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 +3437 -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_p0.02_seed42_level2_raremixbatch16
|
| 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_p0.02_seed42_level2_raremixbatch16
|
| 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": 252825934233600.0,
|
| 4 |
+
"train_loss": 0.5146465574979289,
|
| 5 |
+
"train_runtime": 25682.4245,
|
| 6 |
+
"train_samples": 98881,
|
| 7 |
+
"train_samples_per_second": 1.504,
|
| 8 |
+
"train_steps_per_second": 0.094
|
| 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": 252825934233600.0,
|
| 4 |
+
"train_loss": 0.5146465574979289,
|
| 5 |
+
"train_runtime": 25682.4245,
|
| 6 |
+
"train_samples": 98881,
|
| 7 |
+
"train_samples_per_second": 1.504,
|
| 8 |
+
"train_steps_per_second": 0.094
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3437 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2415,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.00041407867494824016,
|
| 13 |
+
"grad_norm": 24.186954732535973,
|
| 14 |
+
"learning_rate": 4.132231404958678e-08,
|
| 15 |
+
"loss": 1.427,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.002070393374741201,
|
| 20 |
+
"grad_norm": 24.196130418494608,
|
| 21 |
+
"learning_rate": 2.066115702479339e-07,
|
| 22 |
+
"loss": 1.4047,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.004140786749482402,
|
| 27 |
+
"grad_norm": 16.3762981570342,
|
| 28 |
+
"learning_rate": 4.132231404958678e-07,
|
| 29 |
+
"loss": 1.3799,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.006211180124223602,
|
| 34 |
+
"grad_norm": 8.425206725694215,
|
| 35 |
+
"learning_rate": 6.198347107438018e-07,
|
| 36 |
+
"loss": 1.2594,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.008281573498964804,
|
| 41 |
+
"grad_norm": 10.05371068200267,
|
| 42 |
+
"learning_rate": 8.264462809917356e-07,
|
| 43 |
+
"loss": 1.1298,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.010351966873706004,
|
| 48 |
+
"grad_norm": 4.721175355009842,
|
| 49 |
+
"learning_rate": 1.0330578512396695e-06,
|
| 50 |
+
"loss": 1.0296,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.012422360248447204,
|
| 55 |
+
"grad_norm": 3.3182689234643394,
|
| 56 |
+
"learning_rate": 1.2396694214876035e-06,
|
| 57 |
+
"loss": 0.9653,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.014492753623188406,
|
| 62 |
+
"grad_norm": 3.504456631337904,
|
| 63 |
+
"learning_rate": 1.4462809917355372e-06,
|
| 64 |
+
"loss": 0.9407,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.016563146997929608,
|
| 69 |
+
"grad_norm": 2.9960316182859783,
|
| 70 |
+
"learning_rate": 1.6528925619834712e-06,
|
| 71 |
+
"loss": 0.9251,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.018633540372670808,
|
| 76 |
+
"grad_norm": 3.1240932779362622,
|
| 77 |
+
"learning_rate": 1.859504132231405e-06,
|
| 78 |
+
"loss": 0.9188,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.020703933747412008,
|
| 83 |
+
"grad_norm": 3.2513730400449203,
|
| 84 |
+
"learning_rate": 2.066115702479339e-06,
|
| 85 |
+
"loss": 0.8944,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.022774327122153208,
|
| 90 |
+
"grad_norm": 3.1786210930130308,
|
| 91 |
+
"learning_rate": 2.2727272727272728e-06,
|
| 92 |
+
"loss": 0.9026,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.024844720496894408,
|
| 97 |
+
"grad_norm": 3.1076677023399464,
|
| 98 |
+
"learning_rate": 2.479338842975207e-06,
|
| 99 |
+
"loss": 0.9026,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.026915113871635612,
|
| 104 |
+
"grad_norm": 3.0823234166313824,
|
| 105 |
+
"learning_rate": 2.6859504132231405e-06,
|
| 106 |
+
"loss": 0.8895,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.028985507246376812,
|
| 111 |
+
"grad_norm": 3.005596270571828,
|
| 112 |
+
"learning_rate": 2.8925619834710743e-06,
|
| 113 |
+
"loss": 0.8802,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.031055900621118012,
|
| 118 |
+
"grad_norm": 3.3025364718701056,
|
| 119 |
+
"learning_rate": 3.0991735537190086e-06,
|
| 120 |
+
"loss": 0.8779,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.033126293995859216,
|
| 125 |
+
"grad_norm": 2.9484146765559256,
|
| 126 |
+
"learning_rate": 3.3057851239669424e-06,
|
| 127 |
+
"loss": 0.8556,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.035196687370600416,
|
| 132 |
+
"grad_norm": 2.991384604651969,
|
| 133 |
+
"learning_rate": 3.5123966942148763e-06,
|
| 134 |
+
"loss": 0.8768,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.037267080745341616,
|
| 139 |
+
"grad_norm": 3.2763200828654044,
|
| 140 |
+
"learning_rate": 3.71900826446281e-06,
|
| 141 |
+
"loss": 0.8693,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.039337474120082816,
|
| 146 |
+
"grad_norm": 2.9849839023650913,
|
| 147 |
+
"learning_rate": 3.925619834710744e-06,
|
| 148 |
+
"loss": 0.8663,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.041407867494824016,
|
| 153 |
+
"grad_norm": 3.0799688517985406,
|
| 154 |
+
"learning_rate": 4.132231404958678e-06,
|
| 155 |
+
"loss": 0.8449,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.043478260869565216,
|
| 160 |
+
"grad_norm": 3.164277927549159,
|
| 161 |
+
"learning_rate": 4.338842975206612e-06,
|
| 162 |
+
"loss": 0.8556,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.045548654244306416,
|
| 167 |
+
"grad_norm": 3.054698785813227,
|
| 168 |
+
"learning_rate": 4.5454545454545455e-06,
|
| 169 |
+
"loss": 0.8406,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.047619047619047616,
|
| 174 |
+
"grad_norm": 3.1517792822498296,
|
| 175 |
+
"learning_rate": 4.75206611570248e-06,
|
| 176 |
+
"loss": 0.845,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.049689440993788817,
|
| 181 |
+
"grad_norm": 3.1935189527896988,
|
| 182 |
+
"learning_rate": 4.958677685950414e-06,
|
| 183 |
+
"loss": 0.8501,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.051759834368530024,
|
| 188 |
+
"grad_norm": 3.2963890157762696,
|
| 189 |
+
"learning_rate": 5.165289256198347e-06,
|
| 190 |
+
"loss": 0.8274,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.053830227743271224,
|
| 195 |
+
"grad_norm": 3.1324161276204303,
|
| 196 |
+
"learning_rate": 5.371900826446281e-06,
|
| 197 |
+
"loss": 0.8291,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.055900621118012424,
|
| 202 |
+
"grad_norm": 3.010837639139227,
|
| 203 |
+
"learning_rate": 5.578512396694216e-06,
|
| 204 |
+
"loss": 0.8415,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.057971014492753624,
|
| 209 |
+
"grad_norm": 3.0560184585554295,
|
| 210 |
+
"learning_rate": 5.785123966942149e-06,
|
| 211 |
+
"loss": 0.8273,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.060041407867494824,
|
| 216 |
+
"grad_norm": 3.1627396501144025,
|
| 217 |
+
"learning_rate": 5.991735537190083e-06,
|
| 218 |
+
"loss": 0.8352,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.062111801242236024,
|
| 223 |
+
"grad_norm": 3.3250094405823707,
|
| 224 |
+
"learning_rate": 6.198347107438017e-06,
|
| 225 |
+
"loss": 0.8289,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.06418219461697723,
|
| 230 |
+
"grad_norm": 3.2642262323401003,
|
| 231 |
+
"learning_rate": 6.404958677685951e-06,
|
| 232 |
+
"loss": 0.8111,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.06625258799171843,
|
| 237 |
+
"grad_norm": 2.980902395191929,
|
| 238 |
+
"learning_rate": 6.611570247933885e-06,
|
| 239 |
+
"loss": 0.819,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.06832298136645963,
|
| 244 |
+
"grad_norm": 3.0419508898389886,
|
| 245 |
+
"learning_rate": 6.818181818181818e-06,
|
| 246 |
+
"loss": 0.8177,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.07039337474120083,
|
| 251 |
+
"grad_norm": 3.1843886993600385,
|
| 252 |
+
"learning_rate": 7.0247933884297525e-06,
|
| 253 |
+
"loss": 0.8312,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.07246376811594203,
|
| 258 |
+
"grad_norm": 3.1854141596433867,
|
| 259 |
+
"learning_rate": 7.231404958677687e-06,
|
| 260 |
+
"loss": 0.813,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.07453416149068323,
|
| 265 |
+
"grad_norm": 3.2876165638717443,
|
| 266 |
+
"learning_rate": 7.43801652892562e-06,
|
| 267 |
+
"loss": 0.8384,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.07660455486542443,
|
| 272 |
+
"grad_norm": 2.9781450388628192,
|
| 273 |
+
"learning_rate": 7.644628099173555e-06,
|
| 274 |
+
"loss": 0.8094,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.07867494824016563,
|
| 279 |
+
"grad_norm": 3.348906096067214,
|
| 280 |
+
"learning_rate": 7.851239669421489e-06,
|
| 281 |
+
"loss": 0.8049,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.08074534161490683,
|
| 286 |
+
"grad_norm": 3.1092493369410783,
|
| 287 |
+
"learning_rate": 8.057851239669421e-06,
|
| 288 |
+
"loss": 0.826,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.08281573498964803,
|
| 293 |
+
"grad_norm": 3.056952614753785,
|
| 294 |
+
"learning_rate": 8.264462809917356e-06,
|
| 295 |
+
"loss": 0.8139,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.08488612836438923,
|
| 300 |
+
"grad_norm": 3.1015829113018234,
|
| 301 |
+
"learning_rate": 8.47107438016529e-06,
|
| 302 |
+
"loss": 0.8271,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.08695652173913043,
|
| 307 |
+
"grad_norm": 3.0740038363605926,
|
| 308 |
+
"learning_rate": 8.677685950413224e-06,
|
| 309 |
+
"loss": 0.801,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.08902691511387163,
|
| 314 |
+
"grad_norm": 2.9037810245033997,
|
| 315 |
+
"learning_rate": 8.884297520661158e-06,
|
| 316 |
+
"loss": 0.8095,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.09109730848861283,
|
| 321 |
+
"grad_norm": 3.027706124969457,
|
| 322 |
+
"learning_rate": 9.090909090909091e-06,
|
| 323 |
+
"loss": 0.8125,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.09316770186335403,
|
| 328 |
+
"grad_norm": 3.1151507400215745,
|
| 329 |
+
"learning_rate": 9.297520661157025e-06,
|
| 330 |
+
"loss": 0.797,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.09523809523809523,
|
| 335 |
+
"grad_norm": 3.053150958204543,
|
| 336 |
+
"learning_rate": 9.50413223140496e-06,
|
| 337 |
+
"loss": 0.7984,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.09730848861283643,
|
| 342 |
+
"grad_norm": 3.0112118668387047,
|
| 343 |
+
"learning_rate": 9.710743801652894e-06,
|
| 344 |
+
"loss": 0.8126,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.09937888198757763,
|
| 349 |
+
"grad_norm": 3.117727242870845,
|
| 350 |
+
"learning_rate": 9.917355371900828e-06,
|
| 351 |
+
"loss": 0.8048,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.10144927536231885,
|
| 356 |
+
"grad_norm": 2.8274397177377812,
|
| 357 |
+
"learning_rate": 9.999952971391373e-06,
|
| 358 |
+
"loss": 0.7992,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.10351966873706005,
|
| 363 |
+
"grad_norm": 3.006340482443734,
|
| 364 |
+
"learning_rate": 9.999665577542406e-06,
|
| 365 |
+
"loss": 0.7823,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.10559006211180125,
|
| 370 |
+
"grad_norm": 2.9753685361571915,
|
| 371 |
+
"learning_rate": 9.999116931848504e-06,
|
| 372 |
+
"loss": 0.804,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.10766045548654245,
|
| 377 |
+
"grad_norm": 2.8643733992658995,
|
| 378 |
+
"learning_rate": 9.99830706297852e-06,
|
| 379 |
+
"loss": 0.7747,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.10973084886128365,
|
| 384 |
+
"grad_norm": 2.8078822435477777,
|
| 385 |
+
"learning_rate": 9.997236013251234e-06,
|
| 386 |
+
"loss": 0.7961,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.11180124223602485,
|
| 391 |
+
"grad_norm": 2.912717313270744,
|
| 392 |
+
"learning_rate": 9.995903838633133e-06,
|
| 393 |
+
"loss": 0.783,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.11387163561076605,
|
| 398 |
+
"grad_norm": 2.8217001344431036,
|
| 399 |
+
"learning_rate": 9.994310608735492e-06,
|
| 400 |
+
"loss": 0.7999,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.11594202898550725,
|
| 405 |
+
"grad_norm": 2.958269779951464,
|
| 406 |
+
"learning_rate": 9.99245640681072e-06,
|
| 407 |
+
"loss": 0.7617,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.11801242236024845,
|
| 412 |
+
"grad_norm": 3.5988430245555416,
|
| 413 |
+
"learning_rate": 9.99034132974804e-06,
|
| 414 |
+
"loss": 0.79,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.12008281573498965,
|
| 419 |
+
"grad_norm": 2.7976040573963856,
|
| 420 |
+
"learning_rate": 9.987965488068398e-06,
|
| 421 |
+
"loss": 0.7506,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.12215320910973085,
|
| 426 |
+
"grad_norm": 2.8825061372075935,
|
| 427 |
+
"learning_rate": 9.985329005918702e-06,
|
| 428 |
+
"loss": 0.7841,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.12422360248447205,
|
| 433 |
+
"grad_norm": 3.05421834044982,
|
| 434 |
+
"learning_rate": 9.982432021065334e-06,
|
| 435 |
+
"loss": 0.7831,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.12629399585921325,
|
| 440 |
+
"grad_norm": 2.7477317097995537,
|
| 441 |
+
"learning_rate": 9.979274684886943e-06,
|
| 442 |
+
"loss": 0.7734,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.12836438923395446,
|
| 447 |
+
"grad_norm": 2.949511742636124,
|
| 448 |
+
"learning_rate": 9.975857162366547e-06,
|
| 449 |
+
"loss": 0.7577,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.13043478260869565,
|
| 454 |
+
"grad_norm": 3.0377639689629885,
|
| 455 |
+
"learning_rate": 9.972179632082899e-06,
|
| 456 |
+
"loss": 0.7725,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.13250517598343686,
|
| 461 |
+
"grad_norm": 3.1353289951358754,
|
| 462 |
+
"learning_rate": 9.968242286201171e-06,
|
| 463 |
+
"loss": 0.7619,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.13457556935817805,
|
| 468 |
+
"grad_norm": 3.1332684478011084,
|
| 469 |
+
"learning_rate": 9.964045330462896e-06,
|
| 470 |
+
"loss": 0.7611,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.13664596273291926,
|
| 475 |
+
"grad_norm": 3.156332097898071,
|
| 476 |
+
"learning_rate": 9.959588984175228e-06,
|
| 477 |
+
"loss": 0.7764,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.13871635610766045,
|
| 482 |
+
"grad_norm": 2.900772606537427,
|
| 483 |
+
"learning_rate": 9.954873480199483e-06,
|
| 484 |
+
"loss": 0.7571,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.14078674948240166,
|
| 489 |
+
"grad_norm": 2.975845497976506,
|
| 490 |
+
"learning_rate": 9.94989906493896e-06,
|
| 491 |
+
"loss": 0.7407,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.14285714285714285,
|
| 496 |
+
"grad_norm": 3.1365253794712364,
|
| 497 |
+
"learning_rate": 9.94466599832608e-06,
|
| 498 |
+
"loss": 0.776,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.14492753623188406,
|
| 503 |
+
"grad_norm": 2.7718028032418522,
|
| 504 |
+
"learning_rate": 9.939174553808793e-06,
|
| 505 |
+
"loss": 0.7441,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.14699792960662525,
|
| 510 |
+
"grad_norm": 2.8486682818864284,
|
| 511 |
+
"learning_rate": 9.933425018336292e-06,
|
| 512 |
+
"loss": 0.7476,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.14906832298136646,
|
| 517 |
+
"grad_norm": 2.9252854160968735,
|
| 518 |
+
"learning_rate": 9.927417692344025e-06,
|
| 519 |
+
"loss": 0.752,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.15113871635610765,
|
| 524 |
+
"grad_norm": 2.9773109330044902,
|
| 525 |
+
"learning_rate": 9.921152889737985e-06,
|
| 526 |
+
"loss": 0.7349,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.15320910973084886,
|
| 531 |
+
"grad_norm": 2.910786730363286,
|
| 532 |
+
"learning_rate": 9.914630937878315e-06,
|
| 533 |
+
"loss": 0.76,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.15527950310559005,
|
| 538 |
+
"grad_norm": 2.620946999109397,
|
| 539 |
+
"learning_rate": 9.907852177562201e-06,
|
| 540 |
+
"loss": 0.7246,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.15734989648033126,
|
| 545 |
+
"grad_norm": 2.7006556546764915,
|
| 546 |
+
"learning_rate": 9.900816963006063e-06,
|
| 547 |
+
"loss": 0.7548,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.15942028985507245,
|
| 552 |
+
"grad_norm": 2.8388206457825054,
|
| 553 |
+
"learning_rate": 9.893525661827043e-06,
|
| 554 |
+
"loss": 0.7334,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.16149068322981366,
|
| 559 |
+
"grad_norm": 2.7395620691398856,
|
| 560 |
+
"learning_rate": 9.885978655023805e-06,
|
| 561 |
+
"loss": 0.7487,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.16356107660455488,
|
| 566 |
+
"grad_norm": 2.7273846238672754,
|
| 567 |
+
"learning_rate": 9.87817633695661e-06,
|
| 568 |
+
"loss": 0.7304,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.16563146997929606,
|
| 573 |
+
"grad_norm": 2.8179008423909524,
|
| 574 |
+
"learning_rate": 9.87011911532673e-06,
|
| 575 |
+
"loss": 0.7419,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.16770186335403728,
|
| 580 |
+
"grad_norm": 3.149985443780922,
|
| 581 |
+
"learning_rate": 9.861807411155126e-06,
|
| 582 |
+
"loss": 0.7371,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.16977225672877846,
|
| 587 |
+
"grad_norm": 2.8768198170816026,
|
| 588 |
+
"learning_rate": 9.853241658760457e-06,
|
| 589 |
+
"loss": 0.7411,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.17184265010351968,
|
| 594 |
+
"grad_norm": 2.7229415130760146,
|
| 595 |
+
"learning_rate": 9.844422305736383e-06,
|
| 596 |
+
"loss": 0.7206,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.17391304347826086,
|
| 601 |
+
"grad_norm": 2.851284715864324,
|
| 602 |
+
"learning_rate": 9.835349812928178e-06,
|
| 603 |
+
"loss": 0.7168,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.17598343685300208,
|
| 608 |
+
"grad_norm": 2.757115924608802,
|
| 609 |
+
"learning_rate": 9.826024654408645e-06,
|
| 610 |
+
"loss": 0.7058,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.17805383022774326,
|
| 615 |
+
"grad_norm": 3.1745359708751812,
|
| 616 |
+
"learning_rate": 9.816447317453353e-06,
|
| 617 |
+
"loss": 0.718,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.18012422360248448,
|
| 622 |
+
"grad_norm": 2.911212797883499,
|
| 623 |
+
"learning_rate": 9.80661830251516e-06,
|
| 624 |
+
"loss": 0.7092,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.18219461697722567,
|
| 629 |
+
"grad_norm": 2.79176343971146,
|
| 630 |
+
"learning_rate": 9.796538123198077e-06,
|
| 631 |
+
"loss": 0.718,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.18426501035196688,
|
| 636 |
+
"grad_norm": 3.0726347632388755,
|
| 637 |
+
"learning_rate": 9.786207306230422e-06,
|
| 638 |
+
"loss": 0.7034,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.18633540372670807,
|
| 643 |
+
"grad_norm": 2.69142104953591,
|
| 644 |
+
"learning_rate": 9.775626391437303e-06,
|
| 645 |
+
"loss": 0.7029,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.18840579710144928,
|
| 650 |
+
"grad_norm": 2.8242682148315366,
|
| 651 |
+
"learning_rate": 9.764795931712396e-06,
|
| 652 |
+
"loss": 0.7148,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.19047619047619047,
|
| 657 |
+
"grad_norm": 3.107425827164937,
|
| 658 |
+
"learning_rate": 9.753716492989076e-06,
|
| 659 |
+
"loss": 0.7065,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.19254658385093168,
|
| 664 |
+
"grad_norm": 2.942574456055828,
|
| 665 |
+
"learning_rate": 9.742388654210822e-06,
|
| 666 |
+
"loss": 0.7061,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.19461697722567287,
|
| 671 |
+
"grad_norm": 2.7983523129620926,
|
| 672 |
+
"learning_rate": 9.730813007300984e-06,
|
| 673 |
+
"loss": 0.7009,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.19668737060041408,
|
| 678 |
+
"grad_norm": 2.748714187875416,
|
| 679 |
+
"learning_rate": 9.718990157131841e-06,
|
| 680 |
+
"loss": 0.6953,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.19875776397515527,
|
| 685 |
+
"grad_norm": 2.7787928902685732,
|
| 686 |
+
"learning_rate": 9.706920721492995e-06,
|
| 687 |
+
"loss": 0.6792,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.20082815734989648,
|
| 692 |
+
"grad_norm": 2.8759174624811843,
|
| 693 |
+
"learning_rate": 9.694605331059094e-06,
|
| 694 |
+
"loss": 0.688,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.2028985507246377,
|
| 699 |
+
"grad_norm": 2.839118884637647,
|
| 700 |
+
"learning_rate": 9.682044629356874e-06,
|
| 701 |
+
"loss": 0.688,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.20496894409937888,
|
| 706 |
+
"grad_norm": 2.7979296188735674,
|
| 707 |
+
"learning_rate": 9.669239272731538e-06,
|
| 708 |
+
"loss": 0.6781,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.2070393374741201,
|
| 713 |
+
"grad_norm": 2.9252774557866124,
|
| 714 |
+
"learning_rate": 9.656189930312443e-06,
|
| 715 |
+
"loss": 0.6664,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.20910973084886128,
|
| 720 |
+
"grad_norm": 2.809436333693126,
|
| 721 |
+
"learning_rate": 9.642897283978157e-06,
|
| 722 |
+
"loss": 0.6406,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.2111801242236025,
|
| 727 |
+
"grad_norm": 2.937022586016644,
|
| 728 |
+
"learning_rate": 9.629362028320808e-06,
|
| 729 |
+
"loss": 0.6732,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.21325051759834368,
|
| 734 |
+
"grad_norm": 2.9843908796613356,
|
| 735 |
+
"learning_rate": 9.615584870609809e-06,
|
| 736 |
+
"loss": 0.6745,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.2153209109730849,
|
| 741 |
+
"grad_norm": 3.0213679386695427,
|
| 742 |
+
"learning_rate": 9.601566530754882e-06,
|
| 743 |
+
"loss": 0.7024,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.21739130434782608,
|
| 748 |
+
"grad_norm": 2.7454641085709515,
|
| 749 |
+
"learning_rate": 9.587307741268452e-06,
|
| 750 |
+
"loss": 0.6654,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.2194616977225673,
|
| 755 |
+
"grad_norm": 2.699259252813139,
|
| 756 |
+
"learning_rate": 9.572809247227366e-06,
|
| 757 |
+
"loss": 0.6816,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.22153209109730848,
|
| 762 |
+
"grad_norm": 2.856456590780441,
|
| 763 |
+
"learning_rate": 9.558071806233955e-06,
|
| 764 |
+
"loss": 0.6807,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.2236024844720497,
|
| 769 |
+
"grad_norm": 2.8884926905246817,
|
| 770 |
+
"learning_rate": 9.54309618837646e-06,
|
| 771 |
+
"loss": 0.675,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.22567287784679088,
|
| 776 |
+
"grad_norm": 2.7159778958415655,
|
| 777 |
+
"learning_rate": 9.52788317618878e-06,
|
| 778 |
+
"loss": 0.6693,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.2277432712215321,
|
| 783 |
+
"grad_norm": 2.7094266982916024,
|
| 784 |
+
"learning_rate": 9.512433564609578e-06,
|
| 785 |
+
"loss": 0.6584,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.22981366459627328,
|
| 790 |
+
"grad_norm": 2.684727550045969,
|
| 791 |
+
"learning_rate": 9.496748160940762e-06,
|
| 792 |
+
"loss": 0.6521,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.2318840579710145,
|
| 797 |
+
"grad_norm": 2.8152646867143747,
|
| 798 |
+
"learning_rate": 9.480827784805278e-06,
|
| 799 |
+
"loss": 0.6631,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.23395445134575568,
|
| 804 |
+
"grad_norm": 2.746998219194102,
|
| 805 |
+
"learning_rate": 9.464673268104299e-06,
|
| 806 |
+
"loss": 0.6649,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.2360248447204969,
|
| 811 |
+
"grad_norm": 2.6681124619837417,
|
| 812 |
+
"learning_rate": 9.448285454973739e-06,
|
| 813 |
+
"loss": 0.6641,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.23809523809523808,
|
| 818 |
+
"grad_norm": 2.7076660662242773,
|
| 819 |
+
"learning_rate": 9.431665201740154e-06,
|
| 820 |
+
"loss": 0.661,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.2401656314699793,
|
| 825 |
+
"grad_norm": 2.79748527458811,
|
| 826 |
+
"learning_rate": 9.414813376876003e-06,
|
| 827 |
+
"loss": 0.6544,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.2422360248447205,
|
| 832 |
+
"grad_norm": 2.6352068976041223,
|
| 833 |
+
"learning_rate": 9.397730860954242e-06,
|
| 834 |
+
"loss": 0.6656,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.2443064182194617,
|
| 839 |
+
"grad_norm": 2.697353214501619,
|
| 840 |
+
"learning_rate": 9.38041854660234e-06,
|
| 841 |
+
"loss": 0.6455,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.2463768115942029,
|
| 846 |
+
"grad_norm": 2.7822792953780184,
|
| 847 |
+
"learning_rate": 9.362877338455611e-06,
|
| 848 |
+
"loss": 0.6572,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.2484472049689441,
|
| 853 |
+
"grad_norm": 2.76479005730922,
|
| 854 |
+
"learning_rate": 9.345108153109963e-06,
|
| 855 |
+
"loss": 0.657,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.2505175983436853,
|
| 860 |
+
"grad_norm": 2.641137724563326,
|
| 861 |
+
"learning_rate": 9.327111919073988e-06,
|
| 862 |
+
"loss": 0.6322,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.2525879917184265,
|
| 867 |
+
"grad_norm": 2.7457640915517962,
|
| 868 |
+
"learning_rate": 9.308889576720453e-06,
|
| 869 |
+
"loss": 0.6263,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.2546583850931677,
|
| 874 |
+
"grad_norm": 2.8809597496679715,
|
| 875 |
+
"learning_rate": 9.290442078237154e-06,
|
| 876 |
+
"loss": 0.6447,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.2567287784679089,
|
| 881 |
+
"grad_norm": 2.649088481356515,
|
| 882 |
+
"learning_rate": 9.271770387577168e-06,
|
| 883 |
+
"loss": 0.6513,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.2587991718426501,
|
| 888 |
+
"grad_norm": 2.723263824128368,
|
| 889 |
+
"learning_rate": 9.252875480408479e-06,
|
| 890 |
+
"loss": 0.6363,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.2608695652173913,
|
| 895 |
+
"grad_norm": 2.643341522308482,
|
| 896 |
+
"learning_rate": 9.233758344062996e-06,
|
| 897 |
+
"loss": 0.6636,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.2629399585921325,
|
| 902 |
+
"grad_norm": 2.760214611828711,
|
| 903 |
+
"learning_rate": 9.21441997748496e-06,
|
| 904 |
+
"loss": 0.6149,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.2650103519668737,
|
| 909 |
+
"grad_norm": 2.7138292470479155,
|
| 910 |
+
"learning_rate": 9.194861391178749e-06,
|
| 911 |
+
"loss": 0.6409,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.2670807453416149,
|
| 916 |
+
"grad_norm": 2.7960171070921853,
|
| 917 |
+
"learning_rate": 9.175083607156067e-06,
|
| 918 |
+
"loss": 0.6291,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.2691511387163561,
|
| 923 |
+
"grad_norm": 2.6781478740703535,
|
| 924 |
+
"learning_rate": 9.155087658882555e-06,
|
| 925 |
+
"loss": 0.6294,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.2712215320910973,
|
| 930 |
+
"grad_norm": 2.78377570516491,
|
| 931 |
+
"learning_rate": 9.134874591223773e-06,
|
| 932 |
+
"loss": 0.6424,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.2732919254658385,
|
| 937 |
+
"grad_norm": 2.6278416215911915,
|
| 938 |
+
"learning_rate": 9.114445460390605e-06,
|
| 939 |
+
"loss": 0.6256,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.2753623188405797,
|
| 944 |
+
"grad_norm": 2.9435581582969244,
|
| 945 |
+
"learning_rate": 9.093801333884076e-06,
|
| 946 |
+
"loss": 0.6396,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.2774327122153209,
|
| 951 |
+
"grad_norm": 2.717501103920811,
|
| 952 |
+
"learning_rate": 9.072943290439566e-06,
|
| 953 |
+
"loss": 0.6334,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.2795031055900621,
|
| 958 |
+
"grad_norm": 2.722616988464195,
|
| 959 |
+
"learning_rate": 9.051872419970439e-06,
|
| 960 |
+
"loss": 0.6278,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.2815734989648033,
|
| 965 |
+
"grad_norm": 2.718385633125168,
|
| 966 |
+
"learning_rate": 9.03058982351109e-06,
|
| 967 |
+
"loss": 0.604,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.2836438923395445,
|
| 972 |
+
"grad_norm": 2.6781823046558384,
|
| 973 |
+
"learning_rate": 9.009096613159426e-06,
|
| 974 |
+
"loss": 0.6092,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.2857142857142857,
|
| 979 |
+
"grad_norm": 2.7824837364041133,
|
| 980 |
+
"learning_rate": 8.987393912018732e-06,
|
| 981 |
+
"loss": 0.6145,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.28778467908902694,
|
| 986 |
+
"grad_norm": 2.7837043936680623,
|
| 987 |
+
"learning_rate": 8.965482854139003e-06,
|
| 988 |
+
"loss": 0.5975,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.2898550724637681,
|
| 993 |
+
"grad_norm": 2.716842889874823,
|
| 994 |
+
"learning_rate": 8.943364584457675e-06,
|
| 995 |
+
"loss": 0.619,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.2919254658385093,
|
| 1000 |
+
"grad_norm": 2.5861342696674106,
|
| 1001 |
+
"learning_rate": 8.921040258739804e-06,
|
| 1002 |
+
"loss": 0.6001,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.2939958592132505,
|
| 1007 |
+
"grad_norm": 2.810107936750979,
|
| 1008 |
+
"learning_rate": 8.898511043517668e-06,
|
| 1009 |
+
"loss": 0.6183,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.29606625258799174,
|
| 1014 |
+
"grad_norm": 2.7398243739315036,
|
| 1015 |
+
"learning_rate": 8.875778116029816e-06,
|
| 1016 |
+
"loss": 0.6021,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.2981366459627329,
|
| 1021 |
+
"grad_norm": 2.70859978292154,
|
| 1022 |
+
"learning_rate": 8.85284266415955e-06,
|
| 1023 |
+
"loss": 0.5953,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.3002070393374741,
|
| 1028 |
+
"grad_norm": 2.783835934538832,
|
| 1029 |
+
"learning_rate": 8.829705886372845e-06,
|
| 1030 |
+
"loss": 0.6013,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.3022774327122153,
|
| 1035 |
+
"grad_norm": 2.5763975267274994,
|
| 1036 |
+
"learning_rate": 8.806368991655747e-06,
|
| 1037 |
+
"loss": 0.5878,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.30434782608695654,
|
| 1042 |
+
"grad_norm": 2.697730711173743,
|
| 1043 |
+
"learning_rate": 8.782833199451177e-06,
|
| 1044 |
+
"loss": 0.6062,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.3064182194616977,
|
| 1049 |
+
"grad_norm": 2.9279171282668024,
|
| 1050 |
+
"learning_rate": 8.759099739595215e-06,
|
| 1051 |
+
"loss": 0.5935,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.3084886128364389,
|
| 1056 |
+
"grad_norm": 2.5900198753888373,
|
| 1057 |
+
"learning_rate": 8.735169852252848e-06,
|
| 1058 |
+
"loss": 0.5846,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.3105590062111801,
|
| 1063 |
+
"grad_norm": 2.6427817212845337,
|
| 1064 |
+
"learning_rate": 8.71104478785315e-06,
|
| 1065 |
+
"loss": 0.5852,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.31262939958592134,
|
| 1070 |
+
"grad_norm": 2.6213352415697724,
|
| 1071 |
+
"learning_rate": 8.686725807023955e-06,
|
| 1072 |
+
"loss": 0.5939,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.3146997929606625,
|
| 1077 |
+
"grad_norm": 2.695782898397081,
|
| 1078 |
+
"learning_rate": 8.662214180525982e-06,
|
| 1079 |
+
"loss": 0.5855,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.3167701863354037,
|
| 1084 |
+
"grad_norm": 2.6027809243052262,
|
| 1085 |
+
"learning_rate": 8.637511189186425e-06,
|
| 1086 |
+
"loss": 0.5807,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.3188405797101449,
|
| 1091 |
+
"grad_norm": 2.769447274560436,
|
| 1092 |
+
"learning_rate": 8.612618123832033e-06,
|
| 1093 |
+
"loss": 0.5815,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.32091097308488614,
|
| 1098 |
+
"grad_norm": 2.7990707978795246,
|
| 1099 |
+
"learning_rate": 8.587536285221656e-06,
|
| 1100 |
+
"loss": 0.5758,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.32298136645962733,
|
| 1105 |
+
"grad_norm": 2.584119759399477,
|
| 1106 |
+
"learning_rate": 8.562266983978278e-06,
|
| 1107 |
+
"loss": 0.5939,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.3250517598343685,
|
| 1112 |
+
"grad_norm": 2.8164096331343647,
|
| 1113 |
+
"learning_rate": 8.536811540520529e-06,
|
| 1114 |
+
"loss": 0.5866,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.32712215320910976,
|
| 1119 |
+
"grad_norm": 4.038986317699086,
|
| 1120 |
+
"learning_rate": 8.511171284993686e-06,
|
| 1121 |
+
"loss": 0.5538,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.32919254658385094,
|
| 1126 |
+
"grad_norm": 2.8362127741980334,
|
| 1127 |
+
"learning_rate": 8.485347557200177e-06,
|
| 1128 |
+
"loss": 0.5678,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.33126293995859213,
|
| 1133 |
+
"grad_norm": 2.772547092346996,
|
| 1134 |
+
"learning_rate": 8.459341706529557e-06,
|
| 1135 |
+
"loss": 0.5762,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.3333333333333333,
|
| 1140 |
+
"grad_norm": 2.523559738817573,
|
| 1141 |
+
"learning_rate": 8.43315509188801e-06,
|
| 1142 |
+
"loss": 0.5616,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.33540372670807456,
|
| 1147 |
+
"grad_norm": 2.734190622994657,
|
| 1148 |
+
"learning_rate": 8.406789081627337e-06,
|
| 1149 |
+
"loss": 0.5747,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.33747412008281574,
|
| 1154 |
+
"grad_norm": 2.5647167102786743,
|
| 1155 |
+
"learning_rate": 8.380245053473452e-06,
|
| 1156 |
+
"loss": 0.579,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.33954451345755693,
|
| 1161 |
+
"grad_norm": 2.7181506919040817,
|
| 1162 |
+
"learning_rate": 8.353524394454388e-06,
|
| 1163 |
+
"loss": 0.5645,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.3416149068322981,
|
| 1168 |
+
"grad_norm": 2.532384721512014,
|
| 1169 |
+
"learning_rate": 8.326628500827826e-06,
|
| 1170 |
+
"loss": 0.5816,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.34368530020703936,
|
| 1175 |
+
"grad_norm": 2.7095413968694593,
|
| 1176 |
+
"learning_rate": 8.299558778008137e-06,
|
| 1177 |
+
"loss": 0.5685,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.34575569358178054,
|
| 1182 |
+
"grad_norm": 2.6888726345421694,
|
| 1183 |
+
"learning_rate": 8.272316640492932e-06,
|
| 1184 |
+
"loss": 0.5608,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.34782608695652173,
|
| 1189 |
+
"grad_norm": 2.6165321145769216,
|
| 1190 |
+
"learning_rate": 8.244903511789158e-06,
|
| 1191 |
+
"loss": 0.5589,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.3498964803312629,
|
| 1196 |
+
"grad_norm": 2.688022401858765,
|
| 1197 |
+
"learning_rate": 8.217320824338713e-06,
|
| 1198 |
+
"loss": 0.567,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.35196687370600416,
|
| 1203 |
+
"grad_norm": 2.6698972691855003,
|
| 1204 |
+
"learning_rate": 8.189570019443597e-06,
|
| 1205 |
+
"loss": 0.5608,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.35403726708074534,
|
| 1210 |
+
"grad_norm": 2.5607516214652395,
|
| 1211 |
+
"learning_rate": 8.161652547190593e-06,
|
| 1212 |
+
"loss": 0.5544,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.35610766045548653,
|
| 1217 |
+
"grad_norm": 2.5093610647637807,
|
| 1218 |
+
"learning_rate": 8.133569866375497e-06,
|
| 1219 |
+
"loss": 0.556,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.3581780538302277,
|
| 1224 |
+
"grad_norm": 2.5888370930695754,
|
| 1225 |
+
"learning_rate": 8.105323444426891e-06,
|
| 1226 |
+
"loss": 0.5382,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.36024844720496896,
|
| 1231 |
+
"grad_norm": 2.502429725724875,
|
| 1232 |
+
"learning_rate": 8.076914757329467e-06,
|
| 1233 |
+
"loss": 0.5515,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.36231884057971014,
|
| 1238 |
+
"grad_norm": 2.513132604961173,
|
| 1239 |
+
"learning_rate": 8.048345289546895e-06,
|
| 1240 |
+
"loss": 0.5504,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.36438923395445133,
|
| 1245 |
+
"grad_norm": 2.613023773895591,
|
| 1246 |
+
"learning_rate": 8.01961653394426e-06,
|
| 1247 |
+
"loss": 0.5514,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.36645962732919257,
|
| 1252 |
+
"grad_norm": 2.8379598375055024,
|
| 1253 |
+
"learning_rate": 7.990729991710046e-06,
|
| 1254 |
+
"loss": 0.5359,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.36853002070393376,
|
| 1259 |
+
"grad_norm": 2.6353250882091075,
|
| 1260 |
+
"learning_rate": 7.9616871722777e-06,
|
| 1261 |
+
"loss": 0.5607,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.37060041407867494,
|
| 1266 |
+
"grad_norm": 2.695376888325521,
|
| 1267 |
+
"learning_rate": 7.932489593246764e-06,
|
| 1268 |
+
"loss": 0.5425,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.37267080745341613,
|
| 1273 |
+
"grad_norm": 2.59578776213556,
|
| 1274 |
+
"learning_rate": 7.903138780303556e-06,
|
| 1275 |
+
"loss": 0.5513,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.3747412008281574,
|
| 1280 |
+
"grad_norm": 2.6357999566585075,
|
| 1281 |
+
"learning_rate": 7.873636267141463e-06,
|
| 1282 |
+
"loss": 0.5189,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.37681159420289856,
|
| 1287 |
+
"grad_norm": 2.610338652616046,
|
| 1288 |
+
"learning_rate": 7.843983595380793e-06,
|
| 1289 |
+
"loss": 0.5564,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.37888198757763975,
|
| 1294 |
+
"grad_norm": 2.7200267334644566,
|
| 1295 |
+
"learning_rate": 7.814182314488225e-06,
|
| 1296 |
+
"loss": 0.5291,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.38095238095238093,
|
| 1301 |
+
"grad_norm": 2.55575507245088,
|
| 1302 |
+
"learning_rate": 7.784233981695835e-06,
|
| 1303 |
+
"loss": 0.5491,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.3830227743271222,
|
| 1308 |
+
"grad_norm": 2.741846594136525,
|
| 1309 |
+
"learning_rate": 7.754140161919732e-06,
|
| 1310 |
+
"loss": 0.5344,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.38509316770186336,
|
| 1315 |
+
"grad_norm": 2.6008608315377555,
|
| 1316 |
+
"learning_rate": 7.72390242767828e-06,
|
| 1317 |
+
"loss": 0.5124,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.38716356107660455,
|
| 1322 |
+
"grad_norm": 2.8320649567171197,
|
| 1323 |
+
"learning_rate": 7.693522359009931e-06,
|
| 1324 |
+
"loss": 0.5248,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.38923395445134573,
|
| 1329 |
+
"grad_norm": 2.603319566659037,
|
| 1330 |
+
"learning_rate": 7.663001543390657e-06,
|
| 1331 |
+
"loss": 0.525,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.391304347826087,
|
| 1336 |
+
"grad_norm": 2.560918488979843,
|
| 1337 |
+
"learning_rate": 7.63234157565101e-06,
|
| 1338 |
+
"loss": 0.5381,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.39337474120082816,
|
| 1343 |
+
"grad_norm": 2.708900648839639,
|
| 1344 |
+
"learning_rate": 7.601544057892769e-06,
|
| 1345 |
+
"loss": 0.5188,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.39544513457556935,
|
| 1350 |
+
"grad_norm": 2.557453498959152,
|
| 1351 |
+
"learning_rate": 7.570610599405242e-06,
|
| 1352 |
+
"loss": 0.5403,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.39751552795031053,
|
| 1357 |
+
"grad_norm": 2.568997745239728,
|
| 1358 |
+
"learning_rate": 7.539542816581157e-06,
|
| 1359 |
+
"loss": 0.5306,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.3995859213250518,
|
| 1364 |
+
"grad_norm": 2.849320989394085,
|
| 1365 |
+
"learning_rate": 7.508342332832213e-06,
|
| 1366 |
+
"loss": 0.5137,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.40165631469979296,
|
| 1371 |
+
"grad_norm": 2.4808674765715066,
|
| 1372 |
+
"learning_rate": 7.477010778504241e-06,
|
| 1373 |
+
"loss": 0.5207,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.40372670807453415,
|
| 1378 |
+
"grad_norm": 2.560284660092028,
|
| 1379 |
+
"learning_rate": 7.445549790792021e-06,
|
| 1380 |
+
"loss": 0.5192,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.4057971014492754,
|
| 1385 |
+
"grad_norm": 2.553120554037133,
|
| 1386 |
+
"learning_rate": 7.413961013653725e-06,
|
| 1387 |
+
"loss": 0.5114,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.4078674948240166,
|
| 1392 |
+
"grad_norm": 2.495494520049296,
|
| 1393 |
+
"learning_rate": 7.3822460977250145e-06,
|
| 1394 |
+
"loss": 0.5136,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.40993788819875776,
|
| 1399 |
+
"grad_norm": 2.6954833669770726,
|
| 1400 |
+
"learning_rate": 7.350406700232794e-06,
|
| 1401 |
+
"loss": 0.5264,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.41200828157349895,
|
| 1406 |
+
"grad_norm": 2.7381082998519983,
|
| 1407 |
+
"learning_rate": 7.318444484908606e-06,
|
| 1408 |
+
"loss": 0.5052,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.4140786749482402,
|
| 1413 |
+
"grad_norm": 2.550035768584111,
|
| 1414 |
+
"learning_rate": 7.286361121901706e-06,
|
| 1415 |
+
"loss": 0.5133,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.4161490683229814,
|
| 1420 |
+
"grad_norm": 3.142045365448196,
|
| 1421 |
+
"learning_rate": 7.254158287691775e-06,
|
| 1422 |
+
"loss": 0.51,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.41821946169772256,
|
| 1427 |
+
"grad_norm": 2.558804554803966,
|
| 1428 |
+
"learning_rate": 7.221837665001335e-06,
|
| 1429 |
+
"loss": 0.5168,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.42028985507246375,
|
| 1434 |
+
"grad_norm": 2.5530037450130343,
|
| 1435 |
+
"learning_rate": 7.189400942707804e-06,
|
| 1436 |
+
"loss": 0.506,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.422360248447205,
|
| 1441 |
+
"grad_norm": 2.5368973829874775,
|
| 1442 |
+
"learning_rate": 7.1568498157552576e-06,
|
| 1443 |
+
"loss": 0.5105,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.4244306418219462,
|
| 1448 |
+
"grad_norm": 2.4485501120829034,
|
| 1449 |
+
"learning_rate": 7.124185985065856e-06,
|
| 1450 |
+
"loss": 0.5219,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.42650103519668736,
|
| 1455 |
+
"grad_norm": 2.5431441032943836,
|
| 1456 |
+
"learning_rate": 7.091411157450965e-06,
|
| 1457 |
+
"loss": 0.5068,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.42857142857142855,
|
| 1462 |
+
"grad_norm": 2.463753095314535,
|
| 1463 |
+
"learning_rate": 7.0585270455219654e-06,
|
| 1464 |
+
"loss": 0.4885,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.4306418219461698,
|
| 1469 |
+
"grad_norm": 2.6318223236200873,
|
| 1470 |
+
"learning_rate": 7.025535367600771e-06,
|
| 1471 |
+
"loss": 0.5078,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.432712215320911,
|
| 1476 |
+
"grad_norm": 2.600283059347969,
|
| 1477 |
+
"learning_rate": 6.992437847630031e-06,
|
| 1478 |
+
"loss": 0.5008,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.43478260869565216,
|
| 1483 |
+
"grad_norm": 2.508163120217184,
|
| 1484 |
+
"learning_rate": 6.95923621508305e-06,
|
| 1485 |
+
"loss": 0.4939,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.43685300207039335,
|
| 1490 |
+
"grad_norm": 2.660009117404832,
|
| 1491 |
+
"learning_rate": 6.92593220487342e-06,
|
| 1492 |
+
"loss": 0.4967,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.4389233954451346,
|
| 1497 |
+
"grad_norm": 2.524467102115331,
|
| 1498 |
+
"learning_rate": 6.892527557264358e-06,
|
| 1499 |
+
"loss": 0.5016,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.4409937888198758,
|
| 1504 |
+
"grad_norm": 2.5265295780956722,
|
| 1505 |
+
"learning_rate": 6.859024017777779e-06,
|
| 1506 |
+
"loss": 0.5029,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.44306418219461696,
|
| 1511 |
+
"grad_norm": 2.765020322815882,
|
| 1512 |
+
"learning_rate": 6.825423337103074e-06,
|
| 1513 |
+
"loss": 0.4803,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.4451345755693582,
|
| 1518 |
+
"grad_norm": 2.704943538082429,
|
| 1519 |
+
"learning_rate": 6.791727271005642e-06,
|
| 1520 |
+
"loss": 0.4967,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.4472049689440994,
|
| 1525 |
+
"grad_norm": 2.609473963979335,
|
| 1526 |
+
"learning_rate": 6.757937580235138e-06,
|
| 1527 |
+
"loss": 0.4907,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.4492753623188406,
|
| 1532 |
+
"grad_norm": 2.551608044305943,
|
| 1533 |
+
"learning_rate": 6.724056030433464e-06,
|
| 1534 |
+
"loss": 0.4928,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.45134575569358176,
|
| 1539 |
+
"grad_norm": 2.5603760994887255,
|
| 1540 |
+
"learning_rate": 6.690084392042514e-06,
|
| 1541 |
+
"loss": 0.4904,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.453416149068323,
|
| 1546 |
+
"grad_norm": 2.489337193865874,
|
| 1547 |
+
"learning_rate": 6.656024440211662e-06,
|
| 1548 |
+
"loss": 0.4751,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.4554865424430642,
|
| 1553 |
+
"grad_norm": 2.476219428614687,
|
| 1554 |
+
"learning_rate": 6.621877954704996e-06,
|
| 1555 |
+
"loss": 0.4808,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.4575569358178054,
|
| 1560 |
+
"grad_norm": 2.51661088671244,
|
| 1561 |
+
"learning_rate": 6.5876467198083235e-06,
|
| 1562 |
+
"loss": 0.4838,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.45962732919254656,
|
| 1567 |
+
"grad_norm": 2.625540683850569,
|
| 1568 |
+
"learning_rate": 6.553332524235937e-06,
|
| 1569 |
+
"loss": 0.4857,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.4616977225672878,
|
| 1574 |
+
"grad_norm": 2.6692887392725533,
|
| 1575 |
+
"learning_rate": 6.518937161037144e-06,
|
| 1576 |
+
"loss": 0.4766,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.463768115942029,
|
| 1581 |
+
"grad_norm": 2.451852906338064,
|
| 1582 |
+
"learning_rate": 6.484462427502572e-06,
|
| 1583 |
+
"loss": 0.4689,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.4658385093167702,
|
| 1588 |
+
"grad_norm": 2.528928504026905,
|
| 1589 |
+
"learning_rate": 6.44991012507026e-06,
|
| 1590 |
+
"loss": 0.47,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.46790890269151136,
|
| 1595 |
+
"grad_norm": 2.6023195877779797,
|
| 1596 |
+
"learning_rate": 6.415282059231518e-06,
|
| 1597 |
+
"loss": 0.472,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.4699792960662526,
|
| 1602 |
+
"grad_norm": 2.515973456448443,
|
| 1603 |
+
"learning_rate": 6.380580039436586e-06,
|
| 1604 |
+
"loss": 0.4524,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.4720496894409938,
|
| 1609 |
+
"grad_norm": 2.4273434418887607,
|
| 1610 |
+
"learning_rate": 6.345805879000087e-06,
|
| 1611 |
+
"loss": 0.4662,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.474120082815735,
|
| 1616 |
+
"grad_norm": 2.5041098379950375,
|
| 1617 |
+
"learning_rate": 6.31096139500627e-06,
|
| 1618 |
+
"loss": 0.4684,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.47619047619047616,
|
| 1623 |
+
"grad_norm": 2.890849402884173,
|
| 1624 |
+
"learning_rate": 6.2760484082140604e-06,
|
| 1625 |
+
"loss": 0.4817,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.4782608695652174,
|
| 1630 |
+
"grad_norm": 2.4406551900826843,
|
| 1631 |
+
"learning_rate": 6.24106874296192e-06,
|
| 1632 |
+
"loss": 0.479,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.4803312629399586,
|
| 1637 |
+
"grad_norm": 2.581296500428043,
|
| 1638 |
+
"learning_rate": 6.20602422707252e-06,
|
| 1639 |
+
"loss": 0.4626,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.4824016563146998,
|
| 1644 |
+
"grad_norm": 2.462136348090071,
|
| 1645 |
+
"learning_rate": 6.1709166917572264e-06,
|
| 1646 |
+
"loss": 0.4796,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.484472049689441,
|
| 1651 |
+
"grad_norm": 2.6803789589303646,
|
| 1652 |
+
"learning_rate": 6.135747971520412e-06,
|
| 1653 |
+
"loss": 0.4728,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.4865424430641822,
|
| 1658 |
+
"grad_norm": 2.4077364719321985,
|
| 1659 |
+
"learning_rate": 6.100519904063597e-06,
|
| 1660 |
+
"loss": 0.4623,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.4886128364389234,
|
| 1665 |
+
"grad_norm": 2.4101411199940315,
|
| 1666 |
+
"learning_rate": 6.0652343301894345e-06,
|
| 1667 |
+
"loss": 0.4657,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.4906832298136646,
|
| 1672 |
+
"grad_norm": 2.5188929556410415,
|
| 1673 |
+
"learning_rate": 6.029893093705492e-06,
|
| 1674 |
+
"loss": 0.46,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.4927536231884058,
|
| 1679 |
+
"grad_norm": 2.3677717501653563,
|
| 1680 |
+
"learning_rate": 5.99449804132794e-06,
|
| 1681 |
+
"loss": 0.4757,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.494824016563147,
|
| 1686 |
+
"grad_norm": 2.5710734020272157,
|
| 1687 |
+
"learning_rate": 5.959051022585025e-06,
|
| 1688 |
+
"loss": 0.473,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.4968944099378882,
|
| 1693 |
+
"grad_norm": 2.836194471253979,
|
| 1694 |
+
"learning_rate": 5.923553889720447e-06,
|
| 1695 |
+
"loss": 0.4504,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.4989648033126294,
|
| 1700 |
+
"grad_norm": 2.5393687325157073,
|
| 1701 |
+
"learning_rate": 5.888008497596553e-06,
|
| 1702 |
+
"loss": 0.4609,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.5010351966873706,
|
| 1707 |
+
"grad_norm": 2.4510897506943845,
|
| 1708 |
+
"learning_rate": 5.852416703597431e-06,
|
| 1709 |
+
"loss": 0.4696,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5031055900621118,
|
| 1714 |
+
"grad_norm": 2.4297048949306435,
|
| 1715 |
+
"learning_rate": 5.816780367531841e-06,
|
| 1716 |
+
"loss": 0.4469,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.505175983436853,
|
| 1721 |
+
"grad_norm": 2.456463228901228,
|
| 1722 |
+
"learning_rate": 5.781101351536041e-06,
|
| 1723 |
+
"loss": 0.4541,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.5072463768115942,
|
| 1728 |
+
"grad_norm": 2.44277926327073,
|
| 1729 |
+
"learning_rate": 5.745381519976477e-06,
|
| 1730 |
+
"loss": 0.4504,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.5093167701863354,
|
| 1735 |
+
"grad_norm": 2.5010485725693328,
|
| 1736 |
+
"learning_rate": 5.7096227393523716e-06,
|
| 1737 |
+
"loss": 0.4613,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.5113871635610766,
|
| 1742 |
+
"grad_norm": 2.6469453128675613,
|
| 1743 |
+
"learning_rate": 5.673826878198181e-06,
|
| 1744 |
+
"loss": 0.4493,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.5134575569358178,
|
| 1749 |
+
"grad_norm": 2.4138425627962636,
|
| 1750 |
+
"learning_rate": 5.637995806985961e-06,
|
| 1751 |
+
"loss": 0.4546,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.515527950310559,
|
| 1756 |
+
"grad_norm": 2.7188944082485924,
|
| 1757 |
+
"learning_rate": 5.602131398027637e-06,
|
| 1758 |
+
"loss": 0.4527,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.5175983436853002,
|
| 1763 |
+
"grad_norm": 2.3956639841305893,
|
| 1764 |
+
"learning_rate": 5.566235525377155e-06,
|
| 1765 |
+
"loss": 0.4456,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.5196687370600414,
|
| 1770 |
+
"grad_norm": 2.667940884594386,
|
| 1771 |
+
"learning_rate": 5.530310064732559e-06,
|
| 1772 |
+
"loss": 0.4475,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.5217391304347826,
|
| 1777 |
+
"grad_norm": 2.568790625515596,
|
| 1778 |
+
"learning_rate": 5.494356893337985e-06,
|
| 1779 |
+
"loss": 0.4603,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.5238095238095238,
|
| 1784 |
+
"grad_norm": 2.537093225337417,
|
| 1785 |
+
"learning_rate": 5.4583778898855576e-06,
|
| 1786 |
+
"loss": 0.4517,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.525879917184265,
|
| 1791 |
+
"grad_norm": 2.377214097836386,
|
| 1792 |
+
"learning_rate": 5.422374934417228e-06,
|
| 1793 |
+
"loss": 0.4334,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.5279503105590062,
|
| 1798 |
+
"grad_norm": 2.5322732831198307,
|
| 1799 |
+
"learning_rate": 5.386349908226538e-06,
|
| 1800 |
+
"loss": 0.443,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.5300207039337475,
|
| 1805 |
+
"grad_norm": 2.275812013448177,
|
| 1806 |
+
"learning_rate": 5.350304693760301e-06,
|
| 1807 |
+
"loss": 0.452,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.5320910973084886,
|
| 1812 |
+
"grad_norm": 2.490540393564678,
|
| 1813 |
+
"learning_rate": 5.314241174520251e-06,
|
| 1814 |
+
"loss": 0.4571,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.5341614906832298,
|
| 1819 |
+
"grad_norm": 2.303916018471547,
|
| 1820 |
+
"learning_rate": 5.2781612349646175e-06,
|
| 1821 |
+
"loss": 0.4437,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.5362318840579711,
|
| 1826 |
+
"grad_norm": 2.4742224611265367,
|
| 1827 |
+
"learning_rate": 5.242066760409653e-06,
|
| 1828 |
+
"loss": 0.4419,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.5383022774327122,
|
| 1833 |
+
"grad_norm": 2.6048926561003323,
|
| 1834 |
+
"learning_rate": 5.205959636931121e-06,
|
| 1835 |
+
"loss": 0.4334,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.5403726708074534,
|
| 1840 |
+
"grad_norm": 2.4996195423198975,
|
| 1841 |
+
"learning_rate": 5.169841751265738e-06,
|
| 1842 |
+
"loss": 0.4477,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.5424430641821946,
|
| 1847 |
+
"grad_norm": 2.436615222988275,
|
| 1848 |
+
"learning_rate": 5.13371499071259e-06,
|
| 1849 |
+
"loss": 0.4227,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.5445134575569358,
|
| 1854 |
+
"grad_norm": 2.4895563915580468,
|
| 1855 |
+
"learning_rate": 5.09758124303451e-06,
|
| 1856 |
+
"loss": 0.4382,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.546583850931677,
|
| 1861 |
+
"grad_norm": 2.457090055709984,
|
| 1862 |
+
"learning_rate": 5.0614423963594295e-06,
|
| 1863 |
+
"loss": 0.4383,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.5486542443064182,
|
| 1868 |
+
"grad_norm": 2.46529676282277,
|
| 1869 |
+
"learning_rate": 5.0253003390817264e-06,
|
| 1870 |
+
"loss": 0.4404,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.5507246376811594,
|
| 1875 |
+
"grad_norm": 2.509514530174065,
|
| 1876 |
+
"learning_rate": 4.989156959763551e-06,
|
| 1877 |
+
"loss": 0.432,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.5527950310559007,
|
| 1882 |
+
"grad_norm": 2.3489886489521568,
|
| 1883 |
+
"learning_rate": 4.953014147036121e-06,
|
| 1884 |
+
"loss": 0.4238,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.5548654244306418,
|
| 1889 |
+
"grad_norm": 2.548572605165313,
|
| 1890 |
+
"learning_rate": 4.9168737895010615e-06,
|
| 1891 |
+
"loss": 0.4385,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.556935817805383,
|
| 1896 |
+
"grad_norm": 2.5754573965566188,
|
| 1897 |
+
"learning_rate": 4.880737775631698e-06,
|
| 1898 |
+
"loss": 0.4188,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.5590062111801242,
|
| 1903 |
+
"grad_norm": 2.4663169631010633,
|
| 1904 |
+
"learning_rate": 4.844607993674382e-06,
|
| 1905 |
+
"loss": 0.4277,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.5610766045548654,
|
| 1910 |
+
"grad_norm": 2.5893970305781124,
|
| 1911 |
+
"learning_rate": 4.808486331549824e-06,
|
| 1912 |
+
"loss": 0.4246,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.5631469979296067,
|
| 1917 |
+
"grad_norm": 2.5461127745316987,
|
| 1918 |
+
"learning_rate": 4.772374676754444e-06,
|
| 1919 |
+
"loss": 0.4269,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.5652173913043478,
|
| 1924 |
+
"grad_norm": 2.4510878027098784,
|
| 1925 |
+
"learning_rate": 4.736274916261741e-06,
|
| 1926 |
+
"loss": 0.4227,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.567287784679089,
|
| 1931 |
+
"grad_norm": 2.4337437356788882,
|
| 1932 |
+
"learning_rate": 4.700188936423683e-06,
|
| 1933 |
+
"loss": 0.4301,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.5693581780538303,
|
| 1938 |
+
"grad_norm": 2.616623513135904,
|
| 1939 |
+
"learning_rate": 4.664118622872157e-06,
|
| 1940 |
+
"loss": 0.4234,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.5714285714285714,
|
| 1945 |
+
"grad_norm": 2.3671956167402874,
|
| 1946 |
+
"learning_rate": 4.628065860420417e-06,
|
| 1947 |
+
"loss": 0.4269,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.5734989648033126,
|
| 1952 |
+
"grad_norm": 2.484279489532884,
|
| 1953 |
+
"learning_rate": 4.592032532964611e-06,
|
| 1954 |
+
"loss": 0.4354,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.5755693581780539,
|
| 1959 |
+
"grad_norm": 2.4601006150341416,
|
| 1960 |
+
"learning_rate": 4.556020523385326e-06,
|
| 1961 |
+
"loss": 0.4245,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.577639751552795,
|
| 1966 |
+
"grad_norm": 2.6097660668896454,
|
| 1967 |
+
"learning_rate": 4.520031713449215e-06,
|
| 1968 |
+
"loss": 0.4295,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.5797101449275363,
|
| 1973 |
+
"grad_norm": 2.4072399455791658,
|
| 1974 |
+
"learning_rate": 4.484067983710653e-06,
|
| 1975 |
+
"loss": 0.4234,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.5817805383022774,
|
| 1980 |
+
"grad_norm": 2.448489377931204,
|
| 1981 |
+
"learning_rate": 4.448131213413485e-06,
|
| 1982 |
+
"loss": 0.4203,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.5838509316770186,
|
| 1987 |
+
"grad_norm": 2.443669450647813,
|
| 1988 |
+
"learning_rate": 4.4122232803928145e-06,
|
| 1989 |
+
"loss": 0.4197,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.5859213250517599,
|
| 1994 |
+
"grad_norm": 2.4564289974274485,
|
| 1995 |
+
"learning_rate": 4.376346060976888e-06,
|
| 1996 |
+
"loss": 0.4136,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.587991718426501,
|
| 2001 |
+
"grad_norm": 2.5200444579378933,
|
| 2002 |
+
"learning_rate": 4.340501429889053e-06,
|
| 2003 |
+
"loss": 0.4151,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.5900621118012422,
|
| 2008 |
+
"grad_norm": 2.48165117649947,
|
| 2009 |
+
"learning_rate": 4.30469126014978e-06,
|
| 2010 |
+
"loss": 0.402,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.5921325051759835,
|
| 2015 |
+
"grad_norm": 2.4558613956242845,
|
| 2016 |
+
"learning_rate": 4.268917422978811e-06,
|
| 2017 |
+
"loss": 0.4059,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.5942028985507246,
|
| 2022 |
+
"grad_norm": 2.478521562259034,
|
| 2023 |
+
"learning_rate": 4.233181787697364e-06,
|
| 2024 |
+
"loss": 0.4234,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.5962732919254659,
|
| 2029 |
+
"grad_norm": 2.3579770809913656,
|
| 2030 |
+
"learning_rate": 4.197486221630467e-06,
|
| 2031 |
+
"loss": 0.4177,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.598343685300207,
|
| 2036 |
+
"grad_norm": 2.440943795986475,
|
| 2037 |
+
"learning_rate": 4.161832590009371e-06,
|
| 2038 |
+
"loss": 0.4136,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.6004140786749482,
|
| 2043 |
+
"grad_norm": 2.3924731794257434,
|
| 2044 |
+
"learning_rate": 4.1262227558740915e-06,
|
| 2045 |
+
"loss": 0.4097,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.6024844720496895,
|
| 2050 |
+
"grad_norm": 2.3853646352098914,
|
| 2051 |
+
"learning_rate": 4.090658579976058e-06,
|
| 2052 |
+
"loss": 0.4125,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.6045548654244306,
|
| 2057 |
+
"grad_norm": 2.476730861427627,
|
| 2058 |
+
"learning_rate": 4.055141920680883e-06,
|
| 2059 |
+
"loss": 0.411,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.6066252587991718,
|
| 2064 |
+
"grad_norm": 2.338110722995676,
|
| 2065 |
+
"learning_rate": 4.019674633871246e-06,
|
| 2066 |
+
"loss": 0.4053,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.6086956521739131,
|
| 2071 |
+
"grad_norm": 2.3987001096487726,
|
| 2072 |
+
"learning_rate": 3.984258572849926e-06,
|
| 2073 |
+
"loss": 0.4136,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.6107660455486542,
|
| 2078 |
+
"grad_norm": 2.3639908244617422,
|
| 2079 |
+
"learning_rate": 3.9488955882429605e-06,
|
| 2080 |
+
"loss": 0.4093,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.6128364389233955,
|
| 2085 |
+
"grad_norm": 2.422313766220588,
|
| 2086 |
+
"learning_rate": 3.9135875279029366e-06,
|
| 2087 |
+
"loss": 0.4104,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.6149068322981367,
|
| 2092 |
+
"grad_norm": 2.4728098446592854,
|
| 2093 |
+
"learning_rate": 3.878336236812432e-06,
|
| 2094 |
+
"loss": 0.3814,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.6169772256728778,
|
| 2099 |
+
"grad_norm": 2.3708630478506176,
|
| 2100 |
+
"learning_rate": 3.84314355698762e-06,
|
| 2101 |
+
"loss": 0.4105,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.6190476190476191,
|
| 2106 |
+
"grad_norm": 2.4275761087290304,
|
| 2107 |
+
"learning_rate": 3.8080113273820024e-06,
|
| 2108 |
+
"loss": 0.4058,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.6211180124223602,
|
| 2113 |
+
"grad_norm": 2.4421724912048077,
|
| 2114 |
+
"learning_rate": 3.7729413837903285e-06,
|
| 2115 |
+
"loss": 0.3913,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.6231884057971014,
|
| 2120 |
+
"grad_norm": 2.4517011088536,
|
| 2121 |
+
"learning_rate": 3.7379355587526543e-06,
|
| 2122 |
+
"loss": 0.3962,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.6252587991718427,
|
| 2127 |
+
"grad_norm": 2.4069249372010946,
|
| 2128 |
+
"learning_rate": 3.702995681458605e-06,
|
| 2129 |
+
"loss": 0.3941,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.6273291925465838,
|
| 2134 |
+
"grad_norm": 2.323725479679109,
|
| 2135 |
+
"learning_rate": 3.668123577651773e-06,
|
| 2136 |
+
"loss": 0.3984,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.629399585921325,
|
| 2141 |
+
"grad_norm": 2.3190346458499884,
|
| 2142 |
+
"learning_rate": 3.633321069534325e-06,
|
| 2143 |
+
"loss": 0.3977,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.6314699792960663,
|
| 2148 |
+
"grad_norm": 2.335078659342428,
|
| 2149 |
+
"learning_rate": 3.5985899756717872e-06,
|
| 2150 |
+
"loss": 0.4047,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.6335403726708074,
|
| 2155 |
+
"grad_norm": 2.455638307213381,
|
| 2156 |
+
"learning_rate": 3.563932110898015e-06,
|
| 2157 |
+
"loss": 0.3912,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.6356107660455487,
|
| 2162 |
+
"grad_norm": 2.3702634454791993,
|
| 2163 |
+
"learning_rate": 3.5293492862203595e-06,
|
| 2164 |
+
"loss": 0.3847,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.6376811594202898,
|
| 2169 |
+
"grad_norm": 2.3580779842802753,
|
| 2170 |
+
"learning_rate": 3.494843308725032e-06,
|
| 2171 |
+
"loss": 0.3955,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.639751552795031,
|
| 2176 |
+
"grad_norm": 2.328047105272821,
|
| 2177 |
+
"learning_rate": 3.460415981482691e-06,
|
| 2178 |
+
"loss": 0.3915,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.6418219461697723,
|
| 2183 |
+
"grad_norm": 2.4764402089768476,
|
| 2184 |
+
"learning_rate": 3.426069103454209e-06,
|
| 2185 |
+
"loss": 0.3886,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.6438923395445134,
|
| 2190 |
+
"grad_norm": 2.454108156460191,
|
| 2191 |
+
"learning_rate": 3.3918044693966802e-06,
|
| 2192 |
+
"loss": 0.3842,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.6459627329192547,
|
| 2197 |
+
"grad_norm": 2.441202869593845,
|
| 2198 |
+
"learning_rate": 3.357623869769628e-06,
|
| 2199 |
+
"loss": 0.3724,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.6480331262939959,
|
| 2204 |
+
"grad_norm": 2.4755133325255256,
|
| 2205 |
+
"learning_rate": 3.3235290906414575e-06,
|
| 2206 |
+
"loss": 0.3822,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.650103519668737,
|
| 2211 |
+
"grad_norm": 2.4311073492384336,
|
| 2212 |
+
"learning_rate": 3.289521913596117e-06,
|
| 2213 |
+
"loss": 0.3908,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.6521739130434783,
|
| 2218 |
+
"grad_norm": 2.381193945822083,
|
| 2219 |
+
"learning_rate": 3.255604115640014e-06,
|
| 2220 |
+
"loss": 0.3825,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.6542443064182195,
|
| 2225 |
+
"grad_norm": 2.350792909332454,
|
| 2226 |
+
"learning_rate": 3.2217774691091437e-06,
|
| 2227 |
+
"loss": 0.3946,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.6563146997929606,
|
| 2232 |
+
"grad_norm": 2.390554035306188,
|
| 2233 |
+
"learning_rate": 3.1880437415764936e-06,
|
| 2234 |
+
"loss": 0.3993,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.6583850931677019,
|
| 2239 |
+
"grad_norm": 2.329131133970712,
|
| 2240 |
+
"learning_rate": 3.1544046957596746e-06,
|
| 2241 |
+
"loss": 0.3884,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.660455486542443,
|
| 2246 |
+
"grad_norm": 2.5542279992917902,
|
| 2247 |
+
"learning_rate": 3.1208620894288105e-06,
|
| 2248 |
+
"loss": 0.3908,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.6625258799171843,
|
| 2253 |
+
"grad_norm": 2.3701687553879243,
|
| 2254 |
+
"learning_rate": 3.087417675314687e-06,
|
| 2255 |
+
"loss": 0.3823,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.6645962732919255,
|
| 2260 |
+
"grad_norm": 2.225525354523903,
|
| 2261 |
+
"learning_rate": 3.0540732010171704e-06,
|
| 2262 |
+
"loss": 0.3733,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.6666666666666666,
|
| 2267 |
+
"grad_norm": 2.366938192597044,
|
| 2268 |
+
"learning_rate": 3.0208304089138873e-06,
|
| 2269 |
+
"loss": 0.383,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.6687370600414079,
|
| 2274 |
+
"grad_norm": 2.4030474398123953,
|
| 2275 |
+
"learning_rate": 2.9876910360691707e-06,
|
| 2276 |
+
"loss": 0.3919,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.6708074534161491,
|
| 2281 |
+
"grad_norm": 2.4354010216355992,
|
| 2282 |
+
"learning_rate": 2.9546568141433007e-06,
|
| 2283 |
+
"loss": 0.3868,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.6728778467908902,
|
| 2288 |
+
"grad_norm": 2.356177911915648,
|
| 2289 |
+
"learning_rate": 2.921729469302018e-06,
|
| 2290 |
+
"loss": 0.3803,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.6749482401656315,
|
| 2295 |
+
"grad_norm": 2.455281456899517,
|
| 2296 |
+
"learning_rate": 2.8889107221263203e-06,
|
| 2297 |
+
"loss": 0.3713,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.6770186335403726,
|
| 2302 |
+
"grad_norm": 2.221742782748604,
|
| 2303 |
+
"learning_rate": 2.856202287522556e-06,
|
| 2304 |
+
"loss": 0.3766,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.6790890269151139,
|
| 2309 |
+
"grad_norm": 2.6070082952771787,
|
| 2310 |
+
"learning_rate": 2.8236058746328203e-06,
|
| 2311 |
+
"loss": 0.3743,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.6811594202898551,
|
| 2316 |
+
"grad_norm": 2.3423694128380292,
|
| 2317 |
+
"learning_rate": 2.7911231867456345e-06,
|
| 2318 |
+
"loss": 0.3684,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.6832298136645962,
|
| 2323 |
+
"grad_norm": 2.242939946393628,
|
| 2324 |
+
"learning_rate": 2.7587559212069554e-06,
|
| 2325 |
+
"loss": 0.4181,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.6853002070393375,
|
| 2330 |
+
"grad_norm": 2.4538446272076775,
|
| 2331 |
+
"learning_rate": 2.72650576933147e-06,
|
| 2332 |
+
"loss": 0.3695,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.6873706004140787,
|
| 2337 |
+
"grad_norm": 2.248569538212587,
|
| 2338 |
+
"learning_rate": 2.6943744163142292e-06,
|
| 2339 |
+
"loss": 0.3634,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.6894409937888198,
|
| 2344 |
+
"grad_norm": 2.4019909678857494,
|
| 2345 |
+
"learning_rate": 2.662363541142585e-06,
|
| 2346 |
+
"loss": 0.3766,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.6915113871635611,
|
| 2351 |
+
"grad_norm": 2.571206708691807,
|
| 2352 |
+
"learning_rate": 2.6304748165084503e-06,
|
| 2353 |
+
"loss": 0.375,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.6935817805383023,
|
| 2358 |
+
"grad_norm": 2.434809314229936,
|
| 2359 |
+
"learning_rate": 2.5987099087209035e-06,
|
| 2360 |
+
"loss": 0.3693,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.6956521739130435,
|
| 2365 |
+
"grad_norm": 2.339716705035725,
|
| 2366 |
+
"learning_rate": 2.5670704776191158e-06,
|
| 2367 |
+
"loss": 0.3673,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.6977225672877847,
|
| 2372 |
+
"grad_norm": 2.442360671231122,
|
| 2373 |
+
"learning_rate": 2.5355581764856186e-06,
|
| 2374 |
+
"loss": 0.37,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.6997929606625258,
|
| 2379 |
+
"grad_norm": 2.356305225228509,
|
| 2380 |
+
"learning_rate": 2.5041746519599036e-06,
|
| 2381 |
+
"loss": 0.3722,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.7018633540372671,
|
| 2386 |
+
"grad_norm": 2.400871546850089,
|
| 2387 |
+
"learning_rate": 2.4729215439523945e-06,
|
| 2388 |
+
"loss": 0.378,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.7039337474120083,
|
| 2393 |
+
"grad_norm": 2.485733684370957,
|
| 2394 |
+
"learning_rate": 2.441800485558748e-06,
|
| 2395 |
+
"loss": 0.368,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.7060041407867494,
|
| 2400 |
+
"grad_norm": 2.4153165474956295,
|
| 2401 |
+
"learning_rate": 2.410813102974512e-06,
|
| 2402 |
+
"loss": 0.374,
|
| 2403 |
+
"step": 1705
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.7080745341614907,
|
| 2407 |
+
"grad_norm": 2.417914099527471,
|
| 2408 |
+
"learning_rate": 2.3799610154101603e-06,
|
| 2409 |
+
"loss": 0.3649,
|
| 2410 |
+
"step": 1710
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.7101449275362319,
|
| 2414 |
+
"grad_norm": 2.421566323056135,
|
| 2415 |
+
"learning_rate": 2.3492458350064805e-06,
|
| 2416 |
+
"loss": 0.3722,
|
| 2417 |
+
"step": 1715
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.7122153209109731,
|
| 2421 |
+
"grad_norm": 110.14431209902904,
|
| 2422 |
+
"learning_rate": 2.3186691667503315e-06,
|
| 2423 |
+
"loss": 0.376,
|
| 2424 |
+
"step": 1720
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.7142857142857143,
|
| 2428 |
+
"grad_norm": 2.498987084667395,
|
| 2429 |
+
"learning_rate": 2.288232608390778e-06,
|
| 2430 |
+
"loss": 0.367,
|
| 2431 |
+
"step": 1725
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.7163561076604554,
|
| 2435 |
+
"grad_norm": 2.377271205692582,
|
| 2436 |
+
"learning_rate": 2.2579377503555967e-06,
|
| 2437 |
+
"loss": 0.364,
|
| 2438 |
+
"step": 1730
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.7184265010351967,
|
| 2442 |
+
"grad_norm": 2.40592016919288,
|
| 2443 |
+
"learning_rate": 2.227786175668181e-06,
|
| 2444 |
+
"loss": 0.3661,
|
| 2445 |
+
"step": 1735
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.7204968944099379,
|
| 2449 |
+
"grad_norm": 2.4853659060167748,
|
| 2450 |
+
"learning_rate": 2.1977794598648106e-06,
|
| 2451 |
+
"loss": 0.3566,
|
| 2452 |
+
"step": 1740
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.722567287784679,
|
| 2456 |
+
"grad_norm": 2.5196941482373556,
|
| 2457 |
+
"learning_rate": 2.167919170912333e-06,
|
| 2458 |
+
"loss": 0.3679,
|
| 2459 |
+
"step": 1745
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.7246376811594203,
|
| 2463 |
+
"grad_norm": 2.3428574169962872,
|
| 2464 |
+
"learning_rate": 2.138206869126222e-06,
|
| 2465 |
+
"loss": 0.3534,
|
| 2466 |
+
"step": 1750
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.7267080745341615,
|
| 2470 |
+
"grad_norm": 2.3854557500475675,
|
| 2471 |
+
"learning_rate": 2.108644107089055e-06,
|
| 2472 |
+
"loss": 0.3584,
|
| 2473 |
+
"step": 1755
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.7287784679089027,
|
| 2477 |
+
"grad_norm": 2.5623905943162635,
|
| 2478 |
+
"learning_rate": 2.0792324295693782e-06,
|
| 2479 |
+
"loss": 0.3663,
|
| 2480 |
+
"step": 1760
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.7308488612836439,
|
| 2484 |
+
"grad_norm": 2.1463966743523235,
|
| 2485 |
+
"learning_rate": 2.0499733734409884e-06,
|
| 2486 |
+
"loss": 0.3689,
|
| 2487 |
+
"step": 1765
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.7329192546583851,
|
| 2491 |
+
"grad_norm": 2.3434379762931234,
|
| 2492 |
+
"learning_rate": 2.020868467602623e-06,
|
| 2493 |
+
"loss": 0.3501,
|
| 2494 |
+
"step": 1770
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.7349896480331263,
|
| 2498 |
+
"grad_norm": 2.466102573223308,
|
| 2499 |
+
"learning_rate": 1.9919192328980703e-06,
|
| 2500 |
+
"loss": 0.3406,
|
| 2501 |
+
"step": 1775
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.7370600414078675,
|
| 2505 |
+
"grad_norm": 2.441564402693419,
|
| 2506 |
+
"learning_rate": 1.9631271820367058e-06,
|
| 2507 |
+
"loss": 0.3555,
|
| 2508 |
+
"step": 1780
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.7391304347826086,
|
| 2512 |
+
"grad_norm": 2.474880211520262,
|
| 2513 |
+
"learning_rate": 1.9344938195144353e-06,
|
| 2514 |
+
"loss": 0.351,
|
| 2515 |
+
"step": 1785
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.7412008281573499,
|
| 2519 |
+
"grad_norm": 2.482728988847221,
|
| 2520 |
+
"learning_rate": 1.9060206415350896e-06,
|
| 2521 |
+
"loss": 0.3638,
|
| 2522 |
+
"step": 1790
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.7432712215320911,
|
| 2526 |
+
"grad_norm": 2.3721058207362,
|
| 2527 |
+
"learning_rate": 1.8777091359322386e-06,
|
| 2528 |
+
"loss": 0.3553,
|
| 2529 |
+
"step": 1795
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.7453416149068323,
|
| 2533 |
+
"grad_norm": 2.3780878490180735,
|
| 2534 |
+
"learning_rate": 1.8495607820914451e-06,
|
| 2535 |
+
"loss": 0.3563,
|
| 2536 |
+
"step": 1800
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.7474120082815735,
|
| 2540 |
+
"grad_norm": 2.5716332485081383,
|
| 2541 |
+
"learning_rate": 1.8215770508729602e-06,
|
| 2542 |
+
"loss": 0.3569,
|
| 2543 |
+
"step": 1805
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.7494824016563147,
|
| 2547 |
+
"grad_norm": 2.353530194343243,
|
| 2548 |
+
"learning_rate": 1.7937594045348666e-06,
|
| 2549 |
+
"loss": 0.3394,
|
| 2550 |
+
"step": 1810
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.7515527950310559,
|
| 2554 |
+
"grad_norm": 2.3569526754259504,
|
| 2555 |
+
"learning_rate": 1.7661092966566728e-06,
|
| 2556 |
+
"loss": 0.3481,
|
| 2557 |
+
"step": 1815
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.7536231884057971,
|
| 2561 |
+
"grad_norm": 2.649114709642696,
|
| 2562 |
+
"learning_rate": 1.7386281720633569e-06,
|
| 2563 |
+
"loss": 0.3643,
|
| 2564 |
+
"step": 1820
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.7556935817805382,
|
| 2568 |
+
"grad_norm": 2.460060042283697,
|
| 2569 |
+
"learning_rate": 1.7113174667498612e-06,
|
| 2570 |
+
"loss": 0.3576,
|
| 2571 |
+
"step": 1825
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.7577639751552795,
|
| 2575 |
+
"grad_norm": 2.339978220484055,
|
| 2576 |
+
"learning_rate": 1.6841786078060674e-06,
|
| 2577 |
+
"loss": 0.3558,
|
| 2578 |
+
"step": 1830
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.7598343685300207,
|
| 2582 |
+
"grad_norm": 2.4953672860958975,
|
| 2583 |
+
"learning_rate": 1.6572130133422204e-06,
|
| 2584 |
+
"loss": 0.3457,
|
| 2585 |
+
"step": 1835
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.7619047619047619,
|
| 2589 |
+
"grad_norm": 2.426133939030405,
|
| 2590 |
+
"learning_rate": 1.630422092414823e-06,
|
| 2591 |
+
"loss": 0.3482,
|
| 2592 |
+
"step": 1840
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.7639751552795031,
|
| 2596 |
+
"grad_norm": 2.459777432451568,
|
| 2597 |
+
"learning_rate": 1.6038072449530118e-06,
|
| 2598 |
+
"loss": 0.35,
|
| 2599 |
+
"step": 1845
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.7660455486542443,
|
| 2603 |
+
"grad_norm": 2.368741592685829,
|
| 2604 |
+
"learning_rate": 1.5773698616854065e-06,
|
| 2605 |
+
"loss": 0.3467,
|
| 2606 |
+
"step": 1850
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.7681159420289855,
|
| 2610 |
+
"grad_norm": 2.2896780011923488,
|
| 2611 |
+
"learning_rate": 1.5511113240674364e-06,
|
| 2612 |
+
"loss": 0.358,
|
| 2613 |
+
"step": 1855
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.7701863354037267,
|
| 2617 |
+
"grad_norm": 2.3730136202177143,
|
| 2618 |
+
"learning_rate": 1.5250330042091543e-06,
|
| 2619 |
+
"loss": 0.3481,
|
| 2620 |
+
"step": 1860
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.772256728778468,
|
| 2624 |
+
"grad_norm": 2.464479876236929,
|
| 2625 |
+
"learning_rate": 1.4991362648035368e-06,
|
| 2626 |
+
"loss": 0.3431,
|
| 2627 |
+
"step": 1865
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.7743271221532091,
|
| 2631 |
+
"grad_norm": 2.391982798119773,
|
| 2632 |
+
"learning_rate": 1.4734224590552844e-06,
|
| 2633 |
+
"loss": 0.3368,
|
| 2634 |
+
"step": 1870
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.7763975155279503,
|
| 2638 |
+
"grad_norm": 2.431774830394956,
|
| 2639 |
+
"learning_rate": 1.4478929306101025e-06,
|
| 2640 |
+
"loss": 0.3509,
|
| 2641 |
+
"step": 1875
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.7784679089026915,
|
| 2645 |
+
"grad_norm": 2.2880451270645614,
|
| 2646 |
+
"learning_rate": 1.4225490134844967e-06,
|
| 2647 |
+
"loss": 0.3376,
|
| 2648 |
+
"step": 1880
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.7805383022774327,
|
| 2652 |
+
"grad_norm": 2.541490709582719,
|
| 2653 |
+
"learning_rate": 1.3973920319960654e-06,
|
| 2654 |
+
"loss": 0.3422,
|
| 2655 |
+
"step": 1885
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.782608695652174,
|
| 2659 |
+
"grad_norm": 2.486644761227762,
|
| 2660 |
+
"learning_rate": 1.3724233006942972e-06,
|
| 2661 |
+
"loss": 0.3486,
|
| 2662 |
+
"step": 1890
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.7846790890269151,
|
| 2666 |
+
"grad_norm": 2.222907731738531,
|
| 2667 |
+
"learning_rate": 1.347644124291883e-06,
|
| 2668 |
+
"loss": 0.3401,
|
| 2669 |
+
"step": 1895
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.7867494824016563,
|
| 2673 |
+
"grad_norm": 2.503229276614086,
|
| 2674 |
+
"learning_rate": 1.323055797596532e-06,
|
| 2675 |
+
"loss": 0.3423,
|
| 2676 |
+
"step": 1900
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.7888198757763976,
|
| 2680 |
+
"grad_norm": 2.44333238565998,
|
| 2681 |
+
"learning_rate": 1.2986596054433255e-06,
|
| 2682 |
+
"loss": 0.3428,
|
| 2683 |
+
"step": 1905
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.7908902691511387,
|
| 2687 |
+
"grad_norm": 2.2546076468429592,
|
| 2688 |
+
"learning_rate": 1.2744568226275662e-06,
|
| 2689 |
+
"loss": 0.3367,
|
| 2690 |
+
"step": 1910
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.7929606625258799,
|
| 2694 |
+
"grad_norm": 2.241016793849244,
|
| 2695 |
+
"learning_rate": 1.2504487138381782e-06,
|
| 2696 |
+
"loss": 0.3301,
|
| 2697 |
+
"step": 1915
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.7950310559006211,
|
| 2701 |
+
"grad_norm": 2.472605862322622,
|
| 2702 |
+
"learning_rate": 1.2266365335916086e-06,
|
| 2703 |
+
"loss": 0.3378,
|
| 2704 |
+
"step": 1920
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.7971014492753623,
|
| 2708 |
+
"grad_norm": 2.235182498951894,
|
| 2709 |
+
"learning_rate": 1.203021526166287e-06,
|
| 2710 |
+
"loss": 0.3387,
|
| 2711 |
+
"step": 1925
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.7991718426501035,
|
| 2715 |
+
"grad_norm": 2.3511754008831516,
|
| 2716 |
+
"learning_rate": 1.179604925537598e-06,
|
| 2717 |
+
"loss": 0.3328,
|
| 2718 |
+
"step": 1930
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.8012422360248447,
|
| 2722 |
+
"grad_norm": 2.4302698271655863,
|
| 2723 |
+
"learning_rate": 1.1563879553134072e-06,
|
| 2724 |
+
"loss": 0.3479,
|
| 2725 |
+
"step": 1935
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.8033126293995859,
|
| 2729 |
+
"grad_norm": 2.4051091564094724,
|
| 2730 |
+
"learning_rate": 1.1333718286701162e-06,
|
| 2731 |
+
"loss": 0.345,
|
| 2732 |
+
"step": 1940
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.8053830227743272,
|
| 2736 |
+
"grad_norm": 2.4216508859806742,
|
| 2737 |
+
"learning_rate": 1.110557748289275e-06,
|
| 2738 |
+
"loss": 0.3362,
|
| 2739 |
+
"step": 1945
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.8074534161490683,
|
| 2743 |
+
"grad_norm": 2.518070754468628,
|
| 2744 |
+
"learning_rate": 1.0879469062947369e-06,
|
| 2745 |
+
"loss": 0.3406,
|
| 2746 |
+
"step": 1950
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.8095238095238095,
|
| 2750 |
+
"grad_norm": 2.4384803076253476,
|
| 2751 |
+
"learning_rate": 1.0655404841903648e-06,
|
| 2752 |
+
"loss": 0.3558,
|
| 2753 |
+
"step": 1955
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.8115942028985508,
|
| 2757 |
+
"grad_norm": 2.6344448059599412,
|
| 2758 |
+
"learning_rate": 1.0433396527982893e-06,
|
| 2759 |
+
"loss": 0.3346,
|
| 2760 |
+
"step": 1960
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.8136645962732919,
|
| 2764 |
+
"grad_norm": 2.3007277716039396,
|
| 2765 |
+
"learning_rate": 1.0213455721977356e-06,
|
| 2766 |
+
"loss": 0.3347,
|
| 2767 |
+
"step": 1965
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.8157349896480331,
|
| 2771 |
+
"grad_norm": 2.3043896070313137,
|
| 2772 |
+
"learning_rate": 9.995593916643981e-07,
|
| 2773 |
+
"loss": 0.3273,
|
| 2774 |
+
"step": 1970
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.8178053830227743,
|
| 2778 |
+
"grad_norm": 2.4269750741096465,
|
| 2779 |
+
"learning_rate": 9.779822496103902e-07,
|
| 2780 |
+
"loss": 0.3288,
|
| 2781 |
+
"step": 1975
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.8198757763975155,
|
| 2785 |
+
"grad_norm": 2.370989013074573,
|
| 2786 |
+
"learning_rate": 9.566152735247576e-07,
|
| 2787 |
+
"loss": 0.3322,
|
| 2788 |
+
"step": 1980
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.8219461697722568,
|
| 2792 |
+
"grad_norm": 2.2322391420672254,
|
| 2793 |
+
"learning_rate": 9.354595799145627e-07,
|
| 2794 |
+
"loss": 0.3302,
|
| 2795 |
+
"step": 1985
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.8240165631469979,
|
| 2799 |
+
"grad_norm": 2.3138733868403896,
|
| 2800 |
+
"learning_rate": 9.145162742465424e-07,
|
| 2801 |
+
"loss": 0.3281,
|
| 2802 |
+
"step": 1990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.8260869565217391,
|
| 2806 |
+
"grad_norm": 2.218401312393179,
|
| 2807 |
+
"learning_rate": 8.937864508893396e-07,
|
| 2808 |
+
"loss": 0.3263,
|
| 2809 |
+
"step": 1995
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.8281573498964804,
|
| 2813 |
+
"grad_norm": 2.501783210452481,
|
| 2814 |
+
"learning_rate": 8.732711930563259e-07,
|
| 2815 |
+
"loss": 0.3383,
|
| 2816 |
+
"step": 2000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.8302277432712215,
|
| 2820 |
+
"grad_norm": 2.339035923113422,
|
| 2821 |
+
"learning_rate": 8.529715727489912e-07,
|
| 2822 |
+
"loss": 0.3331,
|
| 2823 |
+
"step": 2005
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.8322981366459627,
|
| 2827 |
+
"grad_norm": 2.419266862744491,
|
| 2828 |
+
"learning_rate": 8.32888650700936e-07,
|
| 2829 |
+
"loss": 0.3326,
|
| 2830 |
+
"step": 2010
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.8343685300207039,
|
| 2834 |
+
"grad_norm": 2.3535716478661994,
|
| 2835 |
+
"learning_rate": 8.130234763224342e-07,
|
| 2836 |
+
"loss": 0.3383,
|
| 2837 |
+
"step": 2015
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.8364389233954451,
|
| 2841 |
+
"grad_norm": 2.348419640739673,
|
| 2842 |
+
"learning_rate": 7.93377087645607e-07,
|
| 2843 |
+
"loss": 0.3277,
|
| 2844 |
+
"step": 2020
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.8385093167701864,
|
| 2848 |
+
"grad_norm": 2.5091270230983422,
|
| 2849 |
+
"learning_rate": 7.739505112701751e-07,
|
| 2850 |
+
"loss": 0.3375,
|
| 2851 |
+
"step": 2025
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.8405797101449275,
|
| 2855 |
+
"grad_norm": 2.4359136230459524,
|
| 2856 |
+
"learning_rate": 7.547447623098191e-07,
|
| 2857 |
+
"loss": 0.3258,
|
| 2858 |
+
"step": 2030
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.8426501035196687,
|
| 2862 |
+
"grad_norm": 2.432506753046153,
|
| 2863 |
+
"learning_rate": 7.357608443391312e-07,
|
| 2864 |
+
"loss": 0.3291,
|
| 2865 |
+
"step": 2035
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.84472049689441,
|
| 2869 |
+
"grad_norm": 2.5646276059325706,
|
| 2870 |
+
"learning_rate": 7.169997493411774e-07,
|
| 2871 |
+
"loss": 0.3344,
|
| 2872 |
+
"step": 2040
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.8467908902691511,
|
| 2876 |
+
"grad_norm": 2.3486928663224,
|
| 2877 |
+
"learning_rate": 6.984624576556647e-07,
|
| 2878 |
+
"loss": 0.3271,
|
| 2879 |
+
"step": 2045
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.8488612836438924,
|
| 2883 |
+
"grad_norm": 2.203166381807686,
|
| 2884 |
+
"learning_rate": 6.801499379277115e-07,
|
| 2885 |
+
"loss": 0.3167,
|
| 2886 |
+
"step": 2050
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.8509316770186336,
|
| 2890 |
+
"grad_norm": 2.2436664305021385,
|
| 2891 |
+
"learning_rate": 6.620631470572314e-07,
|
| 2892 |
+
"loss": 0.3307,
|
| 2893 |
+
"step": 2055
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.8530020703933747,
|
| 2897 |
+
"grad_norm": 2.3594907816627875,
|
| 2898 |
+
"learning_rate": 6.442030301489338e-07,
|
| 2899 |
+
"loss": 0.3164,
|
| 2900 |
+
"step": 2060
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.855072463768116,
|
| 2904 |
+
"grad_norm": 2.256989647657127,
|
| 2905 |
+
"learning_rate": 6.265705204629402e-07,
|
| 2906 |
+
"loss": 0.3133,
|
| 2907 |
+
"step": 2065
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.8571428571428571,
|
| 2911 |
+
"grad_norm": 2.3571020694555798,
|
| 2912 |
+
"learning_rate": 6.091665393660112e-07,
|
| 2913 |
+
"loss": 0.3203,
|
| 2914 |
+
"step": 2070
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.8592132505175983,
|
| 2918 |
+
"grad_norm": 2.3091661794248157,
|
| 2919 |
+
"learning_rate": 5.919919962834069e-07,
|
| 2920 |
+
"loss": 0.3224,
|
| 2921 |
+
"step": 2075
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.8612836438923396,
|
| 2925 |
+
"grad_norm": 2.2740550403647894,
|
| 2926 |
+
"learning_rate": 5.750477886513656e-07,
|
| 2927 |
+
"loss": 0.3285,
|
| 2928 |
+
"step": 2080
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.8633540372670807,
|
| 2932 |
+
"grad_norm": 2.256740351456034,
|
| 2933 |
+
"learning_rate": 5.58334801870209e-07,
|
| 2934 |
+
"loss": 0.3178,
|
| 2935 |
+
"step": 2085
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.865424430641822,
|
| 2939 |
+
"grad_norm": 2.385870564024283,
|
| 2940 |
+
"learning_rate": 5.418539092580727e-07,
|
| 2941 |
+
"loss": 0.329,
|
| 2942 |
+
"step": 2090
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.8674948240165632,
|
| 2946 |
+
"grad_norm": 2.215079196146054,
|
| 2947 |
+
"learning_rate": 5.256059720052787e-07,
|
| 2948 |
+
"loss": 0.3303,
|
| 2949 |
+
"step": 2095
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.8695652173913043,
|
| 2953 |
+
"grad_norm": 2.2942059659033833,
|
| 2954 |
+
"learning_rate": 5.095918391293303e-07,
|
| 2955 |
+
"loss": 0.3279,
|
| 2956 |
+
"step": 2100
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.8716356107660456,
|
| 2960 |
+
"grad_norm": 2.5461731820432076,
|
| 2961 |
+
"learning_rate": 4.938123474305473e-07,
|
| 2962 |
+
"loss": 0.3222,
|
| 2963 |
+
"step": 2105
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.8737060041407867,
|
| 2967 |
+
"grad_norm": 2.4819509581487846,
|
| 2968 |
+
"learning_rate": 4.782683214483413e-07,
|
| 2969 |
+
"loss": 0.3185,
|
| 2970 |
+
"step": 2110
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.8757763975155279,
|
| 2974 |
+
"grad_norm": 2.357273267066927,
|
| 2975 |
+
"learning_rate": 4.629605734181331e-07,
|
| 2976 |
+
"loss": 0.316,
|
| 2977 |
+
"step": 2115
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.8778467908902692,
|
| 2981 |
+
"grad_norm": 2.26836858721712,
|
| 2982 |
+
"learning_rate": 4.478899032289047e-07,
|
| 2983 |
+
"loss": 0.3191,
|
| 2984 |
+
"step": 2120
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.8799171842650103,
|
| 2988 |
+
"grad_norm": 2.339531232436072,
|
| 2989 |
+
"learning_rate": 4.3305709838140755e-07,
|
| 2990 |
+
"loss": 0.3221,
|
| 2991 |
+
"step": 2125
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.8819875776397516,
|
| 2995 |
+
"grad_norm": 2.354666153598989,
|
| 2996 |
+
"learning_rate": 4.1846293394700645e-07,
|
| 2997 |
+
"loss": 0.3291,
|
| 2998 |
+
"step": 2130
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.8840579710144928,
|
| 3002 |
+
"grad_norm": 2.5023494902865484,
|
| 3003 |
+
"learning_rate": 4.041081725271856e-07,
|
| 3004 |
+
"loss": 0.3319,
|
| 3005 |
+
"step": 2135
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.8861283643892339,
|
| 3009 |
+
"grad_norm": 2.660892225845494,
|
| 3010 |
+
"learning_rate": 3.8999356421369426e-07,
|
| 3011 |
+
"loss": 0.3273,
|
| 3012 |
+
"step": 2140
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.8881987577639752,
|
| 3016 |
+
"grad_norm": 2.3866941174661993,
|
| 3017 |
+
"learning_rate": 3.7611984654935707e-07,
|
| 3018 |
+
"loss": 0.3218,
|
| 3019 |
+
"step": 2145
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.8902691511387164,
|
| 3023 |
+
"grad_norm": 2.3348475610176,
|
| 3024 |
+
"learning_rate": 3.6248774448952695e-07,
|
| 3025 |
+
"loss": 0.3191,
|
| 3026 |
+
"step": 2150
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.8923395445134575,
|
| 3030 |
+
"grad_norm": 2.441065668351126,
|
| 3031 |
+
"learning_rate": 3.490979703642117e-07,
|
| 3032 |
+
"loss": 0.336,
|
| 3033 |
+
"step": 2155
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.8944099378881988,
|
| 3037 |
+
"grad_norm": 2.273245811499413,
|
| 3038 |
+
"learning_rate": 3.35951223840848e-07,
|
| 3039 |
+
"loss": 0.3136,
|
| 3040 |
+
"step": 2160
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.8964803312629399,
|
| 3044 |
+
"grad_norm": 2.359301186608738,
|
| 3045 |
+
"learning_rate": 3.2304819188773815e-07,
|
| 3046 |
+
"loss": 0.3153,
|
| 3047 |
+
"step": 2165
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.8985507246376812,
|
| 3051 |
+
"grad_norm": 2.4349069583068226,
|
| 3052 |
+
"learning_rate": 3.1038954873816027e-07,
|
| 3053 |
+
"loss": 0.325,
|
| 3054 |
+
"step": 2170
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.9006211180124224,
|
| 3058 |
+
"grad_norm": 2.221649283047898,
|
| 3059 |
+
"learning_rate": 2.9797595585512986e-07,
|
| 3060 |
+
"loss": 0.3162,
|
| 3061 |
+
"step": 2175
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.9026915113871635,
|
| 3065 |
+
"grad_norm": 2.4365346639466976,
|
| 3066 |
+
"learning_rate": 2.858080618968423e-07,
|
| 3067 |
+
"loss": 0.3211,
|
| 3068 |
+
"step": 2180
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.9047619047619048,
|
| 3072 |
+
"grad_norm": 2.2949548389538226,
|
| 3073 |
+
"learning_rate": 2.738865026827714e-07,
|
| 3074 |
+
"loss": 0.3083,
|
| 3075 |
+
"step": 2185
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.906832298136646,
|
| 3079 |
+
"grad_norm": 2.354756885436125,
|
| 3080 |
+
"learning_rate": 2.6221190116045126e-07,
|
| 3081 |
+
"loss": 0.3213,
|
| 3082 |
+
"step": 2190
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.9089026915113871,
|
| 3086 |
+
"grad_norm": 2.385659076635714,
|
| 3087 |
+
"learning_rate": 2.507848673729224e-07,
|
| 3088 |
+
"loss": 0.3175,
|
| 3089 |
+
"step": 2195
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 0.9109730848861284,
|
| 3093 |
+
"grad_norm": 2.4203042137247333,
|
| 3094 |
+
"learning_rate": 2.3960599842685394e-07,
|
| 3095 |
+
"loss": 0.3243,
|
| 3096 |
+
"step": 2200
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 0.9130434782608695,
|
| 3100 |
+
"grad_norm": 2.4504849617910796,
|
| 3101 |
+
"learning_rate": 2.2867587846134188e-07,
|
| 3102 |
+
"loss": 0.3164,
|
| 3103 |
+
"step": 2205
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 0.9151138716356108,
|
| 3107 |
+
"grad_norm": 2.3686013787646165,
|
| 3108 |
+
"learning_rate": 2.179950786173879e-07,
|
| 3109 |
+
"loss": 0.3174,
|
| 3110 |
+
"step": 2210
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 0.917184265010352,
|
| 3114 |
+
"grad_norm": 2.271535894048282,
|
| 3115 |
+
"learning_rate": 2.07564157008055e-07,
|
| 3116 |
+
"loss": 0.3223,
|
| 3117 |
+
"step": 2215
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 0.9192546583850931,
|
| 3121 |
+
"grad_norm": 2.3541613140984894,
|
| 3122 |
+
"learning_rate": 1.9738365868930188e-07,
|
| 3123 |
+
"loss": 0.3207,
|
| 3124 |
+
"step": 2220
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.9213250517598344,
|
| 3128 |
+
"grad_norm": 2.3941528051254504,
|
| 3129 |
+
"learning_rate": 1.8745411563150128e-07,
|
| 3130 |
+
"loss": 0.3161,
|
| 3131 |
+
"step": 2225
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 0.9233954451345756,
|
| 3135 |
+
"grad_norm": 2.261376584339555,
|
| 3136 |
+
"learning_rate": 1.7777604669164727e-07,
|
| 3137 |
+
"loss": 0.3138,
|
| 3138 |
+
"step": 2230
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 0.9254658385093167,
|
| 3142 |
+
"grad_norm": 2.3190443101075973,
|
| 3143 |
+
"learning_rate": 1.6834995758623685e-07,
|
| 3144 |
+
"loss": 0.3155,
|
| 3145 |
+
"step": 2235
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 0.927536231884058,
|
| 3149 |
+
"grad_norm": 2.4743221595041147,
|
| 3150 |
+
"learning_rate": 1.5917634086484778e-07,
|
| 3151 |
+
"loss": 0.3194,
|
| 3152 |
+
"step": 2240
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 0.9296066252587992,
|
| 3156 |
+
"grad_norm": 2.3971910739689695,
|
| 3157 |
+
"learning_rate": 1.5025567588439925e-07,
|
| 3158 |
+
"loss": 0.3185,
|
| 3159 |
+
"step": 2245
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 0.9316770186335404,
|
| 3163 |
+
"grad_norm": 2.3405920624750096,
|
| 3164 |
+
"learning_rate": 1.415884287841074e-07,
|
| 3165 |
+
"loss": 0.3274,
|
| 3166 |
+
"step": 2250
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.9337474120082816,
|
| 3170 |
+
"grad_norm": 2.479303603019371,
|
| 3171 |
+
"learning_rate": 1.331750524611225e-07,
|
| 3172 |
+
"loss": 0.3145,
|
| 3173 |
+
"step": 2255
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 0.9358178053830227,
|
| 3177 |
+
"grad_norm": 2.426253073867152,
|
| 3178 |
+
"learning_rate": 1.2501598654686532e-07,
|
| 3179 |
+
"loss": 0.3147,
|
| 3180 |
+
"step": 2260
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 0.937888198757764,
|
| 3184 |
+
"grad_norm": 2.3156992646831136,
|
| 3185 |
+
"learning_rate": 1.1711165738405749e-07,
|
| 3186 |
+
"loss": 0.3205,
|
| 3187 |
+
"step": 2265
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 0.9399585921325052,
|
| 3191 |
+
"grad_norm": 2.5003604779718707,
|
| 3192 |
+
"learning_rate": 1.0946247800443899e-07,
|
| 3193 |
+
"loss": 0.318,
|
| 3194 |
+
"step": 2270
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 0.9420289855072463,
|
| 3198 |
+
"grad_norm": 2.4811177398484157,
|
| 3199 |
+
"learning_rate": 1.0206884810718964e-07,
|
| 3200 |
+
"loss": 0.3211,
|
| 3201 |
+
"step": 2275
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 0.9440993788819876,
|
| 3205 |
+
"grad_norm": 2.3564288178465156,
|
| 3206 |
+
"learning_rate": 9.493115403803876e-08,
|
| 3207 |
+
"loss": 0.3215,
|
| 3208 |
+
"step": 2280
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 0.9461697722567288,
|
| 3212 |
+
"grad_norm": 2.275026862745937,
|
| 3213 |
+
"learning_rate": 8.804976876908178e-08,
|
| 3214 |
+
"loss": 0.3184,
|
| 3215 |
+
"step": 2285
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 0.94824016563147,
|
| 3219 |
+
"grad_norm": 2.363783479958005,
|
| 3220 |
+
"learning_rate": 8.142505187928861e-08,
|
| 3221 |
+
"loss": 0.3299,
|
| 3222 |
+
"step": 2290
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 0.9503105590062112,
|
| 3226 |
+
"grad_norm": 2.3282452854981224,
|
| 3227 |
+
"learning_rate": 7.505734953571486e-08,
|
| 3228 |
+
"loss": 0.3104,
|
| 3229 |
+
"step": 2295
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 0.9523809523809523,
|
| 3233 |
+
"grad_norm": 2.5588066429543237,
|
| 3234 |
+
"learning_rate": 6.894699447541231e-08,
|
| 3235 |
+
"loss": 0.3216,
|
| 3236 |
+
"step": 2300
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 0.9544513457556936,
|
| 3240 |
+
"grad_norm": 2.3198859845159956,
|
| 3241 |
+
"learning_rate": 6.309430598804234e-08,
|
| 3242 |
+
"loss": 0.3097,
|
| 3243 |
+
"step": 2305
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 0.9565217391304348,
|
| 3247 |
+
"grad_norm": 2.41344387465687,
|
| 3248 |
+
"learning_rate": 5.7499589899193134e-08,
|
| 3249 |
+
"loss": 0.3153,
|
| 3250 |
+
"step": 2310
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 0.9585921325051759,
|
| 3254 |
+
"grad_norm": 2.5822922013181175,
|
| 3255 |
+
"learning_rate": 5.216313855439858e-08,
|
| 3256 |
+
"loss": 0.3193,
|
| 3257 |
+
"step": 2315
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 0.9606625258799172,
|
| 3261 |
+
"grad_norm": 2.422135687600398,
|
| 3262 |
+
"learning_rate": 4.7085230803860513e-08,
|
| 3263 |
+
"loss": 0.3185,
|
| 3264 |
+
"step": 2320
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 0.9627329192546584,
|
| 3268 |
+
"grad_norm": 2.4718246510689803,
|
| 3269 |
+
"learning_rate": 4.2266131987880876e-08,
|
| 3270 |
+
"loss": 0.3295,
|
| 3271 |
+
"step": 2325
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 0.9648033126293996,
|
| 3275 |
+
"grad_norm": 2.495970339302664,
|
| 3276 |
+
"learning_rate": 3.770609392299285e-08,
|
| 3277 |
+
"loss": 0.3132,
|
| 3278 |
+
"step": 2330
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 0.9668737060041408,
|
| 3282 |
+
"grad_norm": 2.4117713089348682,
|
| 3283 |
+
"learning_rate": 3.340535488880525e-08,
|
| 3284 |
+
"loss": 0.3098,
|
| 3285 |
+
"step": 2335
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 0.968944099378882,
|
| 3289 |
+
"grad_norm": 2.4964819297620653,
|
| 3290 |
+
"learning_rate": 2.9364139615550846e-08,
|
| 3291 |
+
"loss": 0.3168,
|
| 3292 |
+
"step": 2340
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 0.9710144927536232,
|
| 3296 |
+
"grad_norm": 2.3391947122526577,
|
| 3297 |
+
"learning_rate": 2.5582659272341827e-08,
|
| 3298 |
+
"loss": 0.3196,
|
| 3299 |
+
"step": 2345
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 0.9730848861283644,
|
| 3303 |
+
"grad_norm": 2.506425184281738,
|
| 3304 |
+
"learning_rate": 2.206111145613865e-08,
|
| 3305 |
+
"loss": 0.3099,
|
| 3306 |
+
"step": 2350
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 0.9751552795031055,
|
| 3310 |
+
"grad_norm": 2.179329803243906,
|
| 3311 |
+
"learning_rate": 1.8799680181421087e-08,
|
| 3312 |
+
"loss": 0.3177,
|
| 3313 |
+
"step": 2355
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 0.9772256728778468,
|
| 3317 |
+
"grad_norm": 2.3142698292838753,
|
| 3318 |
+
"learning_rate": 1.5798535870575893e-08,
|
| 3319 |
+
"loss": 0.3234,
|
| 3320 |
+
"step": 2360
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 0.979296066252588,
|
| 3324 |
+
"grad_norm": 2.383754154510841,
|
| 3325 |
+
"learning_rate": 1.30578353449895e-08,
|
| 3326 |
+
"loss": 0.3127,
|
| 3327 |
+
"step": 2365
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 0.9813664596273292,
|
| 3331 |
+
"grad_norm": 2.370736953831053,
|
| 3332 |
+
"learning_rate": 1.0577721816854015e-08,
|
| 3333 |
+
"loss": 0.3091,
|
| 3334 |
+
"step": 2370
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 0.9834368530020704,
|
| 3338 |
+
"grad_norm": 2.517651342295925,
|
| 3339 |
+
"learning_rate": 8.35832488168431e-09,
|
| 3340 |
+
"loss": 0.3173,
|
| 3341 |
+
"step": 2375
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 0.9855072463768116,
|
| 3345 |
+
"grad_norm": 2.3821914184277295,
|
| 3346 |
+
"learning_rate": 6.399760511546227e-09,
|
| 3347 |
+
"loss": 0.3207,
|
| 3348 |
+
"step": 2380
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 0.9875776397515528,
|
| 3352 |
+
"grad_norm": 2.315898281851513,
|
| 3353 |
+
"learning_rate": 4.7021310489953065e-09,
|
| 3354 |
+
"loss": 0.3185,
|
| 3355 |
+
"step": 2385
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.989648033126294,
|
| 3359 |
+
"grad_norm": 2.4911092964288644,
|
| 3360 |
+
"learning_rate": 3.265525201731623e-09,
|
| 3361 |
+
"loss": 0.3188,
|
| 3362 |
+
"step": 2390
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.9917184265010351,
|
| 3366 |
+
"grad_norm": 2.332775121742113,
|
| 3367 |
+
"learning_rate": 2.090018037960717e-09,
|
| 3368 |
+
"loss": 0.3149,
|
| 3369 |
+
"step": 2395
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.9937888198757764,
|
| 3373 |
+
"grad_norm": 2.269380119205785,
|
| 3374 |
+
"learning_rate": 1.1756709824750633e-09,
|
| 3375 |
+
"loss": 0.321,
|
| 3376 |
+
"step": 2400
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 0.9958592132505176,
|
| 3380 |
+
"grad_norm": 2.1851128802901725,
|
| 3381 |
+
"learning_rate": 5.225318134399749e-10,
|
| 3382 |
+
"loss": 0.3055,
|
| 3383 |
+
"step": 2405
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.9979296066252588,
|
| 3387 |
+
"grad_norm": 2.347230158447268,
|
| 3388 |
+
"learning_rate": 1.306346599011521e-10,
|
| 3389 |
+
"loss": 0.3134,
|
| 3390 |
+
"step": 2410
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 1.0,
|
| 3394 |
+
"grad_norm": 2.5576033751474125,
|
| 3395 |
+
"learning_rate": 0.0,
|
| 3396 |
+
"loss": 0.3164,
|
| 3397 |
+
"step": 2415
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 1.0,
|
| 3401 |
+
"eval_runtime": 2.7619,
|
| 3402 |
+
"eval_samples_per_second": 3.621,
|
| 3403 |
+
"eval_steps_per_second": 1.086,
|
| 3404 |
+
"step": 2415
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 1.0,
|
| 3408 |
+
"step": 2415,
|
| 3409 |
+
"total_flos": 252825934233600.0,
|
| 3410 |
+
"train_loss": 0.5146465574979289,
|
| 3411 |
+
"train_runtime": 25682.4245,
|
| 3412 |
+
"train_samples_per_second": 1.504,
|
| 3413 |
+
"train_steps_per_second": 0.094
|
| 3414 |
+
}
|
| 3415 |
+
],
|
| 3416 |
+
"logging_steps": 5,
|
| 3417 |
+
"max_steps": 2415,
|
| 3418 |
+
"num_input_tokens_seen": 0,
|
| 3419 |
+
"num_train_epochs": 1,
|
| 3420 |
+
"save_steps": 100,
|
| 3421 |
+
"stateful_callbacks": {
|
| 3422 |
+
"TrainerControl": {
|
| 3423 |
+
"args": {
|
| 3424 |
+
"should_epoch_stop": false,
|
| 3425 |
+
"should_evaluate": false,
|
| 3426 |
+
"should_log": false,
|
| 3427 |
+
"should_save": true,
|
| 3428 |
+
"should_training_stop": true
|
| 3429 |
+
},
|
| 3430 |
+
"attributes": {}
|
| 3431 |
+
}
|
| 3432 |
+
},
|
| 3433 |
+
"total_flos": 252825934233600.0,
|
| 3434 |
+
"train_batch_size": 4,
|
| 3435 |
+
"trial_name": null,
|
| 3436 |
+
"trial_params": null
|
| 3437 |
+
}
|