845070425,
|
| 1084 |
+
"grad_norm": 2.8212453417581127,
|
| 1085 |
+
"learning_rate": 8.636316725539445e-06,
|
| 1086 |
+
"loss": 0.5866,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.31897265948632975,
|
| 1091 |
+
"grad_norm": 2.6417442757955247,
|
| 1092 |
+
"learning_rate": 8.611403030622074e-06,
|
| 1093 |
+
"loss": 0.58,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.32104391052195524,
|
| 1098 |
+
"grad_norm": 2.757161123839532,
|
| 1099 |
+
"learning_rate": 8.586300452136895e-06,
|
| 1100 |
+
"loss": 0.5707,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.3231151615575808,
|
| 1105 |
+
"grad_norm": 2.555247097752916,
|
| 1106 |
+
"learning_rate": 8.561010302998734e-06,
|
| 1107 |
+
"loss": 0.5925,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.3251864125932063,
|
| 1112 |
+
"grad_norm": 2.8366581801399655,
|
| 1113 |
+
"learning_rate": 8.535533905932739e-06,
|
| 1114 |
+
"loss": 0.5736,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.3272576636288318,
|
| 1119 |
+
"grad_norm": 3.056206842303661,
|
| 1120 |
+
"learning_rate": 8.509872593405189e-06,
|
| 1121 |
+
"loss": 0.5779,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.32932891466445735,
|
| 1126 |
+
"grad_norm": 2.53237830007663,
|
| 1127 |
+
"learning_rate": 8.484027707553818e-06,
|
| 1128 |
+
"loss": 0.5708,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.33140016570008285,
|
| 1133 |
+
"grad_norm": 2.8417713226772165,
|
| 1134 |
+
"learning_rate": 8.458000600117604e-06,
|
| 1135 |
+
"loss": 0.5814,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.33347141673570835,
|
| 1140 |
+
"grad_norm": 2.824242675911942,
|
| 1141 |
+
"learning_rate": 8.43179263236608e-06,
|
| 1142 |
+
"loss": 0.5631,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.3355426677713339,
|
| 1147 |
+
"grad_norm": 2.6595589053586792,
|
| 1148 |
+
"learning_rate": 8.40540517502813e-06,
|
| 1149 |
+
"loss": 0.5762,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.3376139188069594,
|
| 1154 |
+
"grad_norm": 2.5496305281888767,
|
| 1155 |
+
"learning_rate": 8.378839608220304e-06,
|
| 1156 |
+
"loss": 0.5708,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.3396851698425849,
|
| 1161 |
+
"grad_norm": 2.7369851726349457,
|
| 1162 |
+
"learning_rate": 8.35209732137463e-06,
|
| 1163 |
+
"loss": 0.5511,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.34175642087821045,
|
| 1168 |
+
"grad_norm": 2.7383476860613425,
|
| 1169 |
+
"learning_rate": 8.32517971316595e-06,
|
| 1170 |
+
"loss": 0.5746,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.34382767191383595,
|
| 1175 |
+
"grad_norm": 2.543223602393049,
|
| 1176 |
+
"learning_rate": 8.298088191438753e-06,
|
| 1177 |
+
"loss": 0.5545,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.34589892294946145,
|
| 1182 |
+
"grad_norm": 2.5789286430343945,
|
| 1183 |
+
"learning_rate": 8.270824173133563e-06,
|
| 1184 |
+
"loss": 0.5589,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.347970173985087,
|
| 1189 |
+
"grad_norm": 2.663964077331657,
|
| 1190 |
+
"learning_rate": 8.243389084212808e-06,
|
| 1191 |
+
"loss": 0.5844,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.3500414250207125,
|
| 1196 |
+
"grad_norm": 2.701713888365594,
|
| 1197 |
+
"learning_rate": 8.215784359586257e-06,
|
| 1198 |
+
"loss": 0.5335,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.352112676056338,
|
| 1203 |
+
"grad_norm": 2.5757700396462617,
|
| 1204 |
+
"learning_rate": 8.188011443035962e-06,
|
| 1205 |
+
"loss": 0.5691,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.35418392709196356,
|
| 1210 |
+
"grad_norm": 3.2685206527006967,
|
| 1211 |
+
"learning_rate": 8.160071787140742e-06,
|
| 1212 |
+
"loss": 0.5604,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.35625517812758906,
|
| 1217 |
+
"grad_norm": 2.6808681382303283,
|
| 1218 |
+
"learning_rate": 8.131966853200226e-06,
|
| 1219 |
+
"loss": 0.5589,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.35832642916321455,
|
| 1224 |
+
"grad_norm": 2.5865759561179695,
|
| 1225 |
+
"learning_rate": 8.103698111158405e-06,
|
| 1226 |
+
"loss": 0.5572,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.3603976801988401,
|
| 1231 |
+
"grad_norm": 2.621331974867394,
|
| 1232 |
+
"learning_rate": 8.075267039526764e-06,
|
| 1233 |
+
"loss": 0.5488,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.3624689312344656,
|
| 1238 |
+
"grad_norm": 2.728050169878493,
|
| 1239 |
+
"learning_rate": 8.046675125306948e-06,
|
| 1240 |
+
"loss": 0.5559,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.36454018227009116,
|
| 1245 |
+
"grad_norm": 2.706287204854472,
|
| 1246 |
+
"learning_rate": 8.017923863912989e-06,
|
| 1247 |
+
"loss": 0.5658,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.36661143330571666,
|
| 1252 |
+
"grad_norm": 2.6490252323634573,
|
| 1253 |
+
"learning_rate": 7.989014759093095e-06,
|
| 1254 |
+
"loss": 0.5414,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.36868268434134216,
|
| 1259 |
+
"grad_norm": 2.523456167360872,
|
| 1260 |
+
"learning_rate": 7.959949322850994e-06,
|
| 1261 |
+
"loss": 0.5441,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.3707539353769677,
|
| 1266 |
+
"grad_norm": 2.5740101976508964,
|
| 1267 |
+
"learning_rate": 7.930729075366867e-06,
|
| 1268 |
+
"loss": 0.5361,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.3728251864125932,
|
| 1273 |
+
"grad_norm": 2.7523704722735736,
|
| 1274 |
+
"learning_rate": 7.901355544917827e-06,
|
| 1275 |
+
"loss": 0.5459,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.3748964374482187,
|
| 1280 |
+
"grad_norm": 2.7136466139331716,
|
| 1281 |
+
"learning_rate": 7.87183026779799e-06,
|
| 1282 |
+
"loss": 0.5373,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.37696768848384427,
|
| 1287 |
+
"grad_norm": 2.5974139734998807,
|
| 1288 |
+
"learning_rate": 7.842154788238124e-06,
|
| 1289 |
+
"loss": 0.5163,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.37903893951946976,
|
| 1294 |
+
"grad_norm": 2.5980610883027624,
|
| 1295 |
+
"learning_rate": 7.812330658324884e-06,
|
| 1296 |
+
"loss": 0.5404,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.38111019055509526,
|
| 1301 |
+
"grad_norm": 2.9020873692936036,
|
| 1302 |
+
"learning_rate": 7.782359437919644e-06,
|
| 1303 |
+
"loss": 0.528,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.3831814415907208,
|
| 1308 |
+
"grad_norm": 2.8690861560512473,
|
| 1309 |
+
"learning_rate": 7.75224269457689e-06,
|
| 1310 |
+
"loss": 0.5265,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.3852526926263463,
|
| 1315 |
+
"grad_norm": 2.5570267026915765,
|
| 1316 |
+
"learning_rate": 7.721982003462255e-06,
|
| 1317 |
+
"loss": 0.5247,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.3873239436619718,
|
| 1322 |
+
"grad_norm": 2.6763850509883946,
|
| 1323 |
+
"learning_rate": 7.691578947270122e-06,
|
| 1324 |
+
"loss": 0.534,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.38939519469759737,
|
| 1329 |
+
"grad_norm": 2.5052520741352837,
|
| 1330 |
+
"learning_rate": 7.661035116140856e-06,
|
| 1331 |
+
"loss": 0.5215,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.39146644573322287,
|
| 1336 |
+
"grad_norm": 2.5675785241463687,
|
| 1337 |
+
"learning_rate": 7.63035210757763e-06,
|
| 1338 |
+
"loss": 0.5311,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.39353769676884837,
|
| 1343 |
+
"grad_norm": 2.564582103010566,
|
| 1344 |
+
"learning_rate": 7.599531526362873e-06,
|
| 1345 |
+
"loss": 0.523,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.3956089478044739,
|
| 1350 |
+
"grad_norm": 2.4030300519005827,
|
| 1351 |
+
"learning_rate": 7.568574984474335e-06,
|
| 1352 |
+
"loss": 0.5321,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.3976801988400994,
|
| 1357 |
+
"grad_norm": 2.55360688034303,
|
| 1358 |
+
"learning_rate": 7.537484101000787e-06,
|
| 1359 |
+
"loss": 0.5109,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.3997514498757249,
|
| 1364 |
+
"grad_norm": 2.731502989984868,
|
| 1365 |
+
"learning_rate": 7.506260502057325e-06,
|
| 1366 |
+
"loss": 0.5451,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.40182270091135047,
|
| 1371 |
+
"grad_norm": 2.6079554614607705,
|
| 1372 |
+
"learning_rate": 7.474905820700334e-06,
|
| 1373 |
+
"loss": 0.5368,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.40389395194697597,
|
| 1378 |
+
"grad_norm": 2.574046069630531,
|
| 1379 |
+
"learning_rate": 7.443421696842066e-06,
|
| 1380 |
+
"loss": 0.545,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.40596520298260147,
|
| 1385 |
+
"grad_norm": 3.0200475159212257,
|
| 1386 |
+
"learning_rate": 7.411809777164873e-06,
|
| 1387 |
+
"loss": 0.5291,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.408036454018227,
|
| 1392 |
+
"grad_norm": 2.7452681073557117,
|
| 1393 |
+
"learning_rate": 7.380071715035089e-06,
|
| 1394 |
+
"loss": 0.5049,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.4101077050538525,
|
| 1399 |
+
"grad_norm": 2.62214319189027,
|
| 1400 |
+
"learning_rate": 7.3482091704165405e-06,
|
| 1401 |
+
"loss": 0.5236,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.412178956089478,
|
| 1406 |
+
"grad_norm": 2.7249510221523847,
|
| 1407 |
+
"learning_rate": 7.316223809783745e-06,
|
| 1408 |
+
"loss": 0.5269,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.4142502071251036,
|
| 1413 |
+
"grad_norm": 2.630893735711103,
|
| 1414 |
+
"learning_rate": 7.284117306034733e-06,
|
| 1415 |
+
"loss": 0.5265,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.4163214581607291,
|
| 1420 |
+
"grad_norm": 2.592181156178893,
|
| 1421 |
+
"learning_rate": 7.2518913384035685e-06,
|
| 1422 |
+
"loss": 0.5128,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.4183927091963546,
|
| 1427 |
+
"grad_norm": 2.5689746215808635,
|
| 1428 |
+
"learning_rate": 7.219547592372512e-06,
|
| 1429 |
+
"loss": 0.5102,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.4204639602319801,
|
| 1434 |
+
"grad_norm": 2.443889242394086,
|
| 1435 |
+
"learning_rate": 7.187087759583869e-06,
|
| 1436 |
+
"loss": 0.5106,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.4225352112676056,
|
| 1441 |
+
"grad_norm": 2.4062918533464717,
|
| 1442 |
+
"learning_rate": 7.15451353775151e-06,
|
| 1443 |
+
"loss": 0.5161,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.4246064623032311,
|
| 1448 |
+
"grad_norm": 2.488756641889923,
|
| 1449 |
+
"learning_rate": 7.121826630572084e-06,
|
| 1450 |
+
"loss": 0.5157,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.4266777133388567,
|
| 1455 |
+
"grad_norm": 2.4678699964797435,
|
| 1456 |
+
"learning_rate": 7.089028747635908e-06,
|
| 1457 |
+
"loss": 0.5222,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.4287489643744822,
|
| 1462 |
+
"grad_norm": 2.7985670814019645,
|
| 1463 |
+
"learning_rate": 7.056121604337554e-06,
|
| 1464 |
+
"loss": 0.5109,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.43082021541010773,
|
| 1469 |
+
"grad_norm": 2.750263108942153,
|
| 1470 |
+
"learning_rate": 7.023106921786118e-06,
|
| 1471 |
+
"loss": 0.5087,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.43289146644573323,
|
| 1476 |
+
"grad_norm": 2.527279053044518,
|
| 1477 |
+
"learning_rate": 6.9899864267152275e-06,
|
| 1478 |
+
"loss": 0.5103,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.43496271748135873,
|
| 1483 |
+
"grad_norm": 2.673130489271966,
|
| 1484 |
+
"learning_rate": 6.956761851392706e-06,
|
| 1485 |
+
"loss": 0.5204,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.4370339685169843,
|
| 1490 |
+
"grad_norm": 2.646174732179495,
|
| 1491 |
+
"learning_rate": 6.9234349335299835e-06,
|
| 1492 |
+
"loss": 0.4949,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.4391052195526098,
|
| 1497 |
+
"grad_norm": 2.4710241904168826,
|
| 1498 |
+
"learning_rate": 6.890007416191209e-06,
|
| 1499 |
+
"loss": 0.4976,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.4411764705882353,
|
| 1504 |
+
"grad_norm": 2.513490801734296,
|
| 1505 |
+
"learning_rate": 6.8564810477020835e-06,
|
| 1506 |
+
"loss": 0.4928,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.44324772162386084,
|
| 1511 |
+
"grad_norm": 2.616872712286189,
|
| 1512 |
+
"learning_rate": 6.822857581558423e-06,
|
| 1513 |
+
"loss": 0.4851,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.44531897265948633,
|
| 1518 |
+
"grad_norm": 2.5547631458988995,
|
| 1519 |
+
"learning_rate": 6.789138776334441e-06,
|
| 1520 |
+
"loss": 0.498,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.44739022369511183,
|
| 1525 |
+
"grad_norm": 2.5426717042905294,
|
| 1526 |
+
"learning_rate": 6.7553263955907755e-06,
|
| 1527 |
+
"loss": 0.5021,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.4494614747307374,
|
| 1532 |
+
"grad_norm": 2.762299627144321,
|
| 1533 |
+
"learning_rate": 6.721422207782249e-06,
|
| 1534 |
+
"loss": 0.4894,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.4515327257663629,
|
| 1539 |
+
"grad_norm": 2.6282943687365314,
|
| 1540 |
+
"learning_rate": 6.687427986165379e-06,
|
| 1541 |
+
"loss": 0.4885,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.4536039768019884,
|
| 1546 |
+
"grad_norm": 2.6227170032352505,
|
| 1547 |
+
"learning_rate": 6.653345508705629e-06,
|
| 1548 |
+
"loss": 0.4846,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.45567522783761394,
|
| 1553 |
+
"grad_norm": 2.573980996203933,
|
| 1554 |
+
"learning_rate": 6.6191765579844205e-06,
|
| 1555 |
+
"loss": 0.4677,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.45774647887323944,
|
| 1560 |
+
"grad_norm": 2.760609026653492,
|
| 1561 |
+
"learning_rate": 6.584922921105894e-06,
|
| 1562 |
+
"loss": 0.4704,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.45981772990886494,
|
| 1567 |
+
"grad_norm": 2.5519488960897942,
|
| 1568 |
+
"learning_rate": 6.550586389603451e-06,
|
| 1569 |
+
"loss": 0.4763,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.4618889809444905,
|
| 1574 |
+
"grad_norm": 2.4528803429641814,
|
| 1575 |
+
"learning_rate": 6.5161687593460395e-06,
|
| 1576 |
+
"loss": 0.4854,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.463960231980116,
|
| 1581 |
+
"grad_norm": 2.7025880432763962,
|
| 1582 |
+
"learning_rate": 6.481671830444243e-06,
|
| 1583 |
+
"loss": 0.497,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.4660314830157415,
|
| 1588 |
+
"grad_norm": 2.4770462760723895,
|
| 1589 |
+
"learning_rate": 6.447097407156114e-06,
|
| 1590 |
+
"loss": 0.4799,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.46810273405136704,
|
| 1595 |
+
"grad_norm": 2.416009894405175,
|
| 1596 |
+
"learning_rate": 6.412447297792818e-06,
|
| 1597 |
+
"loss": 0.4772,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.47017398508699254,
|
| 1602 |
+
"grad_norm": 2.422088326310343,
|
| 1603 |
+
"learning_rate": 6.377723314624057e-06,
|
| 1604 |
+
"loss": 0.4684,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.47224523612261804,
|
| 1609 |
+
"grad_norm": 2.6486702874804497,
|
| 1610 |
+
"learning_rate": 6.3429272737832726e-06,
|
| 1611 |
+
"loss": 0.4737,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.4743164871582436,
|
| 1616 |
+
"grad_norm": 2.474167276592783,
|
| 1617 |
+
"learning_rate": 6.308060995172673e-06,
|
| 1618 |
+
"loss": 0.4616,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.4763877381938691,
|
| 1623 |
+
"grad_norm": 2.517931093717021,
|
| 1624 |
+
"learning_rate": 6.273126302368037e-06,
|
| 1625 |
+
"loss": 0.4656,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.4784589892294946,
|
| 1630 |
+
"grad_norm": 2.6624977734074062,
|
| 1631 |
+
"learning_rate": 6.238125022523343e-06,
|
| 1632 |
+
"loss": 0.4747,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.48053024026512015,
|
| 1637 |
+
"grad_norm": 2.501404850938584,
|
| 1638 |
+
"learning_rate": 6.203058986275207e-06,
|
| 1639 |
+
"loss": 0.4718,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.48260149130074564,
|
| 1644 |
+
"grad_norm": 2.393771169197383,
|
| 1645 |
+
"learning_rate": 6.1679300276471285e-06,
|
| 1646 |
+
"loss": 0.4685,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.48467274233637114,
|
| 1651 |
+
"grad_norm": 2.579946322003478,
|
| 1652 |
+
"learning_rate": 6.132739983953579e-06,
|
| 1653 |
+
"loss": 0.4616,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.4867439933719967,
|
| 1658 |
+
"grad_norm": 2.5308648946074124,
|
| 1659 |
+
"learning_rate": 6.097490695703896e-06,
|
| 1660 |
+
"loss": 0.4624,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.4888152444076222,
|
| 1665 |
+
"grad_norm": 2.4343897742629497,
|
| 1666 |
+
"learning_rate": 6.062184006506027e-06,
|
| 1667 |
+
"loss": 0.4576,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.4908864954432477,
|
| 1672 |
+
"grad_norm": 2.5040006985626633,
|
| 1673 |
+
"learning_rate": 6.026821762970102e-06,
|
| 1674 |
+
"loss": 0.4503,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.49295774647887325,
|
| 1679 |
+
"grad_norm": 2.637507097654403,
|
| 1680 |
+
"learning_rate": 5.991405814611855e-06,
|
| 1681 |
+
"loss": 0.4665,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.49502899751449875,
|
| 1686 |
+
"grad_norm": 2.410738785571731,
|
| 1687 |
+
"learning_rate": 5.955938013755888e-06,
|
| 1688 |
+
"loss": 0.4745,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.4971002485501243,
|
| 1693 |
+
"grad_norm": 2.5556725277525225,
|
| 1694 |
+
"learning_rate": 5.920420215438794e-06,
|
| 1695 |
+
"loss": 0.4609,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.4991714995857498,
|
| 1700 |
+
"grad_norm": 2.450509349076648,
|
| 1701 |
+
"learning_rate": 5.8848542773121285e-06,
|
| 1702 |
+
"loss": 0.4692,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.5012427506213754,
|
| 1707 |
+
"grad_norm": 2.633730260436242,
|
| 1708 |
+
"learning_rate": 5.849242059545259e-06,
|
| 1709 |
+
"loss": 0.4859,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5033140016570008,
|
| 1714 |
+
"grad_norm": 2.5117312210011784,
|
| 1715 |
+
"learning_rate": 5.81358542472807e-06,
|
| 1716 |
+
"loss": 0.4598,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.5053852526926264,
|
| 1721 |
+
"grad_norm": 2.727727787839533,
|
| 1722 |
+
"learning_rate": 5.777886237773542e-06,
|
| 1723 |
+
"loss": 0.458,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.5074565037282519,
|
| 1728 |
+
"grad_norm": 2.498652640274728,
|
| 1729 |
+
"learning_rate": 5.742146365820223e-06,
|
| 1730 |
+
"loss": 0.4568,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.5095277547638773,
|
| 1735 |
+
"grad_norm": 2.4105214571197564,
|
| 1736 |
+
"learning_rate": 5.706367678134562e-06,
|
| 1737 |
+
"loss": 0.4529,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.5115990057995029,
|
| 1742 |
+
"grad_norm": 2.482609496489558,
|
| 1743 |
+
"learning_rate": 5.670552046013151e-06,
|
| 1744 |
+
"loss": 0.4353,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.5136702568351285,
|
| 1749 |
+
"grad_norm": 2.464474174865273,
|
| 1750 |
+
"learning_rate": 5.634701342684852e-06,
|
| 1751 |
+
"loss": 0.459,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.5157415078707539,
|
| 1756 |
+
"grad_norm": 2.5816384699470394,
|
| 1757 |
+
"learning_rate": 5.598817443212813e-06,
|
| 1758 |
+
"loss": 0.4535,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.5178127589063795,
|
| 1763 |
+
"grad_norm": 2.4360648057548056,
|
| 1764 |
+
"learning_rate": 5.562902224396416e-06,
|
| 1765 |
+
"loss": 0.4516,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.519884009942005,
|
| 1770 |
+
"grad_norm": 2.451723183702262,
|
| 1771 |
+
"learning_rate": 5.526957564673098e-06,
|
| 1772 |
+
"loss": 0.4601,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.5219552609776305,
|
| 1777 |
+
"grad_norm": 2.455817733692636,
|
| 1778 |
+
"learning_rate": 5.49098534402012e-06,
|
| 1779 |
+
"loss": 0.4622,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.524026512013256,
|
| 1784 |
+
"grad_norm": 2.3800711048949017,
|
| 1785 |
+
"learning_rate": 5.454987443856235e-06,
|
| 1786 |
+
"loss": 0.4494,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.5260977630488816,
|
| 1791 |
+
"grad_norm": 2.4719917969907708,
|
| 1792 |
+
"learning_rate": 5.418965746943281e-06,
|
| 1793 |
+
"loss": 0.4546,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.528169014084507,
|
| 1798 |
+
"grad_norm": 2.5259312235567037,
|
| 1799 |
+
"learning_rate": 5.3829221372877175e-06,
|
| 1800 |
+
"loss": 0.4428,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.5302402651201326,
|
| 1805 |
+
"grad_norm": 2.5192222564921867,
|
| 1806 |
+
"learning_rate": 5.34685850004208e-06,
|
| 1807 |
+
"loss": 0.4552,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.5323115161557581,
|
| 1812 |
+
"grad_norm": 2.5214221566614383,
|
| 1813 |
+
"learning_rate": 5.310776721406392e-06,
|
| 1814 |
+
"loss": 0.4447,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.5343827671913836,
|
| 1819 |
+
"grad_norm": 2.512402482362361,
|
| 1820 |
+
"learning_rate": 5.2746786885295034e-06,
|
| 1821 |
+
"loss": 0.4411,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.5364540182270091,
|
| 1826 |
+
"grad_norm": 2.537503773050984,
|
| 1827 |
+
"learning_rate": 5.238566289410396e-06,
|
| 1828 |
+
"loss": 0.4397,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.5385252692626347,
|
| 1833 |
+
"grad_norm": 2.3626937823064145,
|
| 1834 |
+
"learning_rate": 5.2024414127994325e-06,
|
| 1835 |
+
"loss": 0.4425,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.5405965202982601,
|
| 1840 |
+
"grad_norm": 2.5522641550285416,
|
| 1841 |
+
"learning_rate": 5.166305948099574e-06,
|
| 1842 |
+
"loss": 0.4442,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.5426677713338857,
|
| 1847 |
+
"grad_norm": 2.5568212604328453,
|
| 1848 |
+
"learning_rate": 5.13016178526756e-06,
|
| 1849 |
+
"loss": 0.4396,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.5447390223695112,
|
| 1854 |
+
"grad_norm": 2.402388362583747,
|
| 1855 |
+
"learning_rate": 5.094010814715062e-06,
|
| 1856 |
+
"loss": 0.4347,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.5468102734051367,
|
| 1861 |
+
"grad_norm": 2.6483619826882356,
|
| 1862 |
+
"learning_rate": 5.057854927209804e-06,
|
| 1863 |
+
"loss": 0.4275,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.5488815244407622,
|
| 1868 |
+
"grad_norm": 2.5263379021082084,
|
| 1869 |
+
"learning_rate": 5.0216960137766805e-06,
|
| 1870 |
+
"loss": 0.4343,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.5509527754763878,
|
| 1875 |
+
"grad_norm": 2.5196212568619796,
|
| 1876 |
+
"learning_rate": 4.985535965598843e-06,
|
| 1877 |
+
"loss": 0.4342,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.5530240265120132,
|
| 1882 |
+
"grad_norm": 2.625850121251377,
|
| 1883 |
+
"learning_rate": 4.949376673918802e-06,
|
| 1884 |
+
"loss": 0.4408,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.5550952775476388,
|
| 1889 |
+
"grad_norm": 2.3452199896145816,
|
| 1890 |
+
"learning_rate": 4.913220029939491e-06,
|
| 1891 |
+
"loss": 0.4257,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.5571665285832643,
|
| 1896 |
+
"grad_norm": 2.574940433175934,
|
| 1897 |
+
"learning_rate": 4.877067924725368e-06,
|
| 1898 |
+
"loss": 0.4254,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.5592377796188898,
|
| 1903 |
+
"grad_norm": 2.3655525367188197,
|
| 1904 |
+
"learning_rate": 4.840922249103506e-06,
|
| 1905 |
+
"loss": 0.42,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.5613090306545153,
|
| 1910 |
+
"grad_norm": 2.5151014112590473,
|
| 1911 |
+
"learning_rate": 4.804784893564697e-06,
|
| 1912 |
+
"loss": 0.4325,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.5633802816901409,
|
| 1917 |
+
"grad_norm": 2.4340664966958174,
|
| 1918 |
+
"learning_rate": 4.7686577481645745e-06,
|
| 1919 |
+
"loss": 0.4162,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.5654515327257663,
|
| 1924 |
+
"grad_norm": 2.361074940083378,
|
| 1925 |
+
"learning_rate": 4.732542702424759e-06,
|
| 1926 |
+
"loss": 0.4273,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.5675227837613919,
|
| 1931 |
+
"grad_norm": 2.4166634583502065,
|
| 1932 |
+
"learning_rate": 4.696441645234042e-06,
|
| 1933 |
+
"loss": 0.4247,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.5695940347970174,
|
| 1938 |
+
"grad_norm": 2.416910012406247,
|
| 1939 |
+
"learning_rate": 4.660356464749578e-06,
|
| 1940 |
+
"loss": 0.4272,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.5716652858326429,
|
| 1945 |
+
"grad_norm": 2.5029874954691294,
|
| 1946 |
+
"learning_rate": 4.624289048298147e-06,
|
| 1947 |
+
"loss": 0.4115,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.5737365368682684,
|
| 1952 |
+
"grad_norm": 2.5064068268567343,
|
| 1953 |
+
"learning_rate": 4.588241282277428e-06,
|
| 1954 |
+
"loss": 0.4153,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.575807787903894,
|
| 1959 |
+
"grad_norm": 2.497767560214899,
|
| 1960 |
+
"learning_rate": 4.55221505205734e-06,
|
| 1961 |
+
"loss": 0.4184,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.5778790389395194,
|
| 1966 |
+
"grad_norm": 2.4311887305765882,
|
| 1967 |
+
"learning_rate": 4.516212241881448e-06,
|
| 1968 |
+
"loss": 0.4184,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.579950289975145,
|
| 1973 |
+
"grad_norm": 2.463809168583045,
|
| 1974 |
+
"learning_rate": 4.480234734768393e-06,
|
| 1975 |
+
"loss": 0.4346,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.5820215410107705,
|
| 1980 |
+
"grad_norm": 2.4645651971416744,
|
| 1981 |
+
"learning_rate": 4.444284412413418e-06,
|
| 1982 |
+
"loss": 0.4144,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.584092792046396,
|
| 1987 |
+
"grad_norm": 2.4306509639257277,
|
| 1988 |
+
"learning_rate": 4.408363155089952e-06,
|
| 1989 |
+
"loss": 0.412,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.5861640430820215,
|
| 1994 |
+
"grad_norm": 2.378781711346702,
|
| 1995 |
+
"learning_rate": 4.3724728415512585e-06,
|
| 1996 |
+
"loss": 0.4232,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.5882352941176471,
|
| 2001 |
+
"grad_norm": 2.279100422257553,
|
| 2002 |
+
"learning_rate": 4.3366153489321855e-06,
|
| 2003 |
+
"loss": 0.425,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.5903065451532725,
|
| 2008 |
+
"grad_norm": 2.3440151093069055,
|
| 2009 |
+
"learning_rate": 4.30079255265098e-06,
|
| 2010 |
+
"loss": 0.4109,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.5923777961888981,
|
| 2015 |
+
"grad_norm": 2.461552106548784,
|
| 2016 |
+
"learning_rate": 4.265006326311199e-06,
|
| 2017 |
+
"loss": 0.4188,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.5944490472245236,
|
| 2022 |
+
"grad_norm": 2.554483811668929,
|
| 2023 |
+
"learning_rate": 4.229258541603723e-06,
|
| 2024 |
+
"loss": 0.415,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.5965202982601492,
|
| 2029 |
+
"grad_norm": 2.3890846762026685,
|
| 2030 |
+
"learning_rate": 4.1935510682088545e-06,
|
| 2031 |
+
"loss": 0.4105,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.5985915492957746,
|
| 2036 |
+
"grad_norm": 2.4838854669224677,
|
| 2037 |
+
"learning_rate": 4.157885773698535e-06,
|
| 2038 |
+
"loss": 0.4017,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.6006628003314002,
|
| 2043 |
+
"grad_norm": 2.5664440683068914,
|
| 2044 |
+
"learning_rate": 4.122264523438668e-06,
|
| 2045 |
+
"loss": 0.4021,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.6027340513670257,
|
| 2050 |
+
"grad_norm": 2.5100602917679797,
|
| 2051 |
+
"learning_rate": 4.086689180491554e-06,
|
| 2052 |
+
"loss": 0.4087,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.6048053024026512,
|
| 2057 |
+
"grad_norm": 2.464742428363488,
|
| 2058 |
+
"learning_rate": 4.051161605518453e-06,
|
| 2059 |
+
"loss": 0.4121,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.6068765534382767,
|
| 2064 |
+
"grad_norm": 2.448216596422355,
|
| 2065 |
+
"learning_rate": 4.015683656682255e-06,
|
| 2066 |
+
"loss": 0.4171,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.6089478044739023,
|
| 2071 |
+
"grad_norm": 2.463617402393949,
|
| 2072 |
+
"learning_rate": 3.980257189550316e-06,
|
| 2073 |
+
"loss": 0.4114,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.6110190555095277,
|
| 2078 |
+
"grad_norm": 2.4607870559606693,
|
| 2079 |
+
"learning_rate": 3.94488405699739e-06,
|
| 2080 |
+
"loss": 0.4041,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.6130903065451533,
|
| 2085 |
+
"grad_norm": 2.460386225566258,
|
| 2086 |
+
"learning_rate": 3.909566109108727e-06,
|
| 2087 |
+
"loss": 0.4019,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.6151615575807788,
|
| 2092 |
+
"grad_norm": 2.4056278033306358,
|
| 2093 |
+
"learning_rate": 3.874305193083313e-06,
|
| 2094 |
+
"loss": 0.4088,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.6172328086164043,
|
| 2099 |
+
"grad_norm": 2.7408409480377536,
|
| 2100 |
+
"learning_rate": 3.839103153137247e-06,
|
| 2101 |
+
"loss": 0.4043,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.6193040596520298,
|
| 2106 |
+
"grad_norm": 2.4491154252670353,
|
| 2107 |
+
"learning_rate": 3.803961830407297e-06,
|
| 2108 |
+
"loss": 0.4068,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.6213753106876554,
|
| 2113 |
+
"grad_norm": 2.57957185284119,
|
| 2114 |
+
"learning_rate": 3.768883062854598e-06,
|
| 2115 |
+
"loss": 0.3997,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.6234465617232808,
|
| 2120 |
+
"grad_norm": 2.454129058379837,
|
| 2121 |
+
"learning_rate": 3.7338686851685267e-06,
|
| 2122 |
+
"loss": 0.4013,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.6255178127589064,
|
| 2127 |
+
"grad_norm": 2.4197107098430006,
|
| 2128 |
+
"learning_rate": 3.6989205286707398e-06,
|
| 2129 |
+
"loss": 0.4017,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.6275890637945319,
|
| 2134 |
+
"grad_norm": 2.6100868528373593,
|
| 2135 |
+
"learning_rate": 3.664040421219393e-06,
|
| 2136 |
+
"loss": 0.4042,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.6296603148301574,
|
| 2141 |
+
"grad_norm": 2.466076344798711,
|
| 2142 |
+
"learning_rate": 3.6292301871135425e-06,
|
| 2143 |
+
"loss": 0.3953,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.6317315658657829,
|
| 2148 |
+
"grad_norm": 2.295401269763307,
|
| 2149 |
+
"learning_rate": 3.59449164699773e-06,
|
| 2150 |
+
"loss": 0.4103,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.6338028169014085,
|
| 2155 |
+
"grad_norm": 2.3840288567182686,
|
| 2156 |
+
"learning_rate": 3.55982661776676e-06,
|
| 2157 |
+
"loss": 0.3982,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.6358740679370339,
|
| 2162 |
+
"grad_norm": 2.4506295124869912,
|
| 2163 |
+
"learning_rate": 3.5252369124706697e-06,
|
| 2164 |
+
"loss": 0.3987,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.6379453189726595,
|
| 2169 |
+
"grad_norm": 2.3732585245856597,
|
| 2170 |
+
"learning_rate": 3.4907243402199013e-06,
|
| 2171 |
+
"loss": 0.391,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.640016570008285,
|
| 2176 |
+
"grad_norm": 2.356804936776819,
|
| 2177 |
+
"learning_rate": 3.4562907060906908e-06,
|
| 2178 |
+
"loss": 0.4084,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.6420878210439105,
|
| 2183 |
+
"grad_norm": 2.3210923465162647,
|
| 2184 |
+
"learning_rate": 3.4219378110306523e-06,
|
| 2185 |
+
"loss": 0.393,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.644159072079536,
|
| 2190 |
+
"grad_norm": 2.3856628563586275,
|
| 2191 |
+
"learning_rate": 3.3876674517645815e-06,
|
| 2192 |
+
"loss": 0.3807,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.6462303231151616,
|
| 2197 |
+
"grad_norm": 2.4513218289135192,
|
| 2198 |
+
"learning_rate": 3.353481420700495e-06,
|
| 2199 |
+
"loss": 0.392,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.648301574150787,
|
| 2204 |
+
"grad_norm": 2.4846929435894363,
|
| 2205 |
+
"learning_rate": 3.319381505835868e-06,
|
| 2206 |
+
"loss": 0.389,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.6503728251864126,
|
| 2211 |
+
"grad_norm": 2.401342535607237,
|
| 2212 |
+
"learning_rate": 3.285369490664133e-06,
|
| 2213 |
+
"loss": 0.3826,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.6524440762220381,
|
| 2218 |
+
"grad_norm": 2.4190000807324408,
|
| 2219 |
+
"learning_rate": 3.251447154081394e-06,
|
| 2220 |
+
"loss": 0.395,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.6545153272576636,
|
| 2225 |
+
"grad_norm": 2.3877772373335415,
|
| 2226 |
+
"learning_rate": 3.2176162702933816e-06,
|
| 2227 |
+
"loss": 0.4006,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.6565865782932891,
|
| 2232 |
+
"grad_norm": 2.391815584391418,
|
| 2233 |
+
"learning_rate": 3.183878608722669e-06,
|
| 2234 |
+
"loss": 0.3836,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.6586578293289147,
|
| 2239 |
+
"grad_norm": 2.5034068900739834,
|
| 2240 |
+
"learning_rate": 3.150235933916115e-06,
|
| 2241 |
+
"loss": 0.3808,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.6607290803645401,
|
| 2246 |
+
"grad_norm": 2.774335520816645,
|
| 2247 |
+
"learning_rate": 3.1166900054525873e-06,
|
| 2248 |
+
"loss": 0.3912,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.6628003314001657,
|
| 2253 |
+
"grad_norm": 2.4103041919857233,
|
| 2254 |
+
"learning_rate": 3.0832425778509235e-06,
|
| 2255 |
+
"loss": 0.3826,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.6648715824357913,
|
| 2260 |
+
"grad_norm": 2.3239461500431853,
|
| 2261 |
+
"learning_rate": 3.049895400478174e-06,
|
| 2262 |
+
"loss": 0.3891,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.6669428334714167,
|
| 2267 |
+
"grad_norm": 2.2530556952367693,
|
| 2268 |
+
"learning_rate": 3.0166502174581012e-06,
|
| 2269 |
+
"loss": 0.3909,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.6690140845070423,
|
| 2274 |
+
"grad_norm": 2.4287687098232396,
|
| 2275 |
+
"learning_rate": 2.983508767579956e-06,
|
| 2276 |
+
"loss": 0.3875,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.6710853355426678,
|
| 2281 |
+
"grad_norm": 2.3815469152902136,
|
| 2282 |
+
"learning_rate": 2.950472784207544e-06,
|
| 2283 |
+
"loss": 0.3897,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.6731565865782932,
|
| 2288 |
+
"grad_norm": 2.516171148051071,
|
| 2289 |
+
"learning_rate": 2.917543995188562e-06,
|
| 2290 |
+
"loss": 0.3885,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.6752278376139188,
|
| 2295 |
+
"grad_norm": 2.3576789549853205,
|
| 2296 |
+
"learning_rate": 2.8847241227642255e-06,
|
| 2297 |
+
"loss": 0.3873,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.6772990886495444,
|
| 2302 |
+
"grad_norm": 2.385493122403789,
|
| 2303 |
+
"learning_rate": 2.852014883479198e-06,
|
| 2304 |
+
"loss": 0.3822,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.6793703396851698,
|
| 2309 |
+
"grad_norm": 2.3142689230066043,
|
| 2310 |
+
"learning_rate": 2.819417988091814e-06,
|
| 2311 |
+
"loss": 0.3797,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.6814415907207954,
|
| 2316 |
+
"grad_norm": 2.380402531887656,
|
| 2317 |
+
"learning_rate": 2.786935141484586e-06,
|
| 2318 |
+
"loss": 0.3794,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.6835128417564209,
|
| 2323 |
+
"grad_norm": 2.382965883978087,
|
| 2324 |
+
"learning_rate": 2.754568042575061e-06,
|
| 2325 |
+
"loss": 0.363,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.6855840927920464,
|
| 2330 |
+
"grad_norm": 2.4286550466790935,
|
| 2331 |
+
"learning_rate": 2.7223183842269442e-06,
|
| 2332 |
+
"loss": 0.3874,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.6876553438276719,
|
| 2337 |
+
"grad_norm": 2.3689462455130808,
|
| 2338 |
+
"learning_rate": 2.6901878531615677e-06,
|
| 2339 |
+
"loss": 0.3801,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.6897265948632975,
|
| 2344 |
+
"grad_norm": 2.435325781453067,
|
| 2345 |
+
"learning_rate": 2.658178129869672e-06,
|
| 2346 |
+
"loss": 0.3761,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.6917978458989229,
|
| 2351 |
+
"grad_norm": 2.3644739422275833,
|
| 2352 |
+
"learning_rate": 2.6262908885235046e-06,
|
| 2353 |
+
"loss": 0.3731,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.6938690969345485,
|
| 2358 |
+
"grad_norm": 2.4420478405497352,
|
| 2359 |
+
"learning_rate": 2.594527796889265e-06,
|
| 2360 |
+
"loss": 0.3721,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.695940347970174,
|
| 2365 |
+
"grad_norm": 2.380326746264982,
|
| 2366 |
+
"learning_rate": 2.5628905162398797e-06,
|
| 2367 |
+
"loss": 0.3625,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.6980115990057995,
|
| 2372 |
+
"grad_norm": 2.439689809561185,
|
| 2373 |
+
"learning_rate": 2.531380701268108e-06,
|
| 2374 |
+
"loss": 0.3677,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.700082850041425,
|
| 2379 |
+
"grad_norm": 2.4792328667641823,
|
| 2380 |
+
"learning_rate": 2.5000000000000015e-06,
|
| 2381 |
+
"loss": 0.3751,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.7021541010770506,
|
| 2386 |
+
"grad_norm": 2.3577329201592527,
|
| 2387 |
+
"learning_rate": 2.4687500537087027e-06,
|
| 2388 |
+
"loss": 0.3639,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.704225352112676,
|
| 2393 |
+
"grad_norm": 2.397062265682441,
|
| 2394 |
+
"learning_rate": 2.4376324968286154e-06,
|
| 2395 |
+
"loss": 0.371,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.7062966031483016,
|
| 2400 |
+
"grad_norm": 2.5377567487355197,
|
| 2401 |
+
"learning_rate": 2.40664895686991e-06,
|
| 2402 |
+
"loss": 0.3692,
|
| 2403 |
+
"step": 1705
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.7083678541839271,
|
| 2407 |
+
"grad_norm": 2.476875205585498,
|
| 2408 |
+
"learning_rate": 2.375801054333409e-06,
|
| 2409 |
+
"loss": 0.3724,
|
| 2410 |
+
"step": 1710
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.7104391052195526,
|
| 2414 |
+
"grad_norm": 2.3221981126145916,
|
| 2415 |
+
"learning_rate": 2.345090402625822e-06,
|
| 2416 |
+
"loss": 0.3666,
|
| 2417 |
+
"step": 1715
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.7125103562551781,
|
| 2421 |
+
"grad_norm": 2.8524714154103408,
|
| 2422 |
+
"learning_rate": 2.3145186079753685e-06,
|
| 2423 |
+
"loss": 0.379,
|
| 2424 |
+
"step": 1720
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.7145816072908037,
|
| 2428 |
+
"grad_norm": 2.3860769129554447,
|
| 2429 |
+
"learning_rate": 2.2840872693477694e-06,
|
| 2430 |
+
"loss": 0.3641,
|
| 2431 |
+
"step": 1725
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.7166528583264291,
|
| 2435 |
+
"grad_norm": 2.3054527958287143,
|
| 2436 |
+
"learning_rate": 2.253797978362617e-06,
|
| 2437 |
+
"loss": 0.3672,
|
| 2438 |
+
"step": 1730
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.7187241093620547,
|
| 2442 |
+
"grad_norm": 2.3407707603025805,
|
| 2443 |
+
"learning_rate": 2.2236523192101264e-06,
|
| 2444 |
+
"loss": 0.3579,
|
| 2445 |
+
"step": 1735
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.7207953603976802,
|
| 2449 |
+
"grad_norm": 2.293792797634113,
|
| 2450 |
+
"learning_rate": 2.193651868568285e-06,
|
| 2451 |
+
"loss": 0.3591,
|
| 2452 |
+
"step": 1740
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.7228666114333057,
|
| 2456 |
+
"grad_norm": 2.4185753681052202,
|
| 2457 |
+
"learning_rate": 2.16379819552038e-06,
|
| 2458 |
+
"loss": 0.3681,
|
| 2459 |
+
"step": 1745
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.7249378624689312,
|
| 2463 |
+
"grad_norm": 2.524421200334089,
|
| 2464 |
+
"learning_rate": 2.1340928614729445e-06,
|
| 2465 |
+
"loss": 0.3745,
|
| 2466 |
+
"step": 1750
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.7270091135045568,
|
| 2470 |
+
"grad_norm": 2.455923080730161,
|
| 2471 |
+
"learning_rate": 2.1045374200740863e-06,
|
| 2472 |
+
"loss": 0.3604,
|
| 2473 |
+
"step": 1755
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.7290803645401823,
|
| 2477 |
+
"grad_norm": 2.40811466685782,
|
| 2478 |
+
"learning_rate": 2.075133417132223e-06,
|
| 2479 |
+
"loss": 0.3601,
|
| 2480 |
+
"step": 1760
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.7311516155758078,
|
| 2484 |
+
"grad_norm": 2.442010182183811,
|
| 2485 |
+
"learning_rate": 2.045882390535248e-06,
|
| 2486 |
+
"loss": 0.3644,
|
| 2487 |
+
"step": 1765
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.7332228666114333,
|
| 2491 |
+
"grad_norm": 2.498938310765272,
|
| 2492 |
+
"learning_rate": 2.016785870170079e-06,
|
| 2493 |
+
"loss": 0.3659,
|
| 2494 |
+
"step": 1770
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.7352941176470589,
|
| 2498 |
+
"grad_norm": 2.43600758234028,
|
| 2499 |
+
"learning_rate": 1.987845377842656e-06,
|
| 2500 |
+
"loss": 0.3567,
|
| 2501 |
+
"step": 1775
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.7373653686826843,
|
| 2505 |
+
"grad_norm": 2.2394000440439883,
|
| 2506 |
+
"learning_rate": 1.9590624271983406e-06,
|
| 2507 |
+
"loss": 0.3622,
|
| 2508 |
+
"step": 1780
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.7394366197183099,
|
| 2512 |
+
"grad_norm": 2.5234088238169705,
|
| 2513 |
+
"learning_rate": 1.9304385236427505e-06,
|
| 2514 |
+
"loss": 0.3531,
|
| 2515 |
+
"step": 1785
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.7415078707539354,
|
| 2519 |
+
"grad_norm": 2.4931659816433287,
|
| 2520 |
+
"learning_rate": 1.9019751642630252e-06,
|
| 2521 |
+
"loss": 0.3505,
|
| 2522 |
+
"step": 1790
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.7435791217895609,
|
| 2526 |
+
"grad_norm": 2.3924059427659916,
|
| 2527 |
+
"learning_rate": 1.8736738377495196e-06,
|
| 2528 |
+
"loss": 0.3599,
|
| 2529 |
+
"step": 1795
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.7456503728251864,
|
| 2533 |
+
"grad_norm": 2.388334766295047,
|
| 2534 |
+
"learning_rate": 1.8455360243179537e-06,
|
| 2535 |
+
"loss": 0.3603,
|
| 2536 |
+
"step": 1800
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.747721623860812,
|
| 2540 |
+
"grad_norm": 2.4055399882312947,
|
| 2541 |
+
"learning_rate": 1.8175631956319823e-06,
|
| 2542 |
+
"loss": 0.3499,
|
| 2543 |
+
"step": 1805
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.7497928748964374,
|
| 2547 |
+
"grad_norm": 2.3681998025448743,
|
| 2548 |
+
"learning_rate": 1.7897568147262323e-06,
|
| 2549 |
+
"loss": 0.3451,
|
| 2550 |
+
"step": 1810
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.751864125932063,
|
| 2554 |
+
"grad_norm": 2.3399190104028618,
|
| 2555 |
+
"learning_rate": 1.7621183359297817e-06,
|
| 2556 |
+
"loss": 0.3436,
|
| 2557 |
+
"step": 1815
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.7539353769676885,
|
| 2561 |
+
"grad_norm": 2.299763408013497,
|
| 2562 |
+
"learning_rate": 1.7346492047900897e-06,
|
| 2563 |
+
"loss": 0.3552,
|
| 2564 |
+
"step": 1820
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.756006628003314,
|
| 2568 |
+
"grad_norm": 2.304240086627427,
|
| 2569 |
+
"learning_rate": 1.7073508579973996e-06,
|
| 2570 |
+
"loss": 0.3501,
|
| 2571 |
+
"step": 1825
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.7580778790389395,
|
| 2575 |
+
"grad_norm": 2.471808773880512,
|
| 2576 |
+
"learning_rate": 1.6802247233095914e-06,
|
| 2577 |
+
"loss": 0.3575,
|
| 2578 |
+
"step": 1830
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.7601491300745651,
|
| 2582 |
+
"grad_norm": 2.1787659024138453,
|
| 2583 |
+
"learning_rate": 1.6532722194775108e-06,
|
| 2584 |
+
"loss": 0.3416,
|
| 2585 |
+
"step": 1835
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.7622203811101905,
|
| 2589 |
+
"grad_norm": 2.410186390448844,
|
| 2590 |
+
"learning_rate": 1.626494756170765e-06,
|
| 2591 |
+
"loss": 0.3573,
|
| 2592 |
+
"step": 1840
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.7642916321458161,
|
| 2596 |
+
"grad_norm": 2.453866507102359,
|
| 2597 |
+
"learning_rate": 1.5998937339039889e-06,
|
| 2598 |
+
"loss": 0.3512,
|
| 2599 |
+
"step": 1845
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.7663628831814416,
|
| 2603 |
+
"grad_norm": 2.373346324937054,
|
| 2604 |
+
"learning_rate": 1.5734705439636017e-06,
|
| 2605 |
+
"loss": 0.351,
|
| 2606 |
+
"step": 1850
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.7684341342170671,
|
| 2610 |
+
"grad_norm": 2.293953470114846,
|
| 2611 |
+
"learning_rate": 1.5472265683350397e-06,
|
| 2612 |
+
"loss": 0.3566,
|
| 2613 |
+
"step": 1855
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.7705053852526926,
|
| 2617 |
+
"grad_norm": 2.334589469178464,
|
| 2618 |
+
"learning_rate": 1.5211631796304721e-06,
|
| 2619 |
+
"loss": 0.344,
|
| 2620 |
+
"step": 1860
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.7725766362883182,
|
| 2624 |
+
"grad_norm": 2.1977473111231274,
|
| 2625 |
+
"learning_rate": 1.495281741017016e-06,
|
| 2626 |
+
"loss": 0.3466,
|
| 2627 |
+
"step": 1865
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.7746478873239436,
|
| 2631 |
+
"grad_norm": 2.447667991652811,
|
| 2632 |
+
"learning_rate": 1.46958360614543e-06,
|
| 2633 |
+
"loss": 0.3458,
|
| 2634 |
+
"step": 1870
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.7767191383595692,
|
| 2638 |
+
"grad_norm": 2.4293930001790343,
|
| 2639 |
+
"learning_rate": 1.4440701190793278e-06,
|
| 2640 |
+
"loss": 0.3454,
|
| 2641 |
+
"step": 1875
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.7787903893951947,
|
| 2645 |
+
"grad_norm": 2.4169714631028443,
|
| 2646 |
+
"learning_rate": 1.4187426142248723e-06,
|
| 2647 |
+
"loss": 0.3444,
|
| 2648 |
+
"step": 1880
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.7808616404308202,
|
| 2652 |
+
"grad_norm": 2.347141749048018,
|
| 2653 |
+
"learning_rate": 1.3936024162609897e-06,
|
| 2654 |
+
"loss": 0.3433,
|
| 2655 |
+
"step": 1885
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.7829328914664457,
|
| 2659 |
+
"grad_norm": 2.3150017921664934,
|
| 2660 |
+
"learning_rate": 1.3686508400700787e-06,
|
| 2661 |
+
"loss": 0.3424,
|
| 2662 |
+
"step": 1890
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.7850041425020713,
|
| 2666 |
+
"grad_norm": 2.433953451204477,
|
| 2667 |
+
"learning_rate": 1.3438891906692447e-06,
|
| 2668 |
+
"loss": 0.3402,
|
| 2669 |
+
"step": 1895
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.7870753935376967,
|
| 2673 |
+
"grad_norm": 2.4345942045059923,
|
| 2674 |
+
"learning_rate": 1.3193187631420462e-06,
|
| 2675 |
+
"loss": 0.3559,
|
| 2676 |
+
"step": 1900
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.7891466445733223,
|
| 2680 |
+
"grad_norm": 2.385776976242536,
|
| 2681 |
+
"learning_rate": 1.2949408425707566e-06,
|
| 2682 |
+
"loss": 0.3409,
|
| 2683 |
+
"step": 1905
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.7912178956089478,
|
| 2687 |
+
"grad_norm": 2.284820681480646,
|
| 2688 |
+
"learning_rate": 1.2707567039691505e-06,
|
| 2689 |
+
"loss": 0.3323,
|
| 2690 |
+
"step": 1910
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.7932891466445733,
|
| 2694 |
+
"grad_norm": 2.373599316713983,
|
| 2695 |
+
"learning_rate": 1.2467676122158224e-06,
|
| 2696 |
+
"loss": 0.3431,
|
| 2697 |
+
"step": 1915
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.7953603976801988,
|
| 2701 |
+
"grad_norm": 2.277308417116025,
|
| 2702 |
+
"learning_rate": 1.222974821988024e-06,
|
| 2703 |
+
"loss": 0.334,
|
| 2704 |
+
"step": 1920
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.7974316487158244,
|
| 2708 |
+
"grad_norm": 2.3778447080117577,
|
| 2709 |
+
"learning_rate": 1.1993795776960498e-06,
|
| 2710 |
+
"loss": 0.3326,
|
| 2711 |
+
"step": 1925
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.7995028997514498,
|
| 2715 |
+
"grad_norm": 2.452346766698329,
|
| 2716 |
+
"learning_rate": 1.1759831134181504e-06,
|
| 2717 |
+
"loss": 0.3317,
|
| 2718 |
+
"step": 1930
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.8015741507870754,
|
| 2722 |
+
"grad_norm": 2.4247770647591085,
|
| 2723 |
+
"learning_rate": 1.1527866528359805e-06,
|
| 2724 |
+
"loss": 0.3474,
|
| 2725 |
+
"step": 1935
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.8036454018227009,
|
| 2729 |
+
"grad_norm": 2.3076639437103292,
|
| 2730 |
+
"learning_rate": 1.1297914091706086e-06,
|
| 2731 |
+
"loss": 0.3423,
|
| 2732 |
+
"step": 1940
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.8057166528583264,
|
| 2736 |
+
"grad_norm": 2.346460611096731,
|
| 2737 |
+
"learning_rate": 1.1069985851190524e-06,
|
| 2738 |
+
"loss": 0.3373,
|
| 2739 |
+
"step": 1945
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.8077879038939519,
|
| 2743 |
+
"grad_norm": 2.297664318905517,
|
| 2744 |
+
"learning_rate": 1.0844093727913868e-06,
|
| 2745 |
+
"loss": 0.3452,
|
| 2746 |
+
"step": 1950
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.8098591549295775,
|
| 2750 |
+
"grad_norm": 2.3994880509321823,
|
| 2751 |
+
"learning_rate": 1.062024953648384e-06,
|
| 2752 |
+
"loss": 0.3303,
|
| 2753 |
+
"step": 1955
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.8119304059652029,
|
| 2757 |
+
"grad_norm": 2.3943126094691736,
|
| 2758 |
+
"learning_rate": 1.039846498439727e-06,
|
| 2759 |
+
"loss": 0.337,
|
| 2760 |
+
"step": 1960
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.8140016570008285,
|
| 2764 |
+
"grad_norm": 2.394969585780868,
|
| 2765 |
+
"learning_rate": 1.0178751671427755e-06,
|
| 2766 |
+
"loss": 0.333,
|
| 2767 |
+
"step": 1965
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.816072908036454,
|
| 2771 |
+
"grad_norm": 2.2584619131198327,
|
| 2772 |
+
"learning_rate": 9.961121089018933e-07,
|
| 2773 |
+
"loss": 0.343,
|
| 2774 |
+
"step": 1970
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.8181441590720795,
|
| 2778 |
+
"grad_norm": 2.5188681037163305,
|
| 2779 |
+
"learning_rate": 9.745584619683524e-07,
|
| 2780 |
+
"loss": 0.3386,
|
| 2781 |
+
"step": 1975
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.820215410107705,
|
| 2785 |
+
"grad_norm": 2.3562413171777856,
|
| 2786 |
+
"learning_rate": 9.532153536407923e-07,
|
| 2787 |
+
"loss": 0.3415,
|
| 2788 |
+
"step": 1980
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.8222866611433306,
|
| 2792 |
+
"grad_norm": 2.4322075663751153,
|
| 2793 |
+
"learning_rate": 9.320839002062682e-07,
|
| 2794 |
+
"loss": 0.3301,
|
| 2795 |
+
"step": 1985
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.824357912178956,
|
| 2799 |
+
"grad_norm": 2.4874846740356804,
|
| 2800 |
+
"learning_rate": 9.111652068818621e-07,
|
| 2801 |
+
"loss": 0.3433,
|
| 2802 |
+
"step": 1990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.8264291632145816,
|
| 2806 |
+
"grad_norm": 2.312127292110117,
|
| 2807 |
+
"learning_rate": 8.904603677568785e-07,
|
| 2808 |
+
"loss": 0.3331,
|
| 2809 |
+
"step": 1995
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.8285004142502072,
|
| 2813 |
+
"grad_norm": 2.607402264179206,
|
| 2814 |
+
"learning_rate": 8.699704657356195e-07,
|
| 2815 |
+
"loss": 0.3333,
|
| 2816 |
+
"step": 2000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.8305716652858326,
|
| 2820 |
+
"grad_norm": 2.493751130019948,
|
| 2821 |
+
"learning_rate": 8.496965724807516e-07,
|
| 2822 |
+
"loss": 0.3362,
|
| 2823 |
+
"step": 2005
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.8326429163214581,
|
| 2827 |
+
"grad_norm": 2.351799520241766,
|
| 2828 |
+
"learning_rate": 8.296397483572515e-07,
|
| 2829 |
+
"loss": 0.3361,
|
| 2830 |
+
"step": 2010
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.8347141673570837,
|
| 2834 |
+
"grad_norm": 2.3347796942844457,
|
| 2835 |
+
"learning_rate": 8.098010423769503e-07,
|
| 2836 |
+
"loss": 0.3427,
|
| 2837 |
+
"step": 2015
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.8367854183927091,
|
| 2841 |
+
"grad_norm": 2.416514153196825,
|
| 2842 |
+
"learning_rate": 7.901814921436624e-07,
|
| 2843 |
+
"loss": 0.3373,
|
| 2844 |
+
"step": 2020
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.8388566694283347,
|
| 2848 |
+
"grad_norm": 2.27725091962227,
|
| 2849 |
+
"learning_rate": 7.70782123798921e-07,
|
| 2850 |
+
"loss": 0.3375,
|
| 2851 |
+
"step": 2025
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.8409279204639603,
|
| 2855 |
+
"grad_norm": 2.24276056194008,
|
| 2856 |
+
"learning_rate": 7.516039519683105e-07,
|
| 2857 |
+
"loss": 0.3382,
|
| 2858 |
+
"step": 2030
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.8429991714995857,
|
| 2862 |
+
"grad_norm": 2.4106229979629545,
|
| 2863 |
+
"learning_rate": 7.326479797083963e-07,
|
| 2864 |
+
"loss": 0.3288,
|
| 2865 |
+
"step": 2035
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.8450704225352113,
|
| 2869 |
+
"grad_norm": 2.281186671589003,
|
| 2870 |
+
"learning_rate": 7.139151984542636e-07,
|
| 2871 |
+
"loss": 0.3269,
|
| 2872 |
+
"step": 2040
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.8471416735708368,
|
| 2876 |
+
"grad_norm": 2.441879212516951,
|
| 2877 |
+
"learning_rate": 6.954065879676653e-07,
|
| 2878 |
+
"loss": 0.3327,
|
| 2879 |
+
"step": 2045
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.8492129246064622,
|
| 2883 |
+
"grad_norm": 2.460206403334868,
|
| 2884 |
+
"learning_rate": 6.771231162857722e-07,
|
| 2885 |
+
"loss": 0.3375,
|
| 2886 |
+
"step": 2050
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.8512841756420878,
|
| 2890 |
+
"grad_norm": 2.288538904206948,
|
| 2891 |
+
"learning_rate": 6.590657396705525e-07,
|
| 2892 |
+
"loss": 0.3163,
|
| 2893 |
+
"step": 2055
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.8533554266777134,
|
| 2897 |
+
"grad_norm": 2.3886559210526426,
|
| 2898 |
+
"learning_rate": 6.412354025587509e-07,
|
| 2899 |
+
"loss": 0.3356,
|
| 2900 |
+
"step": 2060
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.8554266777133389,
|
| 2904 |
+
"grad_norm": 2.2517960802152337,
|
| 2905 |
+
"learning_rate": 6.236330375124921e-07,
|
| 2906 |
+
"loss": 0.341,
|
| 2907 |
+
"step": 2065
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.8574979287489644,
|
| 2911 |
+
"grad_norm": 2.33394312378329,
|
| 2912 |
+
"learning_rate": 6.062595651705111e-07,
|
| 2913 |
+
"loss": 0.3262,
|
| 2914 |
+
"step": 2070
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.8595691797845899,
|
| 2918 |
+
"grad_norm": 2.4559361971290077,
|
| 2919 |
+
"learning_rate": 5.891158941999959e-07,
|
| 2920 |
+
"loss": 0.3198,
|
| 2921 |
+
"step": 2075
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.8616404308202155,
|
| 2925 |
+
"grad_norm": 2.367468825376486,
|
| 2926 |
+
"learning_rate": 5.722029212490666e-07,
|
| 2927 |
+
"loss": 0.3264,
|
| 2928 |
+
"step": 2080
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.8637116818558409,
|
| 2932 |
+
"grad_norm": 2.4147043980300955,
|
| 2933 |
+
"learning_rate": 5.555215308998779e-07,
|
| 2934 |
+
"loss": 0.3352,
|
| 2935 |
+
"step": 2085
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.8657829328914665,
|
| 2939 |
+
"grad_norm": 2.4293612156766815,
|
| 2940 |
+
"learning_rate": 5.390725956223531e-07,
|
| 2941 |
+
"loss": 0.3209,
|
| 2942 |
+
"step": 2090
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.867854183927092,
|
| 2946 |
+
"grad_norm": 2.409157285292527,
|
| 2947 |
+
"learning_rate": 5.22856975728554e-07,
|
| 2948 |
+
"loss": 0.3249,
|
| 2949 |
+
"step": 2095
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.8699254349627175,
|
| 2953 |
+
"grad_norm": 2.233849266540862,
|
| 2954 |
+
"learning_rate": 5.068755193276798e-07,
|
| 2955 |
+
"loss": 0.3216,
|
| 2956 |
+
"step": 2100
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.871996685998343,
|
| 2960 |
+
"grad_norm": 2.3098128897942076,
|
| 2961 |
+
"learning_rate": 4.911290622817161e-07,
|
| 2962 |
+
"loss": 0.3133,
|
| 2963 |
+
"step": 2105
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.8740679370339686,
|
| 2967 |
+
"grad_norm": 2.382543737963039,
|
| 2968 |
+
"learning_rate": 4.756184281617121e-07,
|
| 2969 |
+
"loss": 0.3212,
|
| 2970 |
+
"step": 2110
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.876139188069594,
|
| 2974 |
+
"grad_norm": 2.67090524996084,
|
| 2975 |
+
"learning_rate": 4.6034442820471037e-07,
|
| 2976 |
+
"loss": 0.3304,
|
| 2977 |
+
"step": 2115
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.8782104391052196,
|
| 2981 |
+
"grad_norm": 2.342709610038623,
|
| 2982 |
+
"learning_rate": 4.4530786127131575e-07,
|
| 2983 |
+
"loss": 0.324,
|
| 2984 |
+
"step": 2120
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.8802816901408451,
|
| 2988 |
+
"grad_norm": 2.4993138899361953,
|
| 2989 |
+
"learning_rate": 4.305095138039106e-07,
|
| 2990 |
+
"loss": 0.328,
|
| 2991 |
+
"step": 2125
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.8823529411764706,
|
| 2995 |
+
"grad_norm": 2.3696872859071183,
|
| 2996 |
+
"learning_rate": 4.159501597855287e-07,
|
| 2997 |
+
"loss": 0.3258,
|
| 2998 |
+
"step": 2130
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.8844241922120961,
|
| 3002 |
+
"grad_norm": 2.4387312234421707,
|
| 3003 |
+
"learning_rate": 4.0163056069936757e-07,
|
| 3004 |
+
"loss": 0.3241,
|
| 3005 |
+
"step": 2135
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.8864954432477217,
|
| 3009 |
+
"grad_norm": 2.3989145869426536,
|
| 3010 |
+
"learning_rate": 3.8755146548896784e-07,
|
| 3011 |
+
"loss": 0.3275,
|
| 3012 |
+
"step": 2140
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.8885666942833471,
|
| 3016 |
+
"grad_norm": 2.4928751733374326,
|
| 3017 |
+
"learning_rate": 3.737136105190337e-07,
|
| 3018 |
+
"loss": 0.3293,
|
| 3019 |
+
"step": 2145
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.8906379453189727,
|
| 3023 |
+
"grad_norm": 2.2972269705841737,
|
| 3024 |
+
"learning_rate": 3.6011771953693044e-07,
|
| 3025 |
+
"loss": 0.3283,
|
| 3026 |
+
"step": 2150
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.8927091963545982,
|
| 3030 |
+
"grad_norm": 2.297085637993506,
|
| 3031 |
+
"learning_rate": 3.4676450363481937e-07,
|
| 3032 |
+
"loss": 0.3243,
|
| 3033 |
+
"step": 2155
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.8947804473902237,
|
| 3037 |
+
"grad_norm": 2.4352041283350756,
|
| 3038 |
+
"learning_rate": 3.336546612124758e-07,
|
| 3039 |
+
"loss": 0.3229,
|
| 3040 |
+
"step": 2160
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.8968516984258492,
|
| 3044 |
+
"grad_norm": 2.4453453128997693,
|
| 3045 |
+
"learning_rate": 3.20788877940757e-07,
|
| 3046 |
+
"loss": 0.3267,
|
| 3047 |
+
"step": 2165
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.8989229494614748,
|
| 3051 |
+
"grad_norm": 2.497196691348891,
|
| 3052 |
+
"learning_rate": 3.081678267257404e-07,
|
| 3053 |
+
"loss": 0.3245,
|
| 3054 |
+
"step": 2170
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.9009942004971002,
|
| 3058 |
+
"grad_norm": 2.294241552326414,
|
| 3059 |
+
"learning_rate": 2.9579216767352815e-07,
|
| 3060 |
+
"loss": 0.3174,
|
| 3061 |
+
"step": 2175
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.9030654515327258,
|
| 3065 |
+
"grad_norm": 2.5123348209381704,
|
| 3066 |
+
"learning_rate": 2.836625480557265e-07,
|
| 3067 |
+
"loss": 0.3185,
|
| 3068 |
+
"step": 2180
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.9051367025683513,
|
| 3072 |
+
"grad_norm": 2.253684937067122,
|
| 3073 |
+
"learning_rate": 2.7177960227558863e-07,
|
| 3074 |
+
"loss": 0.3144,
|
| 3075 |
+
"step": 2185
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.9072079536039768,
|
| 3079 |
+
"grad_norm": 2.1977246969912723,
|
| 3080 |
+
"learning_rate": 2.601439518348331e-07,
|
| 3081 |
+
"loss": 0.3265,
|
| 3082 |
+
"step": 2190
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.9092792046396023,
|
| 3086 |
+
"grad_norm": 2.3594137667020303,
|
| 3087 |
+
"learning_rate": 2.487562053011422e-07,
|
| 3088 |
+
"loss": 0.3182,
|
| 3089 |
+
"step": 2195
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 0.9113504556752279,
|
| 3093 |
+
"grad_norm": 2.4721070329083954,
|
| 3094 |
+
"learning_rate": 2.376169582763288e-07,
|
| 3095 |
+
"loss": 0.3204,
|
| 3096 |
+
"step": 2200
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 0.9134217067108533,
|
| 3100 |
+
"grad_norm": 2.5655843945707306,
|
| 3101 |
+
"learning_rate": 2.2672679336518789e-07,
|
| 3102 |
+
"loss": 0.3286,
|
| 3103 |
+
"step": 2205
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 0.9154929577464789,
|
| 3107 |
+
"grad_norm": 2.2713717970234155,
|
| 3108 |
+
"learning_rate": 2.1608628014502364e-07,
|
| 3109 |
+
"loss": 0.3186,
|
| 3110 |
+
"step": 2210
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 0.9175642087821044,
|
| 3114 |
+
"grad_norm": 2.319604155676886,
|
| 3115 |
+
"learning_rate": 2.0569597513586004e-07,
|
| 3116 |
+
"loss": 0.3189,
|
| 3117 |
+
"step": 2215
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 0.9196354598177299,
|
| 3121 |
+
"grad_norm": 2.582435571084389,
|
| 3122 |
+
"learning_rate": 1.955564217713335e-07,
|
| 3123 |
+
"loss": 0.3248,
|
| 3124 |
+
"step": 2220
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.9217067108533554,
|
| 3128 |
+
"grad_norm": 2.4094176417675532,
|
| 3129 |
+
"learning_rate": 1.8566815037026897e-07,
|
| 3130 |
+
"loss": 0.3371,
|
| 3131 |
+
"step": 2225
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 0.923777961888981,
|
| 3135 |
+
"grad_norm": 2.4509597114145913,
|
| 3136 |
+
"learning_rate": 1.7603167810894662e-07,
|
| 3137 |
+
"loss": 0.3203,
|
| 3138 |
+
"step": 2230
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 0.9258492129246064,
|
| 3142 |
+
"grad_norm": 2.363051351708879,
|
| 3143 |
+
"learning_rate": 1.6664750899404892e-07,
|
| 3144 |
+
"loss": 0.3101,
|
| 3145 |
+
"step": 2235
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 0.927920463960232,
|
| 3149 |
+
"grad_norm": 2.420545920556074,
|
| 3150 |
+
"learning_rate": 1.5751613383630128e-07,
|
| 3151 |
+
"loss": 0.3218,
|
| 3152 |
+
"step": 2240
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 0.9299917149958575,
|
| 3156 |
+
"grad_norm": 2.3859750154616397,
|
| 3157 |
+
"learning_rate": 1.4863803022480362e-07,
|
| 3158 |
+
"loss": 0.3118,
|
| 3159 |
+
"step": 2245
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 0.932062966031483,
|
| 3163 |
+
"grad_norm": 2.1948107719137955,
|
| 3164 |
+
"learning_rate": 1.4001366250204762e-07,
|
| 3165 |
+
"loss": 0.3109,
|
| 3166 |
+
"step": 2250
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.9341342170671085,
|
| 3170 |
+
"grad_norm": 2.27270194473527,
|
| 3171 |
+
"learning_rate": 1.3164348173963392e-07,
|
| 3172 |
+
"loss": 0.3193,
|
| 3173 |
+
"step": 2255
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 0.9362054681027341,
|
| 3177 |
+
"grad_norm": 2.3657310934258833,
|
| 3178 |
+
"learning_rate": 1.235279257146804e-07,
|
| 3179 |
+
"loss": 0.3203,
|
| 3180 |
+
"step": 2260
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 0.9382767191383595,
|
| 3184 |
+
"grad_norm": 2.338711000044841,
|
| 3185 |
+
"learning_rate": 1.1566741888692168e-07,
|
| 3186 |
+
"loss": 0.3177,
|
| 3187 |
+
"step": 2265
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 0.9403479701739851,
|
| 3191 |
+
"grad_norm": 2.5901775745271376,
|
| 3192 |
+
"learning_rate": 1.080623723765134e-07,
|
| 3193 |
+
"loss": 0.3178,
|
| 3194 |
+
"step": 2270
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 0.9424192212096106,
|
| 3198 |
+
"grad_norm": 2.540579688066845,
|
| 3199 |
+
"learning_rate": 1.0071318394252849e-07,
|
| 3200 |
+
"loss": 0.3199,
|
| 3201 |
+
"step": 2275
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 0.9444904722452361,
|
| 3205 |
+
"grad_norm": 2.46660297908585,
|
| 3206 |
+
"learning_rate": 9.362023796215036e-08,
|
| 3207 |
+
"loss": 0.3183,
|
| 3208 |
+
"step": 2280
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 0.9465617232808616,
|
| 3212 |
+
"grad_norm": 2.56480438463466,
|
| 3213 |
+
"learning_rate": 8.678390541057512e-08,
|
| 3214 |
+
"loss": 0.3127,
|
| 3215 |
+
"step": 2285
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 0.9486329743164872,
|
| 3219 |
+
"grad_norm": 2.4346321810323923,
|
| 3220 |
+
"learning_rate": 8.020454384160437e-08,
|
| 3221 |
+
"loss": 0.3217,
|
| 3222 |
+
"step": 2290
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 0.9507042253521126,
|
| 3226 |
+
"grad_norm": 2.4276401722046734,
|
| 3227 |
+
"learning_rate": 7.388249736894615e-08,
|
| 3228 |
+
"loss": 0.3095,
|
| 3229 |
+
"step": 2295
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 0.9527754763877382,
|
| 3233 |
+
"grad_norm": 2.329219841891159,
|
| 3234 |
+
"learning_rate": 6.78180966482156e-08,
|
| 3235 |
+
"loss": 0.319,
|
| 3236 |
+
"step": 2300
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 0.9548467274233637,
|
| 3240 |
+
"grad_norm": 2.6042724587430257,
|
| 3241 |
+
"learning_rate": 6.201165885964311e-08,
|
| 3242 |
+
"loss": 0.317,
|
| 3243 |
+
"step": 2305
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 0.9569179784589892,
|
| 3247 |
+
"grad_norm": 2.2303445759455967,
|
| 3248 |
+
"learning_rate": 5.646348769148491e-08,
|
| 3249 |
+
"loss": 0.3217,
|
| 3250 |
+
"step": 2310
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 0.9589892294946147,
|
| 3254 |
+
"grad_norm": 2.4607182859035728,
|
| 3255 |
+
"learning_rate": 5.117387332413737e-08,
|
| 3256 |
+
"loss": 0.3379,
|
| 3257 |
+
"step": 2315
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 0.9610604805302403,
|
| 3261 |
+
"grad_norm": 2.3872878040403775,
|
| 3262 |
+
"learning_rate": 4.6143092414961396e-08,
|
| 3263 |
+
"loss": 0.3232,
|
| 3264 |
+
"step": 2320
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 0.9631317315658657,
|
| 3268 |
+
"grad_norm": 2.369941738658177,
|
| 3269 |
+
"learning_rate": 4.1371408083815126e-08,
|
| 3270 |
+
"loss": 0.313,
|
| 3271 |
+
"step": 2325
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 0.9652029826014913,
|
| 3275 |
+
"grad_norm": 2.413042323401412,
|
| 3276 |
+
"learning_rate": 3.685906989928656e-08,
|
| 3277 |
+
"loss": 0.3324,
|
| 3278 |
+
"step": 2330
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 0.9672742336371168,
|
| 3282 |
+
"grad_norm": 2.3918886450019747,
|
| 3283 |
+
"learning_rate": 3.2606313865646276e-08,
|
| 3284 |
+
"loss": 0.3213,
|
| 3285 |
+
"step": 2335
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 0.9693454846727423,
|
| 3289 |
+
"grad_norm": 2.212556778563815,
|
| 3290 |
+
"learning_rate": 2.861336241050061e-08,
|
| 3291 |
+
"loss": 0.3136,
|
| 3292 |
+
"step": 2340
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 0.9714167357083678,
|
| 3296 |
+
"grad_norm": 2.3753322114534514,
|
| 3297 |
+
"learning_rate": 2.488042437315985e-08,
|
| 3298 |
+
"loss": 0.3196,
|
| 3299 |
+
"step": 2345
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 0.9734879867439934,
|
| 3303 |
+
"grad_norm": 2.635822549535026,
|
| 3304 |
+
"learning_rate": 2.1407694993714755e-08,
|
| 3305 |
+
"loss": 0.3201,
|
| 3306 |
+
"step": 2350
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 0.9755592377796188,
|
| 3310 |
+
"grad_norm": 2.4223063593868774,
|
| 3311 |
+
"learning_rate": 1.8195355902824174e-08,
|
| 3312 |
+
"loss": 0.3177,
|
| 3313 |
+
"step": 2355
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 0.9776304888152444,
|
| 3317 |
+
"grad_norm": 2.5265938379888864,
|
| 3318 |
+
"learning_rate": 1.5243575112218744e-08,
|
| 3319 |
+
"loss": 0.3208,
|
| 3320 |
+
"step": 2360
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 0.97970173985087,
|
| 3324 |
+
"grad_norm": 2.344598334063139,
|
| 3325 |
+
"learning_rate": 1.2552507005909042e-08,
|
| 3326 |
+
"loss": 0.3103,
|
| 3327 |
+
"step": 2365
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 0.9817729908864954,
|
| 3331 |
+
"grad_norm": 2.3574452536812145,
|
| 3332 |
+
"learning_rate": 1.0122292332114814e-08,
|
| 3333 |
+
"loss": 0.3191,
|
| 3334 |
+
"step": 2370
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 0.9838442419221209,
|
| 3338 |
+
"grad_norm": 2.3406496363432128,
|
| 3339 |
+
"learning_rate": 7.953058195900864e-09,
|
| 3340 |
+
"loss": 0.317,
|
| 3341 |
+
"step": 2375
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 0.9859154929577465,
|
| 3345 |
+
"grad_norm": 2.2200414240286155,
|
| 3346 |
+
"learning_rate": 6.044918052531268e-09,
|
| 3347 |
+
"loss": 0.3079,
|
| 3348 |
+
"step": 2380
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 0.987986743993372,
|
| 3352 |
+
"grad_norm": 2.345671675412989,
|
| 3353 |
+
"learning_rate": 4.397971701533554e-09,
|
| 3354 |
+
"loss": 0.3113,
|
| 3355 |
+
"step": 2385
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.9900579950289975,
|
| 3359 |
+
"grad_norm": 2.4132441952700687,
|
| 3360 |
+
"learning_rate": 3.0123052814812203e-09,
|
| 3361 |
+
"loss": 0.3207,
|
| 3362 |
+
"step": 2390
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.992129246064623,
|
| 3366 |
+
"grad_norm": 2.599715525455967,
|
| 3367 |
+
"learning_rate": 1.887991265486222e-09,
|
| 3368 |
+
"loss": 0.318,
|
| 3369 |
+
"step": 2395
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.9942004971002486,
|
| 3373 |
+
"grad_norm": 2.371853977706459,
|
| 3374 |
+
"learning_rate": 1.025088457409229e-09,
|
| 3375 |
+
"loss": 0.3283,
|
| 3376 |
+
"step": 2400
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 0.996271748135874,
|
| 3380 |
+
"grad_norm": 2.4606089652599876,
|
| 3381 |
+
"learning_rate": 4.2364198878597216e-10,
|
| 3382 |
+
"loss": 0.3211,
|
| 3383 |
+
"step": 2405
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.9983429991714996,
|
| 3387 |
+
"grad_norm": 2.471628534592773,
|
| 3388 |
+
"learning_rate": 8.368331646302353e-11,
|
| 3389 |
+
"loss": 0.3281,
|
| 3390 |
+
"step": 2410
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 1.0,
|
| 3394 |
+
"eval_runtime": 2.6867,
|
| 3395 |
+
"eval_samples_per_second": 3.722,
|
| 3396 |
+
"eval_steps_per_second": 1.117,
|
| 3397 |
+
"step": 2414
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 1.0,
|
| 3401 |
+
"step": 2414,
|
| 3402 |
+
"total_flos": 252721244405760.0,
|
| 3403 |
+
"train_loss": 0.5161569981705382,
|
| 3404 |
+
"train_runtime": 26044.7738,
|
| 3405 |
+
"train_samples_per_second": 1.483,
|
| 3406 |
+
"train_steps_per_second": 0.093
|
| 3407 |
+
}
|
| 3408 |
+
],
|
| 3409 |
+
"logging_steps": 5,
|
| 3410 |
+
"max_steps": 2414,
|
| 3411 |
+
"num_input_tokens_seen": 0,
|
| 3412 |
+
"num_train_epochs": 1,
|
| 3413 |
+
"save_steps": 100,
|
| 3414 |
+
"stateful_callbacks": {
|
| 3415 |
+
"TrainerControl": {
|
| 3416 |
+
"args": {
|
| 3417 |
+
"should_epoch_stop": false,
|
| 3418 |
+
"should_evaluate": false,
|
| 3419 |
+
"should_log": false,
|
| 3420 |
+
"should_save": true,
|
| 3421 |
+
"should_training_stop": true
|
| 3422 |
+
},
|
| 3423 |
+
"attributes": {}
|
| 3424 |
+
}
|
| 3425 |
+
},
|
| 3426 |
+
"total_flos": 252721244405760.0,
|
| 3427 |
+
"train_batch_size": 4,
|
| 3428 |
+
"trial_name": null,
|
| 3429 |
+
"trial_params": null
|
| 3430 |
+
}
|