Model save
Browse files- README.md +60 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- runs/Oct01_08-05-05_COE-CS-sv003/events.out.tfevents.1727769928.COE-CS-sv003.1418179.0 +2 -2
- train_results.json +9 -0
- trainer_state.json +2898 -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: feedback_p0.1_seed42_level2_raredirtybatch16_rubric
|
| 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 |
+
# feedback_p0.1_seed42_level2_raredirtybatch16_rubric
|
| 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 4.45.1
|
| 58 |
+
- Pytorch 2.4.1+cu121
|
| 59 |
+
- Datasets 3.0.1
|
| 60 |
+
- Tokenizers 0.20.0
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 212729730170880.0,
|
| 4 |
+
"train_loss": 0.4844009231218672,
|
| 5 |
+
"train_runtime": 18284.3129,
|
| 6 |
+
"train_samples": 107952,
|
| 7 |
+
"train_samples_per_second": 1.778,
|
| 8 |
+
"train_steps_per_second": 0.111
|
| 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": "4.45.1"
|
| 6 |
+
}
|
runs/Oct01_08-05-05_COE-CS-sv003/events.out.tfevents.1727769928.COE-CS-sv003.1418179.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ad82788c304c8264189bcb78294e805a3ed04606e29c32797fb97618c838748
|
| 3 |
+
size 92171
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 212729730170880.0,
|
| 4 |
+
"train_loss": 0.4844009231218672,
|
| 5 |
+
"train_runtime": 18284.3129,
|
| 6 |
+
"train_samples": 107952,
|
| 7 |
+
"train_samples_per_second": 1.778,
|
| 8 |
+
"train_steps_per_second": 0.111
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2898 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2032,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0004921259842519685,
|
| 13 |
+
"grad_norm": 23.289178620841536,
|
| 14 |
+
"learning_rate": 4.901960784313726e-08,
|
| 15 |
+
"loss": 1.3042,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0024606299212598425,
|
| 20 |
+
"grad_norm": 22.578339014503392,
|
| 21 |
+
"learning_rate": 2.4509803921568627e-07,
|
| 22 |
+
"loss": 1.3271,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.004921259842519685,
|
| 27 |
+
"grad_norm": 15.254145113240074,
|
| 28 |
+
"learning_rate": 4.901960784313725e-07,
|
| 29 |
+
"loss": 1.2787,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0073818897637795275,
|
| 34 |
+
"grad_norm": 11.623341812407018,
|
| 35 |
+
"learning_rate": 7.352941176470589e-07,
|
| 36 |
+
"loss": 1.1561,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.00984251968503937,
|
| 41 |
+
"grad_norm": 9.64376871387551,
|
| 42 |
+
"learning_rate": 9.80392156862745e-07,
|
| 43 |
+
"loss": 1.034,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.012303149606299213,
|
| 48 |
+
"grad_norm": 3.4638336177505513,
|
| 49 |
+
"learning_rate": 1.2254901960784314e-06,
|
| 50 |
+
"loss": 0.9177,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.014763779527559055,
|
| 55 |
+
"grad_norm": 3.5862986630527347,
|
| 56 |
+
"learning_rate": 1.4705882352941177e-06,
|
| 57 |
+
"loss": 0.8909,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.0172244094488189,
|
| 62 |
+
"grad_norm": 2.9675977602993884,
|
| 63 |
+
"learning_rate": 1.715686274509804e-06,
|
| 64 |
+
"loss": 0.8673,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.01968503937007874,
|
| 69 |
+
"grad_norm": 3.0016020610328322,
|
| 70 |
+
"learning_rate": 1.96078431372549e-06,
|
| 71 |
+
"loss": 0.8523,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.02214566929133858,
|
| 76 |
+
"grad_norm": 3.044795051743997,
|
| 77 |
+
"learning_rate": 2.2058823529411767e-06,
|
| 78 |
+
"loss": 0.836,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.024606299212598427,
|
| 83 |
+
"grad_norm": 2.9272556512306047,
|
| 84 |
+
"learning_rate": 2.450980392156863e-06,
|
| 85 |
+
"loss": 0.8255,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.027066929133858268,
|
| 90 |
+
"grad_norm": 2.9326611094737216,
|
| 91 |
+
"learning_rate": 2.696078431372549e-06,
|
| 92 |
+
"loss": 0.8163,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.02952755905511811,
|
| 97 |
+
"grad_norm": 3.1751190158414553,
|
| 98 |
+
"learning_rate": 2.9411764705882355e-06,
|
| 99 |
+
"loss": 0.8098,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.03198818897637795,
|
| 104 |
+
"grad_norm": 3.267001091193129,
|
| 105 |
+
"learning_rate": 3.1862745098039216e-06,
|
| 106 |
+
"loss": 0.7963,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.0344488188976378,
|
| 111 |
+
"grad_norm": 3.2037792452122074,
|
| 112 |
+
"learning_rate": 3.431372549019608e-06,
|
| 113 |
+
"loss": 0.7904,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.036909448818897635,
|
| 118 |
+
"grad_norm": 3.1335298386999817,
|
| 119 |
+
"learning_rate": 3.6764705882352946e-06,
|
| 120 |
+
"loss": 0.7767,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.03937007874015748,
|
| 125 |
+
"grad_norm": 3.0748371474571066,
|
| 126 |
+
"learning_rate": 3.92156862745098e-06,
|
| 127 |
+
"loss": 0.7566,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.041830708661417325,
|
| 132 |
+
"grad_norm": 3.168808058148803,
|
| 133 |
+
"learning_rate": 4.166666666666667e-06,
|
| 134 |
+
"loss": 0.7642,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.04429133858267716,
|
| 139 |
+
"grad_norm": 3.099061017093863,
|
| 140 |
+
"learning_rate": 4.411764705882353e-06,
|
| 141 |
+
"loss": 0.7567,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.04675196850393701,
|
| 146 |
+
"grad_norm": 3.1397025998475545,
|
| 147 |
+
"learning_rate": 4.65686274509804e-06,
|
| 148 |
+
"loss": 0.7489,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.04921259842519685,
|
| 153 |
+
"grad_norm": 3.267009036532827,
|
| 154 |
+
"learning_rate": 4.901960784313726e-06,
|
| 155 |
+
"loss": 0.7428,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.05167322834645669,
|
| 160 |
+
"grad_norm": 3.0814329808676137,
|
| 161 |
+
"learning_rate": 5.147058823529411e-06,
|
| 162 |
+
"loss": 0.739,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.054133858267716536,
|
| 167 |
+
"grad_norm": 2.922079677455855,
|
| 168 |
+
"learning_rate": 5.392156862745098e-06,
|
| 169 |
+
"loss": 0.7285,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.056594488188976375,
|
| 174 |
+
"grad_norm": 3.1140432595060723,
|
| 175 |
+
"learning_rate": 5.637254901960784e-06,
|
| 176 |
+
"loss": 0.7279,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.05905511811023622,
|
| 181 |
+
"grad_norm": 2.890134615014165,
|
| 182 |
+
"learning_rate": 5.882352941176471e-06,
|
| 183 |
+
"loss": 0.7061,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.061515748031496065,
|
| 188 |
+
"grad_norm": 3.003266954139478,
|
| 189 |
+
"learning_rate": 6.1274509803921575e-06,
|
| 190 |
+
"loss": 0.7124,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.0639763779527559,
|
| 195 |
+
"grad_norm": 3.490944793138708,
|
| 196 |
+
"learning_rate": 6.372549019607843e-06,
|
| 197 |
+
"loss": 0.7,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.06643700787401575,
|
| 202 |
+
"grad_norm": 2.957015579169823,
|
| 203 |
+
"learning_rate": 6.61764705882353e-06,
|
| 204 |
+
"loss": 0.7109,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.0688976377952756,
|
| 209 |
+
"grad_norm": 2.9802255807893343,
|
| 210 |
+
"learning_rate": 6.862745098039216e-06,
|
| 211 |
+
"loss": 0.7307,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.07135826771653543,
|
| 216 |
+
"grad_norm": 2.979885982433863,
|
| 217 |
+
"learning_rate": 7.107843137254903e-06,
|
| 218 |
+
"loss": 0.6999,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.07381889763779527,
|
| 223 |
+
"grad_norm": 3.4724625938807483,
|
| 224 |
+
"learning_rate": 7.352941176470589e-06,
|
| 225 |
+
"loss": 0.7018,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.07627952755905512,
|
| 230 |
+
"grad_norm": 2.8756072791555822,
|
| 231 |
+
"learning_rate": 7.598039215686275e-06,
|
| 232 |
+
"loss": 0.6946,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.07874015748031496,
|
| 237 |
+
"grad_norm": 2.888280542800706,
|
| 238 |
+
"learning_rate": 7.84313725490196e-06,
|
| 239 |
+
"loss": 0.6953,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.0812007874015748,
|
| 244 |
+
"grad_norm": 3.0409331177011176,
|
| 245 |
+
"learning_rate": 8.088235294117648e-06,
|
| 246 |
+
"loss": 0.693,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.08366141732283465,
|
| 251 |
+
"grad_norm": 2.905841883072287,
|
| 252 |
+
"learning_rate": 8.333333333333334e-06,
|
| 253 |
+
"loss": 0.6995,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.08612204724409449,
|
| 258 |
+
"grad_norm": 2.8022565504484476,
|
| 259 |
+
"learning_rate": 8.57843137254902e-06,
|
| 260 |
+
"loss": 0.6941,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.08858267716535433,
|
| 265 |
+
"grad_norm": 2.7594498957096523,
|
| 266 |
+
"learning_rate": 8.823529411764707e-06,
|
| 267 |
+
"loss": 0.6852,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.09104330708661418,
|
| 272 |
+
"grad_norm": 2.928739510451242,
|
| 273 |
+
"learning_rate": 9.068627450980392e-06,
|
| 274 |
+
"loss": 0.6849,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.09350393700787402,
|
| 279 |
+
"grad_norm": 2.922241571597126,
|
| 280 |
+
"learning_rate": 9.31372549019608e-06,
|
| 281 |
+
"loss": 0.6961,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.09596456692913385,
|
| 286 |
+
"grad_norm": 2.733962541795604,
|
| 287 |
+
"learning_rate": 9.558823529411766e-06,
|
| 288 |
+
"loss": 0.6964,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.0984251968503937,
|
| 293 |
+
"grad_norm": 3.3436619097327585,
|
| 294 |
+
"learning_rate": 9.803921568627451e-06,
|
| 295 |
+
"loss": 0.6903,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.10088582677165354,
|
| 300 |
+
"grad_norm": 2.819756078395206,
|
| 301 |
+
"learning_rate": 9.999992616075212e-06,
|
| 302 |
+
"loss": 0.6871,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.10334645669291338,
|
| 307 |
+
"grad_norm": 2.8245769395745217,
|
| 308 |
+
"learning_rate": 9.999734180997554e-06,
|
| 309 |
+
"loss": 0.7001,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.10580708661417323,
|
| 314 |
+
"grad_norm": 2.8711422702525207,
|
| 315 |
+
"learning_rate": 9.999106571489132e-06,
|
| 316 |
+
"loss": 0.6851,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.10826771653543307,
|
| 321 |
+
"grad_norm": 3.0388643695545077,
|
| 322 |
+
"learning_rate": 9.998109833891883e-06,
|
| 323 |
+
"loss": 0.6927,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.11072834645669291,
|
| 328 |
+
"grad_norm": 2.7895631269121393,
|
| 329 |
+
"learning_rate": 9.996744041803731e-06,
|
| 330 |
+
"loss": 0.6716,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.11318897637795275,
|
| 335 |
+
"grad_norm": 2.686288810148222,
|
| 336 |
+
"learning_rate": 9.995009296073138e-06,
|
| 337 |
+
"loss": 0.6725,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.1156496062992126,
|
| 342 |
+
"grad_norm": 2.800082841358338,
|
| 343 |
+
"learning_rate": 9.992905724791669e-06,
|
| 344 |
+
"loss": 0.6717,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.11811023622047244,
|
| 349 |
+
"grad_norm": 2.566573753796988,
|
| 350 |
+
"learning_rate": 9.990433483284527e-06,
|
| 351 |
+
"loss": 0.6666,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.12057086614173228,
|
| 356 |
+
"grad_norm": 2.6155377853742467,
|
| 357 |
+
"learning_rate": 9.987592754099086e-06,
|
| 358 |
+
"loss": 0.6724,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.12303149606299213,
|
| 363 |
+
"grad_norm": 2.6609874435959795,
|
| 364 |
+
"learning_rate": 9.984383746991416e-06,
|
| 365 |
+
"loss": 0.6756,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.12549212598425197,
|
| 370 |
+
"grad_norm": 2.651450027647874,
|
| 371 |
+
"learning_rate": 9.980806698910787e-06,
|
| 372 |
+
"loss": 0.6764,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.1279527559055118,
|
| 377 |
+
"grad_norm": 2.738239456171724,
|
| 378 |
+
"learning_rate": 9.976861873982177e-06,
|
| 379 |
+
"loss": 0.6858,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.13041338582677164,
|
| 384 |
+
"grad_norm": 2.5749543275223767,
|
| 385 |
+
"learning_rate": 9.972549563486776e-06,
|
| 386 |
+
"loss": 0.681,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.1328740157480315,
|
| 391 |
+
"grad_norm": 2.54868336015626,
|
| 392 |
+
"learning_rate": 9.967870085840463e-06,
|
| 393 |
+
"loss": 0.6822,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.13533464566929135,
|
| 398 |
+
"grad_norm": 2.348509038708982,
|
| 399 |
+
"learning_rate": 9.962823786570306e-06,
|
| 400 |
+
"loss": 0.6701,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.1377952755905512,
|
| 405 |
+
"grad_norm": 2.545785789782642,
|
| 406 |
+
"learning_rate": 9.95741103828905e-06,
|
| 407 |
+
"loss": 0.6672,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.14025590551181102,
|
| 412 |
+
"grad_norm": 2.724173815762578,
|
| 413 |
+
"learning_rate": 9.951632240667592e-06,
|
| 414 |
+
"loss": 0.6659,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.14271653543307086,
|
| 419 |
+
"grad_norm": 2.6494992216672197,
|
| 420 |
+
"learning_rate": 9.945487820405487e-06,
|
| 421 |
+
"loss": 0.6642,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.1451771653543307,
|
| 426 |
+
"grad_norm": 2.477475682520284,
|
| 427 |
+
"learning_rate": 9.938978231199419e-06,
|
| 428 |
+
"loss": 0.6794,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.14763779527559054,
|
| 433 |
+
"grad_norm": 2.589636626095672,
|
| 434 |
+
"learning_rate": 9.932103953709724e-06,
|
| 435 |
+
"loss": 0.6583,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.1500984251968504,
|
| 440 |
+
"grad_norm": 2.53566392464488,
|
| 441 |
+
"learning_rate": 9.924865495524884e-06,
|
| 442 |
+
"loss": 0.6737,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.15255905511811024,
|
| 447 |
+
"grad_norm": 3.013667820537959,
|
| 448 |
+
"learning_rate": 9.917263391124046e-06,
|
| 449 |
+
"loss": 0.6475,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.15501968503937008,
|
| 454 |
+
"grad_norm": 2.5986666259102837,
|
| 455 |
+
"learning_rate": 9.90929820183757e-06,
|
| 456 |
+
"loss": 0.6496,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.15748031496062992,
|
| 461 |
+
"grad_norm": 3.1918308775019715,
|
| 462 |
+
"learning_rate": 9.900970515805564e-06,
|
| 463 |
+
"loss": 0.6527,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.15994094488188976,
|
| 468 |
+
"grad_norm": 2.58199571141749,
|
| 469 |
+
"learning_rate": 9.892280947934472e-06,
|
| 470 |
+
"loss": 0.6507,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.1624015748031496,
|
| 475 |
+
"grad_norm": 2.5096275929429828,
|
| 476 |
+
"learning_rate": 9.883230139851656e-06,
|
| 477 |
+
"loss": 0.6493,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.16486220472440946,
|
| 482 |
+
"grad_norm": 2.543717236052826,
|
| 483 |
+
"learning_rate": 9.873818759858034e-06,
|
| 484 |
+
"loss": 0.6398,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.1673228346456693,
|
| 489 |
+
"grad_norm": 2.3455715291925987,
|
| 490 |
+
"learning_rate": 9.864047502878717e-06,
|
| 491 |
+
"loss": 0.6509,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.16978346456692914,
|
| 496 |
+
"grad_norm": 2.7443847653291606,
|
| 497 |
+
"learning_rate": 9.853917090411713e-06,
|
| 498 |
+
"loss": 0.6587,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.17224409448818898,
|
| 503 |
+
"grad_norm": 2.574575364550669,
|
| 504 |
+
"learning_rate": 9.84342827047464e-06,
|
| 505 |
+
"loss": 0.6351,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.1747047244094488,
|
| 510 |
+
"grad_norm": 2.555747928874768,
|
| 511 |
+
"learning_rate": 9.832581817549497e-06,
|
| 512 |
+
"loss": 0.6223,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.17716535433070865,
|
| 517 |
+
"grad_norm": 2.3868178771545834,
|
| 518 |
+
"learning_rate": 9.821378532525479e-06,
|
| 519 |
+
"loss": 0.6423,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.1796259842519685,
|
| 524 |
+
"grad_norm": 2.3906560211167434,
|
| 525 |
+
"learning_rate": 9.809819242639841e-06,
|
| 526 |
+
"loss": 0.6203,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.18208661417322836,
|
| 531 |
+
"grad_norm": 2.613672178779019,
|
| 532 |
+
"learning_rate": 9.79790480141681e-06,
|
| 533 |
+
"loss": 0.6397,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.1845472440944882,
|
| 538 |
+
"grad_norm": 2.704435062795677,
|
| 539 |
+
"learning_rate": 9.785636088604571e-06,
|
| 540 |
+
"loss": 0.6401,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.18700787401574803,
|
| 545 |
+
"grad_norm": 2.4110784601265935,
|
| 546 |
+
"learning_rate": 9.773014010110298e-06,
|
| 547 |
+
"loss": 0.6506,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.18946850393700787,
|
| 552 |
+
"grad_norm": 2.3808900476483434,
|
| 553 |
+
"learning_rate": 9.760039497933266e-06,
|
| 554 |
+
"loss": 0.6427,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.1919291338582677,
|
| 559 |
+
"grad_norm": 2.443382230868993,
|
| 560 |
+
"learning_rate": 9.74671351009604e-06,
|
| 561 |
+
"loss": 0.6536,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.19438976377952755,
|
| 566 |
+
"grad_norm": 2.53346201637983,
|
| 567 |
+
"learning_rate": 9.733037030573725e-06,
|
| 568 |
+
"loss": 0.629,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.1968503937007874,
|
| 573 |
+
"grad_norm": 2.7306008040110776,
|
| 574 |
+
"learning_rate": 9.719011069221316e-06,
|
| 575 |
+
"loss": 0.6395,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.19931102362204725,
|
| 580 |
+
"grad_norm": 2.5885107627509316,
|
| 581 |
+
"learning_rate": 9.704636661699133e-06,
|
| 582 |
+
"loss": 0.6142,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.2017716535433071,
|
| 587 |
+
"grad_norm": 2.5758279906754993,
|
| 588 |
+
"learning_rate": 9.68991486939635e-06,
|
| 589 |
+
"loss": 0.6332,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.20423228346456693,
|
| 594 |
+
"grad_norm": 2.527125806186406,
|
| 595 |
+
"learning_rate": 9.674846779352613e-06,
|
| 596 |
+
"loss": 0.6155,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.20669291338582677,
|
| 601 |
+
"grad_norm": 2.352822836622023,
|
| 602 |
+
"learning_rate": 9.659433504177786e-06,
|
| 603 |
+
"loss": 0.6108,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.2091535433070866,
|
| 608 |
+
"grad_norm": 2.428790076631395,
|
| 609 |
+
"learning_rate": 9.643676181969792e-06,
|
| 610 |
+
"loss": 0.6204,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.21161417322834647,
|
| 615 |
+
"grad_norm": 2.5614397854394055,
|
| 616 |
+
"learning_rate": 9.62757597623058e-06,
|
| 617 |
+
"loss": 0.6088,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.2140748031496063,
|
| 622 |
+
"grad_norm": 2.4898996021687227,
|
| 623 |
+
"learning_rate": 9.611134075780209e-06,
|
| 624 |
+
"loss": 0.6211,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.21653543307086615,
|
| 629 |
+
"grad_norm": 2.4088704164520927,
|
| 630 |
+
"learning_rate": 9.59435169466907e-06,
|
| 631 |
+
"loss": 0.6242,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.21899606299212598,
|
| 636 |
+
"grad_norm": 2.5289526680832113,
|
| 637 |
+
"learning_rate": 9.577230072088246e-06,
|
| 638 |
+
"loss": 0.613,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.22145669291338582,
|
| 643 |
+
"grad_norm": 2.40778858369484,
|
| 644 |
+
"learning_rate": 9.559770472277996e-06,
|
| 645 |
+
"loss": 0.6106,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.22391732283464566,
|
| 650 |
+
"grad_norm": 3.00357065325349,
|
| 651 |
+
"learning_rate": 9.541974184434426e-06,
|
| 652 |
+
"loss": 0.6086,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.2263779527559055,
|
| 657 |
+
"grad_norm": 2.420620535949469,
|
| 658 |
+
"learning_rate": 9.523842522614285e-06,
|
| 659 |
+
"loss": 0.604,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.22883858267716536,
|
| 664 |
+
"grad_norm": 2.8983682401439936,
|
| 665 |
+
"learning_rate": 9.505376825637933e-06,
|
| 666 |
+
"loss": 0.5962,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.2312992125984252,
|
| 671 |
+
"grad_norm": 2.6058841172777334,
|
| 672 |
+
"learning_rate": 9.486578456990494e-06,
|
| 673 |
+
"loss": 0.6022,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.23375984251968504,
|
| 678 |
+
"grad_norm": 2.420955967093701,
|
| 679 |
+
"learning_rate": 9.467448804721171e-06,
|
| 680 |
+
"loss": 0.5849,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.23622047244094488,
|
| 685 |
+
"grad_norm": 2.557055641112582,
|
| 686 |
+
"learning_rate": 9.447989281340753e-06,
|
| 687 |
+
"loss": 0.6052,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.23868110236220472,
|
| 692 |
+
"grad_norm": 2.6358527915325625,
|
| 693 |
+
"learning_rate": 9.428201323717327e-06,
|
| 694 |
+
"loss": 0.6066,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.24114173228346455,
|
| 699 |
+
"grad_norm": 2.514164645393038,
|
| 700 |
+
"learning_rate": 9.408086392970167e-06,
|
| 701 |
+
"loss": 0.6057,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.24360236220472442,
|
| 706 |
+
"grad_norm": 2.386864670354864,
|
| 707 |
+
"learning_rate": 9.387645974361858e-06,
|
| 708 |
+
"loss": 0.5921,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.24606299212598426,
|
| 713 |
+
"grad_norm": 2.572659612274251,
|
| 714 |
+
"learning_rate": 9.36688157718862e-06,
|
| 715 |
+
"loss": 0.5766,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.2485236220472441,
|
| 720 |
+
"grad_norm": 2.4584516295055194,
|
| 721 |
+
"learning_rate": 9.345794734668866e-06,
|
| 722 |
+
"loss": 0.5922,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.25098425196850394,
|
| 727 |
+
"grad_norm": 2.27813169588504,
|
| 728 |
+
"learning_rate": 9.324387003829993e-06,
|
| 729 |
+
"loss": 0.6002,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.2534448818897638,
|
| 734 |
+
"grad_norm": 2.513354918132864,
|
| 735 |
+
"learning_rate": 9.302659965393404e-06,
|
| 736 |
+
"loss": 0.5863,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.2559055118110236,
|
| 741 |
+
"grad_norm": 2.57065216369518,
|
| 742 |
+
"learning_rate": 9.280615223657801e-06,
|
| 743 |
+
"loss": 0.5827,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.25836614173228345,
|
| 748 |
+
"grad_norm": 2.508285807949247,
|
| 749 |
+
"learning_rate": 9.258254406380718e-06,
|
| 750 |
+
"loss": 0.6003,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.2608267716535433,
|
| 755 |
+
"grad_norm": 2.393181758359751,
|
| 756 |
+
"learning_rate": 9.23557916465833e-06,
|
| 757 |
+
"loss": 0.5876,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.2632874015748031,
|
| 762 |
+
"grad_norm": 2.45141675908852,
|
| 763 |
+
"learning_rate": 9.212591172803541e-06,
|
| 764 |
+
"loss": 0.5782,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.265748031496063,
|
| 769 |
+
"grad_norm": 2.602808329250647,
|
| 770 |
+
"learning_rate": 9.189292128222355e-06,
|
| 771 |
+
"loss": 0.5817,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.26820866141732286,
|
| 776 |
+
"grad_norm": 2.6264024238291332,
|
| 777 |
+
"learning_rate": 9.165683751288537e-06,
|
| 778 |
+
"loss": 0.5987,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.2706692913385827,
|
| 783 |
+
"grad_norm": 2.5206627459795086,
|
| 784 |
+
"learning_rate": 9.141767785216585e-06,
|
| 785 |
+
"loss": 0.5773,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.27312992125984253,
|
| 790 |
+
"grad_norm": 2.538664530429439,
|
| 791 |
+
"learning_rate": 9.117545995933015e-06,
|
| 792 |
+
"loss": 0.5725,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.2755905511811024,
|
| 797 |
+
"grad_norm": 2.541257403182122,
|
| 798 |
+
"learning_rate": 9.093020171945966e-06,
|
| 799 |
+
"loss": 0.5667,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.2780511811023622,
|
| 804 |
+
"grad_norm": 2.3883496422885937,
|
| 805 |
+
"learning_rate": 9.068192124213135e-06,
|
| 806 |
+
"loss": 0.5681,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.28051181102362205,
|
| 811 |
+
"grad_norm": 2.4768611856821128,
|
| 812 |
+
"learning_rate": 9.043063686008066e-06,
|
| 813 |
+
"loss": 0.5693,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.2829724409448819,
|
| 818 |
+
"grad_norm": 2.531509664449831,
|
| 819 |
+
"learning_rate": 9.017636712784776e-06,
|
| 820 |
+
"loss": 0.5712,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.2854330708661417,
|
| 825 |
+
"grad_norm": 2.4497026467822134,
|
| 826 |
+
"learning_rate": 8.991913082040752e-06,
|
| 827 |
+
"loss": 0.5711,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.28789370078740156,
|
| 832 |
+
"grad_norm": 2.3627002230344027,
|
| 833 |
+
"learning_rate": 8.96589469317832e-06,
|
| 834 |
+
"loss": 0.5746,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.2903543307086614,
|
| 839 |
+
"grad_norm": 2.5848492576021207,
|
| 840 |
+
"learning_rate": 8.9395834673644e-06,
|
| 841 |
+
"loss": 0.5638,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.29281496062992124,
|
| 846 |
+
"grad_norm": 2.7943897726154026,
|
| 847 |
+
"learning_rate": 8.912981347388634e-06,
|
| 848 |
+
"loss": 0.5716,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.2952755905511811,
|
| 853 |
+
"grad_norm": 2.4233634727846547,
|
| 854 |
+
"learning_rate": 8.886090297519956e-06,
|
| 855 |
+
"loss": 0.5589,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.29773622047244097,
|
| 860 |
+
"grad_norm": 2.406617785286132,
|
| 861 |
+
"learning_rate": 8.85891230336153e-06,
|
| 862 |
+
"loss": 0.5743,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.3001968503937008,
|
| 867 |
+
"grad_norm": 2.375668869002511,
|
| 868 |
+
"learning_rate": 8.83144937170415e-06,
|
| 869 |
+
"loss": 0.5536,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.30265748031496065,
|
| 874 |
+
"grad_norm": 2.3455722754718416,
|
| 875 |
+
"learning_rate": 8.803703530378059e-06,
|
| 876 |
+
"loss": 0.5516,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.3051181102362205,
|
| 881 |
+
"grad_norm": 2.5237377829623107,
|
| 882 |
+
"learning_rate": 8.775676828103205e-06,
|
| 883 |
+
"loss": 0.5565,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.3075787401574803,
|
| 888 |
+
"grad_norm": 2.287022202884204,
|
| 889 |
+
"learning_rate": 8.747371334337983e-06,
|
| 890 |
+
"loss": 0.5725,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.31003937007874016,
|
| 895 |
+
"grad_norm": 2.291688895957761,
|
| 896 |
+
"learning_rate": 8.718789139126417e-06,
|
| 897 |
+
"loss": 0.545,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.3125,
|
| 902 |
+
"grad_norm": 2.527855069790082,
|
| 903 |
+
"learning_rate": 8.689932352943837e-06,
|
| 904 |
+
"loss": 0.543,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.31496062992125984,
|
| 909 |
+
"grad_norm": 2.3987364116312535,
|
| 910 |
+
"learning_rate": 8.660803106541044e-06,
|
| 911 |
+
"loss": 0.5475,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.3174212598425197,
|
| 916 |
+
"grad_norm": 2.61276055348719,
|
| 917 |
+
"learning_rate": 8.631403550786979e-06,
|
| 918 |
+
"loss": 0.5488,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.3198818897637795,
|
| 923 |
+
"grad_norm": 2.419050184531339,
|
| 924 |
+
"learning_rate": 8.601735856509903e-06,
|
| 925 |
+
"loss": 0.5476,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.32234251968503935,
|
| 930 |
+
"grad_norm": 2.3768631155511537,
|
| 931 |
+
"learning_rate": 8.571802214337107e-06,
|
| 932 |
+
"loss": 0.5637,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.3248031496062992,
|
| 937 |
+
"grad_norm": 2.4228616659745987,
|
| 938 |
+
"learning_rate": 8.541604834533159e-06,
|
| 939 |
+
"loss": 0.5305,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.32726377952755903,
|
| 944 |
+
"grad_norm": 2.3754663831523875,
|
| 945 |
+
"learning_rate": 8.511145946836704e-06,
|
| 946 |
+
"loss": 0.5331,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.3297244094488189,
|
| 951 |
+
"grad_norm": 2.3398125333991535,
|
| 952 |
+
"learning_rate": 8.48042780029581e-06,
|
| 953 |
+
"loss": 0.5357,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.33218503937007876,
|
| 958 |
+
"grad_norm": 2.4983820414217663,
|
| 959 |
+
"learning_rate": 8.449452663101918e-06,
|
| 960 |
+
"loss": 0.5332,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.3346456692913386,
|
| 965 |
+
"grad_norm": 2.3721219410260397,
|
| 966 |
+
"learning_rate": 8.418222822422348e-06,
|
| 967 |
+
"loss": 0.5417,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.33710629921259844,
|
| 972 |
+
"grad_norm": 2.360196232033206,
|
| 973 |
+
"learning_rate": 8.386740584231431e-06,
|
| 974 |
+
"loss": 0.5461,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.3395669291338583,
|
| 979 |
+
"grad_norm": 2.466049773095662,
|
| 980 |
+
"learning_rate": 8.355008273140222e-06,
|
| 981 |
+
"loss": 0.5253,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.3420275590551181,
|
| 986 |
+
"grad_norm": 2.348269078433154,
|
| 987 |
+
"learning_rate": 8.323028232224863e-06,
|
| 988 |
+
"loss": 0.5304,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.34448818897637795,
|
| 993 |
+
"grad_norm": 2.476352965234454,
|
| 994 |
+
"learning_rate": 8.290802822853576e-06,
|
| 995 |
+
"loss": 0.5349,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.3469488188976378,
|
| 1000 |
+
"grad_norm": 2.279222616967455,
|
| 1001 |
+
"learning_rate": 8.258334424512293e-06,
|
| 1002 |
+
"loss": 0.546,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.3494094488188976,
|
| 1007 |
+
"grad_norm": 2.3637014928251623,
|
| 1008 |
+
"learning_rate": 8.22562543462897e-06,
|
| 1009 |
+
"loss": 0.5371,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.35187007874015747,
|
| 1014 |
+
"grad_norm": 2.619852828253986,
|
| 1015 |
+
"learning_rate": 8.192678268396545e-06,
|
| 1016 |
+
"loss": 0.5237,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.3543307086614173,
|
| 1021 |
+
"grad_norm": 2.5571753512444912,
|
| 1022 |
+
"learning_rate": 8.159495358594627e-06,
|
| 1023 |
+
"loss": 0.5296,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.35679133858267714,
|
| 1028 |
+
"grad_norm": 2.312802552942875,
|
| 1029 |
+
"learning_rate": 8.126079155409845e-06,
|
| 1030 |
+
"loss": 0.5235,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.359251968503937,
|
| 1035 |
+
"grad_norm": 2.5951586860371374,
|
| 1036 |
+
"learning_rate": 8.092432126254933e-06,
|
| 1037 |
+
"loss": 0.5031,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.3617125984251969,
|
| 1042 |
+
"grad_norm": 2.325097293624387,
|
| 1043 |
+
"learning_rate": 8.058556755586537e-06,
|
| 1044 |
+
"loss": 0.5233,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.3641732283464567,
|
| 1049 |
+
"grad_norm": 2.2732745100546725,
|
| 1050 |
+
"learning_rate": 8.024455544721778e-06,
|
| 1051 |
+
"loss": 0.5247,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.36663385826771655,
|
| 1056 |
+
"grad_norm": 2.363542176910271,
|
| 1057 |
+
"learning_rate": 7.990131011653545e-06,
|
| 1058 |
+
"loss": 0.5218,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.3690944881889764,
|
| 1063 |
+
"grad_norm": 2.3003464358723127,
|
| 1064 |
+
"learning_rate": 7.955585690864567e-06,
|
| 1065 |
+
"loss": 0.5116,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.3715551181102362,
|
| 1070 |
+
"grad_norm": 2.3293766926582213,
|
| 1071 |
+
"learning_rate": 7.920822133140285e-06,
|
| 1072 |
+
"loss": 0.5263,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.37401574803149606,
|
| 1077 |
+
"grad_norm": 2.334188982780053,
|
| 1078 |
+
"learning_rate": 7.88584290538049e-06,
|
| 1079 |
+
"loss": 0.4941,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.3764763779527559,
|
| 1084 |
+
"grad_norm": 2.351749135985343,
|
| 1085 |
+
"learning_rate": 7.850650590409795e-06,
|
| 1086 |
+
"loss": 0.5116,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.37893700787401574,
|
| 1091 |
+
"grad_norm": 2.296567696173685,
|
| 1092 |
+
"learning_rate": 7.815247786786919e-06,
|
| 1093 |
+
"loss": 0.5225,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.3813976377952756,
|
| 1098 |
+
"grad_norm": 2.4483931002237975,
|
| 1099 |
+
"learning_rate": 7.779637108612813e-06,
|
| 1100 |
+
"loss": 0.5188,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.3838582677165354,
|
| 1105 |
+
"grad_norm": 2.372343104732536,
|
| 1106 |
+
"learning_rate": 7.743821185337634e-06,
|
| 1107 |
+
"loss": 0.5138,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.38631889763779526,
|
| 1112 |
+
"grad_norm": 2.561786660857743,
|
| 1113 |
+
"learning_rate": 7.7078026615666e-06,
|
| 1114 |
+
"loss": 0.5005,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.3887795275590551,
|
| 1119 |
+
"grad_norm": 2.2678689682364017,
|
| 1120 |
+
"learning_rate": 7.671584196864703e-06,
|
| 1121 |
+
"loss": 0.5144,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.391240157480315,
|
| 1126 |
+
"grad_norm": 2.4118009586518223,
|
| 1127 |
+
"learning_rate": 7.635168465560343e-06,
|
| 1128 |
+
"loss": 0.507,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.3937007874015748,
|
| 1133 |
+
"grad_norm": 2.3699190289261742,
|
| 1134 |
+
"learning_rate": 7.598558156547842e-06,
|
| 1135 |
+
"loss": 0.5058,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.39616141732283466,
|
| 1140 |
+
"grad_norm": 2.3834254512275894,
|
| 1141 |
+
"learning_rate": 7.561755973088917e-06,
|
| 1142 |
+
"loss": 0.5098,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.3986220472440945,
|
| 1147 |
+
"grad_norm": 2.3484210670312304,
|
| 1148 |
+
"learning_rate": 7.52476463261306e-06,
|
| 1149 |
+
"loss": 0.5101,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.40108267716535434,
|
| 1154 |
+
"grad_norm": 2.259743181885874,
|
| 1155 |
+
"learning_rate": 7.487586866516897e-06,
|
| 1156 |
+
"loss": 0.4957,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.4035433070866142,
|
| 1161 |
+
"grad_norm": 2.459196143941421,
|
| 1162 |
+
"learning_rate": 7.450225419962498e-06,
|
| 1163 |
+
"loss": 0.5018,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.406003937007874,
|
| 1168 |
+
"grad_norm": 2.3987573923118326,
|
| 1169 |
+
"learning_rate": 7.412683051674681e-06,
|
| 1170 |
+
"loss": 0.505,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.40846456692913385,
|
| 1175 |
+
"grad_norm": 2.3945193465931527,
|
| 1176 |
+
"learning_rate": 7.374962533737306e-06,
|
| 1177 |
+
"loss": 0.5021,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.4109251968503937,
|
| 1182 |
+
"grad_norm": 2.450543787256226,
|
| 1183 |
+
"learning_rate": 7.3370666513885965e-06,
|
| 1184 |
+
"loss": 0.5045,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.41338582677165353,
|
| 1189 |
+
"grad_norm": 2.3773634574014624,
|
| 1190 |
+
"learning_rate": 7.298998202815474e-06,
|
| 1191 |
+
"loss": 0.4843,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.41584645669291337,
|
| 1196 |
+
"grad_norm": 2.3218222930991312,
|
| 1197 |
+
"learning_rate": 7.260759998946945e-06,
|
| 1198 |
+
"loss": 0.4999,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.4183070866141732,
|
| 1203 |
+
"grad_norm": 2.2812480048868045,
|
| 1204 |
+
"learning_rate": 7.2223548632465424e-06,
|
| 1205 |
+
"loss": 0.4903,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.42076771653543305,
|
| 1210 |
+
"grad_norm": 2.3409310514361246,
|
| 1211 |
+
"learning_rate": 7.183785631503851e-06,
|
| 1212 |
+
"loss": 0.4816,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.42322834645669294,
|
| 1217 |
+
"grad_norm": 2.3088182610498897,
|
| 1218 |
+
"learning_rate": 7.145055151625113e-06,
|
| 1219 |
+
"loss": 0.4754,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.4256889763779528,
|
| 1224 |
+
"grad_norm": 2.3963782413431325,
|
| 1225 |
+
"learning_rate": 7.1061662834229375e-06,
|
| 1226 |
+
"loss": 0.4918,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.4281496062992126,
|
| 1231 |
+
"grad_norm": 2.375549548917402,
|
| 1232 |
+
"learning_rate": 7.0671218984051385e-06,
|
| 1233 |
+
"loss": 0.4797,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.43061023622047245,
|
| 1238 |
+
"grad_norm": 2.340240253487199,
|
| 1239 |
+
"learning_rate": 7.0279248795627156e-06,
|
| 1240 |
+
"loss": 0.4804,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.4330708661417323,
|
| 1245 |
+
"grad_norm": 2.5999252908460564,
|
| 1246 |
+
"learning_rate": 6.988578121156956e-06,
|
| 1247 |
+
"loss": 0.4975,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.43553149606299213,
|
| 1252 |
+
"grad_norm": 2.412860995827959,
|
| 1253 |
+
"learning_rate": 6.9490845285057505e-06,
|
| 1254 |
+
"loss": 0.4757,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.43799212598425197,
|
| 1259 |
+
"grad_norm": 2.3723415230376053,
|
| 1260 |
+
"learning_rate": 6.909447017769047e-06,
|
| 1261 |
+
"loss": 0.485,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.4404527559055118,
|
| 1266 |
+
"grad_norm": 2.2444505424132357,
|
| 1267 |
+
"learning_rate": 6.869668515733536e-06,
|
| 1268 |
+
"loss": 0.4791,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.44291338582677164,
|
| 1273 |
+
"grad_norm": 2.4018284029434627,
|
| 1274 |
+
"learning_rate": 6.829751959596544e-06,
|
| 1275 |
+
"loss": 0.4781,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.4453740157480315,
|
| 1280 |
+
"grad_norm": 2.2719247540147873,
|
| 1281 |
+
"learning_rate": 6.789700296749141e-06,
|
| 1282 |
+
"loss": 0.4664,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.4478346456692913,
|
| 1287 |
+
"grad_norm": 2.5647531601957296,
|
| 1288 |
+
"learning_rate": 6.749516484558518e-06,
|
| 1289 |
+
"loss": 0.4771,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.45029527559055116,
|
| 1294 |
+
"grad_norm": 2.209823332066905,
|
| 1295 |
+
"learning_rate": 6.709203490149615e-06,
|
| 1296 |
+
"loss": 0.4775,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.452755905511811,
|
| 1301 |
+
"grad_norm": 2.3889046698682734,
|
| 1302 |
+
"learning_rate": 6.668764290186039e-06,
|
| 1303 |
+
"loss": 0.4632,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.4552165354330709,
|
| 1308 |
+
"grad_norm": 2.3497617502204142,
|
| 1309 |
+
"learning_rate": 6.628201870650262e-06,
|
| 1310 |
+
"loss": 0.4609,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.45767716535433073,
|
| 1315 |
+
"grad_norm": 2.286479842578479,
|
| 1316 |
+
"learning_rate": 6.587519226623137e-06,
|
| 1317 |
+
"loss": 0.4584,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.46013779527559057,
|
| 1322 |
+
"grad_norm": 2.346679839886123,
|
| 1323 |
+
"learning_rate": 6.546719362062763e-06,
|
| 1324 |
+
"loss": 0.4627,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.4625984251968504,
|
| 1329 |
+
"grad_norm": 2.372517236495747,
|
| 1330 |
+
"learning_rate": 6.50580528958265e-06,
|
| 1331 |
+
"loss": 0.4627,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.46505905511811024,
|
| 1336 |
+
"grad_norm": 2.352894112351414,
|
| 1337 |
+
"learning_rate": 6.464780030229297e-06,
|
| 1338 |
+
"loss": 0.4613,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.4675196850393701,
|
| 1343 |
+
"grad_norm": 2.3155134840690073,
|
| 1344 |
+
"learning_rate": 6.423646613259103e-06,
|
| 1345 |
+
"loss": 0.4643,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.4699803149606299,
|
| 1350 |
+
"grad_norm": 2.2802677404420604,
|
| 1351 |
+
"learning_rate": 6.382408075914698e-06,
|
| 1352 |
+
"loss": 0.4675,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.47244094488188976,
|
| 1357 |
+
"grad_norm": 2.3261149198375577,
|
| 1358 |
+
"learning_rate": 6.341067463200678e-06,
|
| 1359 |
+
"loss": 0.4661,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.4749015748031496,
|
| 1364 |
+
"grad_norm": 2.3024057597107723,
|
| 1365 |
+
"learning_rate": 6.299627827658757e-06,
|
| 1366 |
+
"loss": 0.4695,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.47736220472440943,
|
| 1371 |
+
"grad_norm": 2.3579401662222805,
|
| 1372 |
+
"learning_rate": 6.258092229142383e-06,
|
| 1373 |
+
"loss": 0.4557,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.47982283464566927,
|
| 1378 |
+
"grad_norm": 2.4039007796045624,
|
| 1379 |
+
"learning_rate": 6.216463734590797e-06,
|
| 1380 |
+
"loss": 0.4733,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.4822834645669291,
|
| 1385 |
+
"grad_norm": 2.350383170709395,
|
| 1386 |
+
"learning_rate": 6.174745417802563e-06,
|
| 1387 |
+
"loss": 0.4542,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.484744094488189,
|
| 1392 |
+
"grad_norm": 2.3727437081826195,
|
| 1393 |
+
"learning_rate": 6.132940359208625e-06,
|
| 1394 |
+
"loss": 0.4599,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.48720472440944884,
|
| 1399 |
+
"grad_norm": 2.2638903752102877,
|
| 1400 |
+
"learning_rate": 6.09105164564483e-06,
|
| 1401 |
+
"loss": 0.4581,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.4896653543307087,
|
| 1406 |
+
"grad_norm": 2.391666707671801,
|
| 1407 |
+
"learning_rate": 6.049082370124011e-06,
|
| 1408 |
+
"loss": 0.4335,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.4921259842519685,
|
| 1413 |
+
"grad_norm": 2.285287290642185,
|
| 1414 |
+
"learning_rate": 6.007035631607605e-06,
|
| 1415 |
+
"loss": 0.4322,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.49458661417322836,
|
| 1420 |
+
"grad_norm": 2.3527234691645886,
|
| 1421 |
+
"learning_rate": 5.964914534776814e-06,
|
| 1422 |
+
"loss": 0.4452,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.4970472440944882,
|
| 1427 |
+
"grad_norm": 2.388047521619243,
|
| 1428 |
+
"learning_rate": 5.9227221898033785e-06,
|
| 1429 |
+
"loss": 0.4452,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.49950787401574803,
|
| 1434 |
+
"grad_norm": 2.330957204189204,
|
| 1435 |
+
"learning_rate": 5.880461712119913e-06,
|
| 1436 |
+
"loss": 0.4509,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.5019685039370079,
|
| 1441 |
+
"grad_norm": 2.3767662326352146,
|
| 1442 |
+
"learning_rate": 5.838136222189874e-06,
|
| 1443 |
+
"loss": 0.4417,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.5044291338582677,
|
| 1448 |
+
"grad_norm": 2.3296456305613553,
|
| 1449 |
+
"learning_rate": 5.795748845277143e-06,
|
| 1450 |
+
"loss": 0.4456,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.5068897637795275,
|
| 1455 |
+
"grad_norm": 2.3238395450456317,
|
| 1456 |
+
"learning_rate": 5.75330271121526e-06,
|
| 1457 |
+
"loss": 0.4483,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.5093503937007874,
|
| 1462 |
+
"grad_norm": 2.271731748541768,
|
| 1463 |
+
"learning_rate": 5.710800954176326e-06,
|
| 1464 |
+
"loss": 0.4439,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.5118110236220472,
|
| 1469 |
+
"grad_norm": 2.2730785805154716,
|
| 1470 |
+
"learning_rate": 5.668246712439579e-06,
|
| 1471 |
+
"loss": 0.438,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.5142716535433071,
|
| 1476 |
+
"grad_norm": 2.2984158447258856,
|
| 1477 |
+
"learning_rate": 5.625643128159658e-06,
|
| 1478 |
+
"loss": 0.4371,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.5167322834645669,
|
| 1483 |
+
"grad_norm": 2.138298543480362,
|
| 1484 |
+
"learning_rate": 5.582993347134604e-06,
|
| 1485 |
+
"loss": 0.4358,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.5191929133858267,
|
| 1490 |
+
"grad_norm": 2.2656933007145326,
|
| 1491 |
+
"learning_rate": 5.540300518573564e-06,
|
| 1492 |
+
"loss": 0.4456,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.5216535433070866,
|
| 1497 |
+
"grad_norm": 2.4687742042434437,
|
| 1498 |
+
"learning_rate": 5.4975677948642704e-06,
|
| 1499 |
+
"loss": 0.4292,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.5241141732283464,
|
| 1504 |
+
"grad_norm": 2.2023440824897538,
|
| 1505 |
+
"learning_rate": 5.454798331340261e-06,
|
| 1506 |
+
"loss": 0.4448,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.5265748031496063,
|
| 1511 |
+
"grad_norm": 2.3820756070238844,
|
| 1512 |
+
"learning_rate": 5.4119952860479e-06,
|
| 1513 |
+
"loss": 0.4322,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.5290354330708661,
|
| 1518 |
+
"grad_norm": 2.307966181110738,
|
| 1519 |
+
"learning_rate": 5.369161819513189e-06,
|
| 1520 |
+
"loss": 0.4345,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.531496062992126,
|
| 1525 |
+
"grad_norm": 2.4566374524002006,
|
| 1526 |
+
"learning_rate": 5.3263010945083994e-06,
|
| 1527 |
+
"loss": 0.4342,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.5339566929133859,
|
| 1532 |
+
"grad_norm": 2.304141656628352,
|
| 1533 |
+
"learning_rate": 5.283416275818531e-06,
|
| 1534 |
+
"loss": 0.4304,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.5364173228346457,
|
| 1539 |
+
"grad_norm": 2.3743587538000974,
|
| 1540 |
+
"learning_rate": 5.240510530007641e-06,
|
| 1541 |
+
"loss": 0.4248,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.5388779527559056,
|
| 1546 |
+
"grad_norm": 2.275764809511964,
|
| 1547 |
+
"learning_rate": 5.1975870251850105e-06,
|
| 1548 |
+
"loss": 0.4294,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.5413385826771654,
|
| 1553 |
+
"grad_norm": 2.4755101283685823,
|
| 1554 |
+
"learning_rate": 5.1546489307712345e-06,
|
| 1555 |
+
"loss": 0.4248,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.5437992125984252,
|
| 1560 |
+
"grad_norm": 2.1985695933129406,
|
| 1561 |
+
"learning_rate": 5.111699417264177e-06,
|
| 1562 |
+
"loss": 0.423,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.5462598425196851,
|
| 1567 |
+
"grad_norm": 2.2532688868144537,
|
| 1568 |
+
"learning_rate": 5.06874165600488e-06,
|
| 1569 |
+
"loss": 0.4247,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.5487204724409449,
|
| 1574 |
+
"grad_norm": 2.5064255810867553,
|
| 1575 |
+
"learning_rate": 5.025778818943391e-06,
|
| 1576 |
+
"loss": 0.4171,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.5511811023622047,
|
| 1581 |
+
"grad_norm": 2.223795201431944,
|
| 1582 |
+
"learning_rate": 4.982814078404543e-06,
|
| 1583 |
+
"loss": 0.4136,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.5536417322834646,
|
| 1588 |
+
"grad_norm": 2.321073710905454,
|
| 1589 |
+
"learning_rate": 4.939850606853724e-06,
|
| 1590 |
+
"loss": 0.4128,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.5561023622047244,
|
| 1595 |
+
"grad_norm": 2.352196407309558,
|
| 1596 |
+
"learning_rate": 4.89689157666262e-06,
|
| 1597 |
+
"loss": 0.4225,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.5585629921259843,
|
| 1602 |
+
"grad_norm": 2.3165609074323386,
|
| 1603 |
+
"learning_rate": 4.853940159874972e-06,
|
| 1604 |
+
"loss": 0.4137,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.5610236220472441,
|
| 1609 |
+
"grad_norm": 2.249863447456596,
|
| 1610 |
+
"learning_rate": 4.8109995279723556e-06,
|
| 1611 |
+
"loss": 0.4124,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.5634842519685039,
|
| 1616 |
+
"grad_norm": 2.3324932344927927,
|
| 1617 |
+
"learning_rate": 4.768072851640006e-06,
|
| 1618 |
+
"loss": 0.4092,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.5659448818897638,
|
| 1623 |
+
"grad_norm": 2.175622171626361,
|
| 1624 |
+
"learning_rate": 4.7251633005326935e-06,
|
| 1625 |
+
"loss": 0.4044,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.5684055118110236,
|
| 1630 |
+
"grad_norm": 2.244989030279403,
|
| 1631 |
+
"learning_rate": 4.682274043040682e-06,
|
| 1632 |
+
"loss": 0.4102,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.5708661417322834,
|
| 1637 |
+
"grad_norm": 2.3655854821968583,
|
| 1638 |
+
"learning_rate": 4.639408246055781e-06,
|
| 1639 |
+
"loss": 0.3998,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.5733267716535433,
|
| 1644 |
+
"grad_norm": 2.3662328414472147,
|
| 1645 |
+
"learning_rate": 4.596569074737501e-06,
|
| 1646 |
+
"loss": 0.4209,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.5757874015748031,
|
| 1651 |
+
"grad_norm": 2.2898685582377114,
|
| 1652 |
+
"learning_rate": 4.5537596922793475e-06,
|
| 1653 |
+
"loss": 0.3884,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.578248031496063,
|
| 1658 |
+
"grad_norm": 2.175371768003494,
|
| 1659 |
+
"learning_rate": 4.510983259675252e-06,
|
| 1660 |
+
"loss": 0.413,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.5807086614173228,
|
| 1665 |
+
"grad_norm": 2.174940506600783,
|
| 1666 |
+
"learning_rate": 4.468242935486164e-06,
|
| 1667 |
+
"loss": 0.3977,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.5831692913385826,
|
| 1672 |
+
"grad_norm": 2.2645627809171205,
|
| 1673 |
+
"learning_rate": 4.425541875606837e-06,
|
| 1674 |
+
"loss": 0.4,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.5856299212598425,
|
| 1679 |
+
"grad_norm": 2.3434052194119097,
|
| 1680 |
+
"learning_rate": 4.38288323303279e-06,
|
| 1681 |
+
"loss": 0.416,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.5880905511811023,
|
| 1686 |
+
"grad_norm": 2.1589090436361027,
|
| 1687 |
+
"learning_rate": 4.340270157627496e-06,
|
| 1688 |
+
"loss": 0.3975,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.5905511811023622,
|
| 1693 |
+
"grad_norm": 2.2770509922795226,
|
| 1694 |
+
"learning_rate": 4.29770579588981e-06,
|
| 1695 |
+
"loss": 0.4133,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.593011811023622,
|
| 1700 |
+
"grad_norm": 2.133861488496685,
|
| 1701 |
+
"learning_rate": 4.255193290721626e-06,
|
| 1702 |
+
"loss": 0.3965,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.5954724409448819,
|
| 1707 |
+
"grad_norm": 2.0592654419552474,
|
| 1708 |
+
"learning_rate": 4.2127357811958006e-06,
|
| 1709 |
+
"loss": 0.3944,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5979330708661418,
|
| 1714 |
+
"grad_norm": 2.3061861279285756,
|
| 1715 |
+
"learning_rate": 4.170336402324393e-06,
|
| 1716 |
+
"loss": 0.3989,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.6003937007874016,
|
| 1721 |
+
"grad_norm": 2.1897218075138665,
|
| 1722 |
+
"learning_rate": 4.127998284827148e-06,
|
| 1723 |
+
"loss": 0.3922,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.6028543307086615,
|
| 1728 |
+
"grad_norm": 2.273569046487088,
|
| 1729 |
+
"learning_rate": 4.085724554900359e-06,
|
| 1730 |
+
"loss": 0.3919,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.6053149606299213,
|
| 1735 |
+
"grad_norm": 2.316647904273548,
|
| 1736 |
+
"learning_rate": 4.043518333986009e-06,
|
| 1737 |
+
"loss": 0.4004,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.6077755905511811,
|
| 1742 |
+
"grad_norm": 2.139410821479213,
|
| 1743 |
+
"learning_rate": 4.001382738541291e-06,
|
| 1744 |
+
"loss": 0.3868,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.610236220472441,
|
| 1749 |
+
"grad_norm": 2.4158380958610266,
|
| 1750 |
+
"learning_rate": 3.9593208798085094e-06,
|
| 1751 |
+
"loss": 0.3968,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.6126968503937008,
|
| 1756 |
+
"grad_norm": 2.2664180097106934,
|
| 1757 |
+
"learning_rate": 3.9173358635853285e-06,
|
| 1758 |
+
"loss": 0.3977,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.6151574803149606,
|
| 1763 |
+
"grad_norm": 2.1617748237173995,
|
| 1764 |
+
"learning_rate": 3.875430789995454e-06,
|
| 1765 |
+
"loss": 0.3951,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.6176181102362205,
|
| 1770 |
+
"grad_norm": 2.3139618808394258,
|
| 1771 |
+
"learning_rate": 3.833608753259729e-06,
|
| 1772 |
+
"loss": 0.3883,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.6200787401574803,
|
| 1777 |
+
"grad_norm": 2.2557729958410646,
|
| 1778 |
+
"learning_rate": 3.791872841467643e-06,
|
| 1779 |
+
"loss": 0.3833,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.6225393700787402,
|
| 1784 |
+
"grad_norm": 2.288978630108445,
|
| 1785 |
+
"learning_rate": 3.7502261363493346e-06,
|
| 1786 |
+
"loss": 0.392,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.625,
|
| 1791 |
+
"grad_norm": 2.400211890621822,
|
| 1792 |
+
"learning_rate": 3.708671713048017e-06,
|
| 1793 |
+
"loss": 0.382,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.6274606299212598,
|
| 1798 |
+
"grad_norm": 2.14291748228325,
|
| 1799 |
+
"learning_rate": 3.6672126398929273e-06,
|
| 1800 |
+
"loss": 0.3868,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.6299212598425197,
|
| 1805 |
+
"grad_norm": 2.189481811128098,
|
| 1806 |
+
"learning_rate": 3.625851978172765e-06,
|
| 1807 |
+
"loss": 0.3844,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.6323818897637795,
|
| 1812 |
+
"grad_norm": 2.17033479220634,
|
| 1813 |
+
"learning_rate": 3.5845927819096405e-06,
|
| 1814 |
+
"loss": 0.3767,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.6348425196850394,
|
| 1819 |
+
"grad_norm": 2.1024337319523285,
|
| 1820 |
+
"learning_rate": 3.543438097633577e-06,
|
| 1821 |
+
"loss": 0.385,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.6373031496062992,
|
| 1826 |
+
"grad_norm": 2.1601959551273584,
|
| 1827 |
+
"learning_rate": 3.5023909641575604e-06,
|
| 1828 |
+
"loss": 0.3655,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.639763779527559,
|
| 1833 |
+
"grad_norm": 2.2147245957921324,
|
| 1834 |
+
"learning_rate": 3.4614544123531476e-06,
|
| 1835 |
+
"loss": 0.3846,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.6422244094488189,
|
| 1840 |
+
"grad_norm": 2.2767016812027037,
|
| 1841 |
+
"learning_rate": 3.4206314649266813e-06,
|
| 1842 |
+
"loss": 0.3829,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.6446850393700787,
|
| 1847 |
+
"grad_norm": 2.189301400471183,
|
| 1848 |
+
"learning_rate": 3.3799251361960883e-06,
|
| 1849 |
+
"loss": 0.3793,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.6471456692913385,
|
| 1854 |
+
"grad_norm": 2.2486478577990328,
|
| 1855 |
+
"learning_rate": 3.339338431868311e-06,
|
| 1856 |
+
"loss": 0.3651,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.6496062992125984,
|
| 1861 |
+
"grad_norm": 2.1239701183150186,
|
| 1862 |
+
"learning_rate": 3.29887434881737e-06,
|
| 1863 |
+
"loss": 0.3962,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.6520669291338582,
|
| 1868 |
+
"grad_norm": 2.0943809822170194,
|
| 1869 |
+
"learning_rate": 3.2585358748630725e-06,
|
| 1870 |
+
"loss": 0.3781,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.6545275590551181,
|
| 1875 |
+
"grad_norm": 2.302998816495133,
|
| 1876 |
+
"learning_rate": 3.2183259885504003e-06,
|
| 1877 |
+
"loss": 0.3599,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.656988188976378,
|
| 1882 |
+
"grad_norm": 2.093137463735972,
|
| 1883 |
+
"learning_rate": 3.1782476589295803e-06,
|
| 1884 |
+
"loss": 0.3693,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.6594488188976378,
|
| 1889 |
+
"grad_norm": 2.0537589165415038,
|
| 1890 |
+
"learning_rate": 3.138303845336844e-06,
|
| 1891 |
+
"loss": 0.3767,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.6619094488188977,
|
| 1896 |
+
"grad_norm": 2.1644873169815018,
|
| 1897 |
+
"learning_rate": 3.098497497175925e-06,
|
| 1898 |
+
"loss": 0.3758,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.6643700787401575,
|
| 1903 |
+
"grad_norm": 2.1158460604452904,
|
| 1904 |
+
"learning_rate": 3.0588315537002682e-06,
|
| 1905 |
+
"loss": 0.3715,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.6668307086614174,
|
| 1910 |
+
"grad_norm": 2.312489752052895,
|
| 1911 |
+
"learning_rate": 3.0193089437960043e-06,
|
| 1912 |
+
"loss": 0.3764,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.6692913385826772,
|
| 1917 |
+
"grad_norm": 2.4373363981556686,
|
| 1918 |
+
"learning_rate": 2.9799325857656856e-06,
|
| 1919 |
+
"loss": 0.3639,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.671751968503937,
|
| 1924 |
+
"grad_norm": 2.051650821654651,
|
| 1925 |
+
"learning_rate": 2.940705387112798e-06,
|
| 1926 |
+
"loss": 0.3456,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.6742125984251969,
|
| 1931 |
+
"grad_norm": 2.162923091354617,
|
| 1932 |
+
"learning_rate": 2.901630244327075e-06,
|
| 1933 |
+
"loss": 0.3737,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.6766732283464567,
|
| 1938 |
+
"grad_norm": 2.1826289176609617,
|
| 1939 |
+
"learning_rate": 2.862710042670629e-06,
|
| 1940 |
+
"loss": 0.3777,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.6791338582677166,
|
| 1945 |
+
"grad_norm": 2.271492512262752,
|
| 1946 |
+
"learning_rate": 2.8239476559649013e-06,
|
| 1947 |
+
"loss": 0.3672,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.6815944881889764,
|
| 1952 |
+
"grad_norm": 2.0841809045493442,
|
| 1953 |
+
"learning_rate": 2.7853459463784643e-06,
|
| 1954 |
+
"loss": 0.3658,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.6840551181102362,
|
| 1959 |
+
"grad_norm": 2.220969106748428,
|
| 1960 |
+
"learning_rate": 2.7469077642156844e-06,
|
| 1961 |
+
"loss": 0.3602,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.6865157480314961,
|
| 1966 |
+
"grad_norm": 2.009625282982446,
|
| 1967 |
+
"learning_rate": 2.7086359477062542e-06,
|
| 1968 |
+
"loss": 0.353,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.6889763779527559,
|
| 1973 |
+
"grad_norm": 2.267920231263409,
|
| 1974 |
+
"learning_rate": 2.6705333227956304e-06,
|
| 1975 |
+
"loss": 0.3672,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.6914370078740157,
|
| 1980 |
+
"grad_norm": 2.099753211120495,
|
| 1981 |
+
"learning_rate": 2.6326027029363575e-06,
|
| 1982 |
+
"loss": 0.3657,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.6938976377952756,
|
| 1987 |
+
"grad_norm": 2.3887622578641046,
|
| 1988 |
+
"learning_rate": 2.5948468888803323e-06,
|
| 1989 |
+
"loss": 0.3641,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.6963582677165354,
|
| 1994 |
+
"grad_norm": 2.2574854591596063,
|
| 1995 |
+
"learning_rate": 2.557268668472002e-06,
|
| 1996 |
+
"loss": 0.3681,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.6988188976377953,
|
| 2001 |
+
"grad_norm": 2.1541998550392285,
|
| 2002 |
+
"learning_rate": 2.5198708164425046e-06,
|
| 2003 |
+
"loss": 0.3666,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.7012795275590551,
|
| 2008 |
+
"grad_norm": 2.2114896772148924,
|
| 2009 |
+
"learning_rate": 2.482656094204794e-06,
|
| 2010 |
+
"loss": 0.3721,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.7037401574803149,
|
| 2015 |
+
"grad_norm": 1.999095279733753,
|
| 2016 |
+
"learning_rate": 2.445627249649742e-06,
|
| 2017 |
+
"loss": 0.3533,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.7062007874015748,
|
| 2022 |
+
"grad_norm": 2.447157001336924,
|
| 2023 |
+
"learning_rate": 2.4087870169432263e-06,
|
| 2024 |
+
"loss": 0.3512,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.7086614173228346,
|
| 2029 |
+
"grad_norm": 2.1383643267190875,
|
| 2030 |
+
"learning_rate": 2.372138116324254e-06,
|
| 2031 |
+
"loss": 0.3637,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.7111220472440944,
|
| 2036 |
+
"grad_norm": 2.2411777344710933,
|
| 2037 |
+
"learning_rate": 2.3356832539040976e-06,
|
| 2038 |
+
"loss": 0.3521,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.7135826771653543,
|
| 2043 |
+
"grad_norm": 2.081326294904533,
|
| 2044 |
+
"learning_rate": 2.2994251214664754e-06,
|
| 2045 |
+
"loss": 0.3636,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.7160433070866141,
|
| 2050 |
+
"grad_norm": 2.0265851050443717,
|
| 2051 |
+
"learning_rate": 2.263366396268806e-06,
|
| 2052 |
+
"loss": 0.3631,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.718503937007874,
|
| 2057 |
+
"grad_norm": 2.016445263024121,
|
| 2058 |
+
"learning_rate": 2.227509740844508e-06,
|
| 2059 |
+
"loss": 0.3506,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.7209645669291339,
|
| 2064 |
+
"grad_norm": 1.9859142840819781,
|
| 2065 |
+
"learning_rate": 2.191857802806409e-06,
|
| 2066 |
+
"loss": 0.3492,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.7234251968503937,
|
| 2071 |
+
"grad_norm": 2.131894669259727,
|
| 2072 |
+
"learning_rate": 2.1564132146512494e-06,
|
| 2073 |
+
"loss": 0.361,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.7258858267716536,
|
| 2078 |
+
"grad_norm": 2.054521134481261,
|
| 2079 |
+
"learning_rate": 2.1211785935652974e-06,
|
| 2080 |
+
"loss": 0.353,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.7283464566929134,
|
| 2085 |
+
"grad_norm": 2.195553612057405,
|
| 2086 |
+
"learning_rate": 2.086156541231109e-06,
|
| 2087 |
+
"loss": 0.3541,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.7308070866141733,
|
| 2092 |
+
"grad_norm": 2.078190944421504,
|
| 2093 |
+
"learning_rate": 2.05134964363541e-06,
|
| 2094 |
+
"loss": 0.3532,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.7332677165354331,
|
| 2099 |
+
"grad_norm": 1.9784288837683923,
|
| 2100 |
+
"learning_rate": 2.016760470878158e-06,
|
| 2101 |
+
"loss": 0.3537,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.7357283464566929,
|
| 2106 |
+
"grad_norm": 2.161906206866742,
|
| 2107 |
+
"learning_rate": 1.9823915769827672e-06,
|
| 2108 |
+
"loss": 0.3635,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.7381889763779528,
|
| 2113 |
+
"grad_norm": 2.0523238713671135,
|
| 2114 |
+
"learning_rate": 1.948245499707523e-06,
|
| 2115 |
+
"loss": 0.3486,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.7406496062992126,
|
| 2120 |
+
"grad_norm": 2.08663663246033,
|
| 2121 |
+
"learning_rate": 1.9143247603581925e-06,
|
| 2122 |
+
"loss": 0.3517,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.7431102362204725,
|
| 2127 |
+
"grad_norm": 2.247439527318004,
|
| 2128 |
+
"learning_rate": 1.8806318636018666e-06,
|
| 2129 |
+
"loss": 0.3485,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.7455708661417323,
|
| 2134 |
+
"grad_norm": 2.2948321577508355,
|
| 2135 |
+
"learning_rate": 1.8471692972820027e-06,
|
| 2136 |
+
"loss": 0.3554,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.7480314960629921,
|
| 2141 |
+
"grad_norm": 2.1604348165019065,
|
| 2142 |
+
"learning_rate": 1.8139395322347335e-06,
|
| 2143 |
+
"loss": 0.3489,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.750492125984252,
|
| 2148 |
+
"grad_norm": 2.125600663304225,
|
| 2149 |
+
"learning_rate": 1.780945022106424e-06,
|
| 2150 |
+
"loss": 0.3589,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.7529527559055118,
|
| 2155 |
+
"grad_norm": 1.9910694715906239,
|
| 2156 |
+
"learning_rate": 1.7481882031724929e-06,
|
| 2157 |
+
"loss": 0.3483,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.7554133858267716,
|
| 2162 |
+
"grad_norm": 2.024386438154311,
|
| 2163 |
+
"learning_rate": 1.7156714941575292e-06,
|
| 2164 |
+
"loss": 0.3582,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.7578740157480315,
|
| 2169 |
+
"grad_norm": 2.236178264159934,
|
| 2170 |
+
"learning_rate": 1.6833972960566868e-06,
|
| 2171 |
+
"loss": 0.3535,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.7603346456692913,
|
| 2176 |
+
"grad_norm": 2.0509148837648876,
|
| 2177 |
+
"learning_rate": 1.6513679919583975e-06,
|
| 2178 |
+
"loss": 0.3485,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.7627952755905512,
|
| 2183 |
+
"grad_norm": 1.9020807079292121,
|
| 2184 |
+
"learning_rate": 1.6195859468684199e-06,
|
| 2185 |
+
"loss": 0.3496,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.765255905511811,
|
| 2190 |
+
"grad_norm": 2.1302050953871694,
|
| 2191 |
+
"learning_rate": 1.588053507535195e-06,
|
| 2192 |
+
"loss": 0.3385,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.7677165354330708,
|
| 2197 |
+
"grad_norm": 2.184409457671906,
|
| 2198 |
+
"learning_rate": 1.5567730022765753e-06,
|
| 2199 |
+
"loss": 0.342,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.7701771653543307,
|
| 2204 |
+
"grad_norm": 2.1109842289764735,
|
| 2205 |
+
"learning_rate": 1.5257467408078996e-06,
|
| 2206 |
+
"loss": 0.3461,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.7726377952755905,
|
| 2211 |
+
"grad_norm": 1.9447642578789646,
|
| 2212 |
+
"learning_rate": 1.494977014071441e-06,
|
| 2213 |
+
"loss": 0.3321,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.7750984251968503,
|
| 2218 |
+
"grad_norm": 2.160598206464441,
|
| 2219 |
+
"learning_rate": 1.4644660940672628e-06,
|
| 2220 |
+
"loss": 0.3463,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.7775590551181102,
|
| 2225 |
+
"grad_norm": 1.9954045529577038,
|
| 2226 |
+
"learning_rate": 1.434216233685441e-06,
|
| 2227 |
+
"loss": 0.354,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.78001968503937,
|
| 2232 |
+
"grad_norm": 1.8705318043615158,
|
| 2233 |
+
"learning_rate": 1.4042296665397187e-06,
|
| 2234 |
+
"loss": 0.3354,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.78248031496063,
|
| 2239 |
+
"grad_norm": 2.131526940157942,
|
| 2240 |
+
"learning_rate": 1.374508606802586e-06,
|
| 2241 |
+
"loss": 0.326,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.7849409448818898,
|
| 2246 |
+
"grad_norm": 1.9854582148960758,
|
| 2247 |
+
"learning_rate": 1.3450552490417712e-06,
|
| 2248 |
+
"loss": 0.3286,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.7874015748031497,
|
| 2253 |
+
"grad_norm": 2.005367208929083,
|
| 2254 |
+
"learning_rate": 1.3158717680582128e-06,
|
| 2255 |
+
"loss": 0.3422,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.7898622047244095,
|
| 2260 |
+
"grad_norm": 2.0697719484600614,
|
| 2261 |
+
"learning_rate": 1.286960318725471e-06,
|
| 2262 |
+
"loss": 0.3405,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.7923228346456693,
|
| 2267 |
+
"grad_norm": 2.1676514262257656,
|
| 2268 |
+
"learning_rate": 1.2583230358306053e-06,
|
| 2269 |
+
"loss": 0.3519,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.7947834645669292,
|
| 2274 |
+
"grad_norm": 2.030399197550805,
|
| 2275 |
+
"learning_rate": 1.2299620339165607e-06,
|
| 2276 |
+
"loss": 0.3371,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.797244094488189,
|
| 2281 |
+
"grad_norm": 1.8800725084378718,
|
| 2282 |
+
"learning_rate": 1.201879407126012e-06,
|
| 2283 |
+
"loss": 0.3387,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.7997047244094488,
|
| 2288 |
+
"grad_norm": 2.1095666405799345,
|
| 2289 |
+
"learning_rate": 1.1740772290467518e-06,
|
| 2290 |
+
"loss": 0.343,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.8021653543307087,
|
| 2295 |
+
"grad_norm": 2.036216759579623,
|
| 2296 |
+
"learning_rate": 1.1465575525585743e-06,
|
| 2297 |
+
"loss": 0.3254,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.8046259842519685,
|
| 2302 |
+
"grad_norm": 1.9725419805321742,
|
| 2303 |
+
"learning_rate": 1.119322409681689e-06,
|
| 2304 |
+
"loss": 0.3377,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.8070866141732284,
|
| 2309 |
+
"grad_norm": 2.261947410750145,
|
| 2310 |
+
"learning_rate": 1.0923738114266824e-06,
|
| 2311 |
+
"loss": 0.3414,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.8095472440944882,
|
| 2316 |
+
"grad_norm": 1.9967677016675829,
|
| 2317 |
+
"learning_rate": 1.0657137476460272e-06,
|
| 2318 |
+
"loss": 0.3427,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.812007874015748,
|
| 2323 |
+
"grad_norm": 1.9329332696972708,
|
| 2324 |
+
"learning_rate": 1.0393441868871507e-06,
|
| 2325 |
+
"loss": 0.3309,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.8144685039370079,
|
| 2330 |
+
"grad_norm": 2.080216309757756,
|
| 2331 |
+
"learning_rate": 1.0132670762470875e-06,
|
| 2332 |
+
"loss": 0.3403,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.8169291338582677,
|
| 2337 |
+
"grad_norm": 1.9963077191898775,
|
| 2338 |
+
"learning_rate": 9.874843412286994e-07,
|
| 2339 |
+
"loss": 0.3395,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.8193897637795275,
|
| 2344 |
+
"grad_norm": 2.1468283823628584,
|
| 2345 |
+
"learning_rate": 9.619978855985017e-07,
|
| 2346 |
+
"loss": 0.3365,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.8218503937007874,
|
| 2351 |
+
"grad_norm": 2.030201170967621,
|
| 2352 |
+
"learning_rate": 9.368095912460934e-07,
|
| 2353 |
+
"loss": 0.3343,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.8243110236220472,
|
| 2358 |
+
"grad_norm": 2.144207392793775,
|
| 2359 |
+
"learning_rate": 9.119213180451974e-07,
|
| 2360 |
+
"loss": 0.3365,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.8267716535433071,
|
| 2365 |
+
"grad_norm": 2.139359174863488,
|
| 2366 |
+
"learning_rate": 8.87334903716332e-07,
|
| 2367 |
+
"loss": 0.3344,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.8292322834645669,
|
| 2372 |
+
"grad_norm": 2.0756053748869028,
|
| 2373 |
+
"learning_rate": 8.630521636911171e-07,
|
| 2374 |
+
"loss": 0.3396,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.8316929133858267,
|
| 2379 |
+
"grad_norm": 1.9417094509672648,
|
| 2380 |
+
"learning_rate": 8.390748909782204e-07,
|
| 2381 |
+
"loss": 0.3319,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.8341535433070866,
|
| 2386 |
+
"grad_norm": 1.904699920697159,
|
| 2387 |
+
"learning_rate": 8.154048560309669e-07,
|
| 2388 |
+
"loss": 0.3383,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.8366141732283464,
|
| 2393 |
+
"grad_norm": 1.9020653291530694,
|
| 2394 |
+
"learning_rate": 7.920438066166097e-07,
|
| 2395 |
+
"loss": 0.3349,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.8390748031496063,
|
| 2400 |
+
"grad_norm": 2.00777989664957,
|
| 2401 |
+
"learning_rate": 7.689934676872768e-07,
|
| 2402 |
+
"loss": 0.337,
|
| 2403 |
+
"step": 1705
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.8415354330708661,
|
| 2407 |
+
"grad_norm": 2.1247990851762677,
|
| 2408 |
+
"learning_rate": 7.462555412526062e-07,
|
| 2409 |
+
"loss": 0.3299,
|
| 2410 |
+
"step": 1710
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.843996062992126,
|
| 2414 |
+
"grad_norm": 2.0897419693416004,
|
| 2415 |
+
"learning_rate": 7.238317062540661e-07,
|
| 2416 |
+
"loss": 0.3261,
|
| 2417 |
+
"step": 1715
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.8464566929133859,
|
| 2421 |
+
"grad_norm": 2.243164262785359,
|
| 2422 |
+
"learning_rate": 7.017236184409859e-07,
|
| 2423 |
+
"loss": 0.3378,
|
| 2424 |
+
"step": 1720
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.8489173228346457,
|
| 2428 |
+
"grad_norm": 1.9439608610518924,
|
| 2429 |
+
"learning_rate": 6.799329102482988e-07,
|
| 2430 |
+
"loss": 0.3206,
|
| 2431 |
+
"step": 1725
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.8513779527559056,
|
| 2435 |
+
"grad_norm": 2.001456213276007,
|
| 2436 |
+
"learning_rate": 6.584611906760036e-07,
|
| 2437 |
+
"loss": 0.3224,
|
| 2438 |
+
"step": 1730
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.8538385826771654,
|
| 2442 |
+
"grad_norm": 1.9357999946640365,
|
| 2443 |
+
"learning_rate": 6.373100451703601e-07,
|
| 2444 |
+
"loss": 0.3514,
|
| 2445 |
+
"step": 1735
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.8562992125984252,
|
| 2449 |
+
"grad_norm": 1.8706183982176696,
|
| 2450 |
+
"learning_rate": 6.164810355068179e-07,
|
| 2451 |
+
"loss": 0.3355,
|
| 2452 |
+
"step": 1740
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.8587598425196851,
|
| 2456 |
+
"grad_norm": 1.9372386803072519,
|
| 2457 |
+
"learning_rate": 5.959756996746996e-07,
|
| 2458 |
+
"loss": 0.3383,
|
| 2459 |
+
"step": 1745
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.8612204724409449,
|
| 2463 |
+
"grad_norm": 1.9978519375008006,
|
| 2464 |
+
"learning_rate": 5.757955517636365e-07,
|
| 2465 |
+
"loss": 0.329,
|
| 2466 |
+
"step": 1750
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.8636811023622047,
|
| 2470 |
+
"grad_norm": 2.260582620281468,
|
| 2471 |
+
"learning_rate": 5.559420818517702e-07,
|
| 2472 |
+
"loss": 0.3352,
|
| 2473 |
+
"step": 1755
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.8661417322834646,
|
| 2477 |
+
"grad_norm": 1.905722889331543,
|
| 2478 |
+
"learning_rate": 5.364167558957267e-07,
|
| 2479 |
+
"loss": 0.3241,
|
| 2480 |
+
"step": 1760
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.8686023622047244,
|
| 2484 |
+
"grad_norm": 2.08290721357846,
|
| 2485 |
+
"learning_rate": 5.172210156223745e-07,
|
| 2486 |
+
"loss": 0.3336,
|
| 2487 |
+
"step": 1765
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.8710629921259843,
|
| 2491 |
+
"grad_norm": 2.0264311598023728,
|
| 2492 |
+
"learning_rate": 4.983562784223645e-07,
|
| 2493 |
+
"loss": 0.3191,
|
| 2494 |
+
"step": 1770
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.8735236220472441,
|
| 2498 |
+
"grad_norm": 1.9306929816740388,
|
| 2499 |
+
"learning_rate": 4.798239372454738e-07,
|
| 2500 |
+
"loss": 0.3256,
|
| 2501 |
+
"step": 1775
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.8759842519685039,
|
| 2505 |
+
"grad_norm": 1.9166412725439652,
|
| 2506 |
+
"learning_rate": 4.6162536049775387e-07,
|
| 2507 |
+
"loss": 0.3284,
|
| 2508 |
+
"step": 1780
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.8784448818897638,
|
| 2512 |
+
"grad_norm": 1.9640277798271495,
|
| 2513 |
+
"learning_rate": 4.437618919404851e-07,
|
| 2514 |
+
"loss": 0.3233,
|
| 2515 |
+
"step": 1785
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.8809055118110236,
|
| 2519 |
+
"grad_norm": 1.9101328543862819,
|
| 2520 |
+
"learning_rate": 4.262348505909608e-07,
|
| 2521 |
+
"loss": 0.326,
|
| 2522 |
+
"step": 1790
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.8833661417322834,
|
| 2526 |
+
"grad_norm": 1.9078892149815414,
|
| 2527 |
+
"learning_rate": 4.0904553062508677e-07,
|
| 2528 |
+
"loss": 0.3273,
|
| 2529 |
+
"step": 1795
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.8858267716535433,
|
| 2533 |
+
"grad_norm": 1.92469620169423,
|
| 2534 |
+
"learning_rate": 3.9219520128182087e-07,
|
| 2535 |
+
"loss": 0.3222,
|
| 2536 |
+
"step": 1800
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.8882874015748031,
|
| 2540 |
+
"grad_norm": 1.8872478826399433,
|
| 2541 |
+
"learning_rate": 3.756851067694606e-07,
|
| 2542 |
+
"loss": 0.3233,
|
| 2543 |
+
"step": 1805
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.890748031496063,
|
| 2547 |
+
"grad_norm": 2.1313117581123366,
|
| 2548 |
+
"learning_rate": 3.5951646617376603e-07,
|
| 2549 |
+
"loss": 0.3309,
|
| 2550 |
+
"step": 1810
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.8932086614173228,
|
| 2554 |
+
"grad_norm": 1.857899228482062,
|
| 2555 |
+
"learning_rate": 3.436904733679436e-07,
|
| 2556 |
+
"loss": 0.3231,
|
| 2557 |
+
"step": 1815
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.8956692913385826,
|
| 2561 |
+
"grad_norm": 1.9630046166560744,
|
| 2562 |
+
"learning_rate": 3.2820829692449984e-07,
|
| 2563 |
+
"loss": 0.3314,
|
| 2564 |
+
"step": 1820
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.8981299212598425,
|
| 2568 |
+
"grad_norm": 1.878464206381759,
|
| 2569 |
+
"learning_rate": 3.130710800289416e-07,
|
| 2570 |
+
"loss": 0.315,
|
| 2571 |
+
"step": 1825
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.9005905511811023,
|
| 2575 |
+
"grad_norm": 1.948299936140257,
|
| 2576 |
+
"learning_rate": 2.982799403953801e-07,
|
| 2577 |
+
"loss": 0.3245,
|
| 2578 |
+
"step": 1830
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.9030511811023622,
|
| 2582 |
+
"grad_norm": 1.9803613486774017,
|
| 2583 |
+
"learning_rate": 2.8383597018398876e-07,
|
| 2584 |
+
"loss": 0.3106,
|
| 2585 |
+
"step": 1835
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.905511811023622,
|
| 2589 |
+
"grad_norm": 1.9228830866585032,
|
| 2590 |
+
"learning_rate": 2.697402359203638e-07,
|
| 2591 |
+
"loss": 0.3206,
|
| 2592 |
+
"step": 1840
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.9079724409448819,
|
| 2596 |
+
"grad_norm": 1.8821637790876162,
|
| 2597 |
+
"learning_rate": 2.559937784167743e-07,
|
| 2598 |
+
"loss": 0.3158,
|
| 2599 |
+
"step": 1845
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.9104330708661418,
|
| 2603 |
+
"grad_norm": 2.0523446694418364,
|
| 2604 |
+
"learning_rate": 2.4259761269530667e-07,
|
| 2605 |
+
"loss": 0.3182,
|
| 2606 |
+
"step": 1850
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.9128937007874016,
|
| 2610 |
+
"grad_norm": 1.9067478250870236,
|
| 2611 |
+
"learning_rate": 2.2955272791291894e-07,
|
| 2612 |
+
"loss": 0.33,
|
| 2613 |
+
"step": 1855
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.9153543307086615,
|
| 2617 |
+
"grad_norm": 1.8161748223299183,
|
| 2618 |
+
"learning_rate": 2.1686008728840301e-07,
|
| 2619 |
+
"loss": 0.3195,
|
| 2620 |
+
"step": 1860
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.9178149606299213,
|
| 2624 |
+
"grad_norm": 1.9778227776433959,
|
| 2625 |
+
"learning_rate": 2.0452062803126005e-07,
|
| 2626 |
+
"loss": 0.3229,
|
| 2627 |
+
"step": 1865
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.9202755905511811,
|
| 2631 |
+
"grad_norm": 1.8169343313243027,
|
| 2632 |
+
"learning_rate": 1.9253526127249787e-07,
|
| 2633 |
+
"loss": 0.328,
|
| 2634 |
+
"step": 1870
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.922736220472441,
|
| 2638 |
+
"grad_norm": 1.9674779916821903,
|
| 2639 |
+
"learning_rate": 1.8090487199735663e-07,
|
| 2640 |
+
"loss": 0.3235,
|
| 2641 |
+
"step": 1875
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.9251968503937008,
|
| 2645 |
+
"grad_norm": 2.066617055366464,
|
| 2646 |
+
"learning_rate": 1.6963031897995863e-07,
|
| 2647 |
+
"loss": 0.3216,
|
| 2648 |
+
"step": 1880
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.9276574803149606,
|
| 2652 |
+
"grad_norm": 1.9662299527781584,
|
| 2653 |
+
"learning_rate": 1.5871243471990372e-07,
|
| 2654 |
+
"loss": 0.3279,
|
| 2655 |
+
"step": 1885
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.9301181102362205,
|
| 2659 |
+
"grad_norm": 2.0282199981854414,
|
| 2660 |
+
"learning_rate": 1.4815202538079e-07,
|
| 2661 |
+
"loss": 0.314,
|
| 2662 |
+
"step": 1890
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.9325787401574803,
|
| 2666 |
+
"grad_norm": 1.9766928164218986,
|
| 2667 |
+
"learning_rate": 1.379498707306942e-07,
|
| 2668 |
+
"loss": 0.3152,
|
| 2669 |
+
"step": 1895
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.9350393700787402,
|
| 2673 |
+
"grad_norm": 1.8773781832541783,
|
| 2674 |
+
"learning_rate": 1.28106724084594e-07,
|
| 2675 |
+
"loss": 0.3286,
|
| 2676 |
+
"step": 1900
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.9375,
|
| 2680 |
+
"grad_norm": 1.9756880092340237,
|
| 2681 |
+
"learning_rate": 1.1862331224873902e-07,
|
| 2682 |
+
"loss": 0.3211,
|
| 2683 |
+
"step": 1905
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.9399606299212598,
|
| 2687 |
+
"grad_norm": 1.946646764579541,
|
| 2688 |
+
"learning_rate": 1.0950033546699213e-07,
|
| 2689 |
+
"loss": 0.3243,
|
| 2690 |
+
"step": 1910
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.9424212598425197,
|
| 2694 |
+
"grad_norm": 1.8898884090335948,
|
| 2695 |
+
"learning_rate": 1.0073846736911697e-07,
|
| 2696 |
+
"loss": 0.3274,
|
| 2697 |
+
"step": 1915
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.9448818897637795,
|
| 2701 |
+
"grad_norm": 1.816110689404615,
|
| 2702 |
+
"learning_rate": 9.233835492104326e-08,
|
| 2703 |
+
"loss": 0.3157,
|
| 2704 |
+
"step": 1920
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.9473425196850394,
|
| 2708 |
+
"grad_norm": 1.903238646884656,
|
| 2709 |
+
"learning_rate": 8.430061837709058e-08,
|
| 2710 |
+
"loss": 0.3158,
|
| 2711 |
+
"step": 1925
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.9498031496062992,
|
| 2715 |
+
"grad_norm": 1.8739160354182305,
|
| 2716 |
+
"learning_rate": 7.662585123417609e-08,
|
| 2717 |
+
"loss": 0.3136,
|
| 2718 |
+
"step": 1930
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.952263779527559,
|
| 2722 |
+
"grad_norm": 2.015422876061845,
|
| 2723 |
+
"learning_rate": 6.931462018798407e-08,
|
| 2724 |
+
"loss": 0.3274,
|
| 2725 |
+
"step": 1935
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.9547244094488189,
|
| 2729 |
+
"grad_norm": 1.7724513355973053,
|
| 2730 |
+
"learning_rate": 6.236746509112824e-08,
|
| 2731 |
+
"loss": 0.3288,
|
| 2732 |
+
"step": 1940
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.9571850393700787,
|
| 2736 |
+
"grad_norm": 1.9490022383605574,
|
| 2737 |
+
"learning_rate": 5.5784898913284754e-08,
|
| 2738 |
+
"loss": 0.3268,
|
| 2739 |
+
"step": 1945
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.9596456692913385,
|
| 2743 |
+
"grad_norm": 2.00148527557363,
|
| 2744 |
+
"learning_rate": 4.9567407703319247e-08,
|
| 2745 |
+
"loss": 0.327,
|
| 2746 |
+
"step": 1950
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.9621062992125984,
|
| 2750 |
+
"grad_norm": 1.9582200223538588,
|
| 2751 |
+
"learning_rate": 4.3715450553393765e-08,
|
| 2752 |
+
"loss": 0.3252,
|
| 2753 |
+
"step": 1955
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.9645669291338582,
|
| 2757 |
+
"grad_norm": 1.837922483605557,
|
| 2758 |
+
"learning_rate": 3.8229459565070074e-08,
|
| 2759 |
+
"loss": 0.31,
|
| 2760 |
+
"step": 1960
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.9670275590551181,
|
| 2764 |
+
"grad_norm": 1.910793470177359,
|
| 2765 |
+
"learning_rate": 3.3109839817404564e-08,
|
| 2766 |
+
"loss": 0.318,
|
| 2767 |
+
"step": 1965
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.969488188976378,
|
| 2771 |
+
"grad_norm": 1.8816675788362491,
|
| 2772 |
+
"learning_rate": 2.8356969337035578e-08,
|
| 2773 |
+
"loss": 0.3262,
|
| 2774 |
+
"step": 1970
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.9719488188976378,
|
| 2778 |
+
"grad_norm": 1.991419593684435,
|
| 2779 |
+
"learning_rate": 2.3971199070271234e-08,
|
| 2780 |
+
"loss": 0.3217,
|
| 2781 |
+
"step": 1975
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.9744094488188977,
|
| 2785 |
+
"grad_norm": 1.8312406446386904,
|
| 2786 |
+
"learning_rate": 1.99528528571763e-08,
|
| 2787 |
+
"loss": 0.3267,
|
| 2788 |
+
"step": 1980
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.9768700787401575,
|
| 2792 |
+
"grad_norm": 1.9479093178370912,
|
| 2793 |
+
"learning_rate": 1.6302227407660744e-08,
|
| 2794 |
+
"loss": 0.3209,
|
| 2795 |
+
"step": 1985
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.9793307086614174,
|
| 2799 |
+
"grad_norm": 1.9735060238944024,
|
| 2800 |
+
"learning_rate": 1.3019592279569503e-08,
|
| 2801 |
+
"loss": 0.3243,
|
| 2802 |
+
"step": 1990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.9817913385826772,
|
| 2806 |
+
"grad_norm": 2.043603259800287,
|
| 2807 |
+
"learning_rate": 1.0105189858779507e-08,
|
| 2808 |
+
"loss": 0.323,
|
| 2809 |
+
"step": 1995
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.984251968503937,
|
| 2813 |
+
"grad_norm": 1.852960962616925,
|
| 2814 |
+
"learning_rate": 7.559235341302872e-09,
|
| 2815 |
+
"loss": 0.322,
|
| 2816 |
+
"step": 2000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.9867125984251969,
|
| 2820 |
+
"grad_norm": 1.8539802968962942,
|
| 2821 |
+
"learning_rate": 5.381916717395186e-09,
|
| 2822 |
+
"loss": 0.3222,
|
| 2823 |
+
"step": 2005
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.9891732283464567,
|
| 2827 |
+
"grad_norm": 1.9807602796212282,
|
| 2828 |
+
"learning_rate": 3.573394757676596e-09,
|
| 2829 |
+
"loss": 0.3203,
|
| 2830 |
+
"step": 2010
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.9916338582677166,
|
| 2834 |
+
"grad_norm": 1.9586954229212994,
|
| 2835 |
+
"learning_rate": 2.1338030012596488e-09,
|
| 2836 |
+
"loss": 0.3272,
|
| 2837 |
+
"step": 2015
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.9940944881889764,
|
| 2841 |
+
"grad_norm": 1.966512685382069,
|
| 2842 |
+
"learning_rate": 1.0632477458888401e-09,
|
| 2843 |
+
"loss": 0.3285,
|
| 2844 |
+
"step": 2020
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.9965551181102362,
|
| 2848 |
+
"grad_norm": 1.8305164111174446,
|
| 2849 |
+
"learning_rate": 3.618080400924484e-10,
|
| 2850 |
+
"loss": 0.3272,
|
| 2851 |
+
"step": 2025
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.9990157480314961,
|
| 2855 |
+
"grad_norm": 1.8345259182578462,
|
| 2856 |
+
"learning_rate": 2.9535677343872637e-11,
|
| 2857 |
+
"loss": 0.3234,
|
| 2858 |
+
"step": 2030
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 1.0,
|
| 2862 |
+
"eval_runtime": 3.3017,
|
| 2863 |
+
"eval_samples_per_second": 3.029,
|
| 2864 |
+
"eval_steps_per_second": 0.909,
|
| 2865 |
+
"step": 2032
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 1.0,
|
| 2869 |
+
"step": 2032,
|
| 2870 |
+
"total_flos": 212729730170880.0,
|
| 2871 |
+
"train_loss": 0.4844009231218672,
|
| 2872 |
+
"train_runtime": 18284.3129,
|
| 2873 |
+
"train_samples_per_second": 1.778,
|
| 2874 |
+
"train_steps_per_second": 0.111
|
| 2875 |
+
}
|
| 2876 |
+
],
|
| 2877 |
+
"logging_steps": 5,
|
| 2878 |
+
"max_steps": 2032,
|
| 2879 |
+
"num_input_tokens_seen": 0,
|
| 2880 |
+
"num_train_epochs": 1,
|
| 2881 |
+
"save_steps": 100,
|
| 2882 |
+
"stateful_callbacks": {
|
| 2883 |
+
"TrainerControl": {
|
| 2884 |
+
"args": {
|
| 2885 |
+
"should_epoch_stop": false,
|
| 2886 |
+
"should_evaluate": false,
|
| 2887 |
+
"should_log": false,
|
| 2888 |
+
"should_save": true,
|
| 2889 |
+
"should_training_stop": true
|
| 2890 |
+
},
|
| 2891 |
+
"attributes": {}
|
| 2892 |
+
}
|
| 2893 |
+
},
|
| 2894 |
+
"total_flos": 212729730170880.0,
|
| 2895 |
+
"train_batch_size": 4,
|
| 2896 |
+
"trial_name": null,
|
| 2897 |
+
"trial_params": null
|
| 2898 |
+
}
|