Commit
•
d449b8e
1
Parent(s):
c50c794
Model save
Browse files- README.md +69 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +3111 -0
README.md
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
|
| 3 |
+
library_name: peft
|
| 4 |
+
license: llama3.1
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- sft
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: barc-llama3.1-8b-instruct-lora64-induction-gpt4-desc-4omini20k_lr2e-4_epoch3
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# barc-llama3.1-8b-instruct-lora64-induction-gpt4-desc-4omini20k_lr2e-4_epoch3
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) on an unknown dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.2928
|
| 22 |
+
|
| 23 |
+
## Model description
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Intended uses & limitations
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training and evaluation data
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Training procedure
|
| 36 |
+
|
| 37 |
+
### Training hyperparameters
|
| 38 |
+
|
| 39 |
+
The following hyperparameters were used during training:
|
| 40 |
+
- learning_rate: 0.0002
|
| 41 |
+
- train_batch_size: 8
|
| 42 |
+
- eval_batch_size: 4
|
| 43 |
+
- seed: 42
|
| 44 |
+
- distributed_type: multi-GPU
|
| 45 |
+
- num_devices: 8
|
| 46 |
+
- gradient_accumulation_steps: 2
|
| 47 |
+
- total_train_batch_size: 128
|
| 48 |
+
- total_eval_batch_size: 32
|
| 49 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 50 |
+
- lr_scheduler_type: cosine
|
| 51 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 52 |
+
- num_epochs: 3
|
| 53 |
+
|
| 54 |
+
### Training results
|
| 55 |
+
|
| 56 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 57 |
+
|:-------------:|:------:|:----:|:---------------:|
|
| 58 |
+
| 0.3214 | 0.9966 | 145 | 0.3190 |
|
| 59 |
+
| 0.2761 | 2.0 | 291 | 0.2950 |
|
| 60 |
+
| 0.2592 | 2.9897 | 435 | 0.2928 |
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
### Framework versions
|
| 64 |
+
|
| 65 |
+
- PEFT 0.13.0
|
| 66 |
+
- Transformers 4.45.0.dev0
|
| 67 |
+
- Pytorch 2.4.1+cu121
|
| 68 |
+
- Datasets 3.0.0
|
| 69 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.9896907216494846,
|
| 3 |
+
"total_flos": 1.330526714897367e+17,
|
| 4 |
+
"train_loss": 0.3336494640714821,
|
| 5 |
+
"train_runtime": 4477.4048,
|
| 6 |
+
"train_samples": 18561,
|
| 7 |
+
"train_samples_per_second": 12.436,
|
| 8 |
+
"train_steps_per_second": 0.097
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.9896907216494846,
|
| 3 |
+
"total_flos": 1.330526714897367e+17,
|
| 4 |
+
"train_loss": 0.3336494640714821,
|
| 5 |
+
"train_runtime": 4477.4048,
|
| 6 |
+
"train_samples": 18561,
|
| 7 |
+
"train_samples_per_second": 12.436,
|
| 8 |
+
"train_steps_per_second": 0.097
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3111 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.9896907216494846,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 435,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.006872852233676976,
|
| 13 |
+
"grad_norm": 0.9486322019687681,
|
| 14 |
+
"learning_rate": 4.5454545454545455e-06,
|
| 15 |
+
"loss": 1.3163,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.013745704467353952,
|
| 20 |
+
"grad_norm": 0.9581819297739079,
|
| 21 |
+
"learning_rate": 9.090909090909091e-06,
|
| 22 |
+
"loss": 1.3039,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.020618556701030927,
|
| 27 |
+
"grad_norm": 1.0054838528858523,
|
| 28 |
+
"learning_rate": 1.3636363636363637e-05,
|
| 29 |
+
"loss": 1.3912,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.027491408934707903,
|
| 34 |
+
"grad_norm": 0.973313664446148,
|
| 35 |
+
"learning_rate": 1.8181818181818182e-05,
|
| 36 |
+
"loss": 1.3176,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03436426116838488,
|
| 41 |
+
"grad_norm": 0.9112881771285355,
|
| 42 |
+
"learning_rate": 2.272727272727273e-05,
|
| 43 |
+
"loss": 1.2813,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.041237113402061855,
|
| 48 |
+
"grad_norm": 0.868853801792406,
|
| 49 |
+
"learning_rate": 2.7272727272727273e-05,
|
| 50 |
+
"loss": 1.2778,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.048109965635738834,
|
| 55 |
+
"grad_norm": 0.7999023078891502,
|
| 56 |
+
"learning_rate": 3.181818181818182e-05,
|
| 57 |
+
"loss": 1.2184,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.054982817869415807,
|
| 62 |
+
"grad_norm": 0.6175826046713091,
|
| 63 |
+
"learning_rate": 3.6363636363636364e-05,
|
| 64 |
+
"loss": 1.1459,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.061855670103092786,
|
| 69 |
+
"grad_norm": 0.5099365471572147,
|
| 70 |
+
"learning_rate": 4.0909090909090915e-05,
|
| 71 |
+
"loss": 1.0493,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.06872852233676977,
|
| 76 |
+
"grad_norm": 0.5444792063302974,
|
| 77 |
+
"learning_rate": 4.545454545454546e-05,
|
| 78 |
+
"loss": 1.0206,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.07560137457044673,
|
| 83 |
+
"grad_norm": 0.612097075909489,
|
| 84 |
+
"learning_rate": 5e-05,
|
| 85 |
+
"loss": 0.9556,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.08247422680412371,
|
| 90 |
+
"grad_norm": 0.5965134047415941,
|
| 91 |
+
"learning_rate": 5.4545454545454546e-05,
|
| 92 |
+
"loss": 0.8832,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.08934707903780069,
|
| 97 |
+
"grad_norm": 0.6321048435692244,
|
| 98 |
+
"learning_rate": 5.90909090909091e-05,
|
| 99 |
+
"loss": 0.8113,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.09621993127147767,
|
| 104 |
+
"grad_norm": 0.5433713534011196,
|
| 105 |
+
"learning_rate": 6.363636363636364e-05,
|
| 106 |
+
"loss": 0.7699,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.10309278350515463,
|
| 111 |
+
"grad_norm": 0.5641051016325467,
|
| 112 |
+
"learning_rate": 6.818181818181818e-05,
|
| 113 |
+
"loss": 0.6987,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.10996563573883161,
|
| 118 |
+
"grad_norm": 0.4741800529459741,
|
| 119 |
+
"learning_rate": 7.272727272727273e-05,
|
| 120 |
+
"loss": 0.6418,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.11683848797250859,
|
| 125 |
+
"grad_norm": 0.31919675330271313,
|
| 126 |
+
"learning_rate": 7.727272727272727e-05,
|
| 127 |
+
"loss": 0.5883,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.12371134020618557,
|
| 132 |
+
"grad_norm": 0.2704850432116006,
|
| 133 |
+
"learning_rate": 8.181818181818183e-05,
|
| 134 |
+
"loss": 0.576,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.13058419243986255,
|
| 139 |
+
"grad_norm": 0.26062797028012374,
|
| 140 |
+
"learning_rate": 8.636363636363637e-05,
|
| 141 |
+
"loss": 0.5267,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.13745704467353953,
|
| 146 |
+
"grad_norm": 0.18740805347521763,
|
| 147 |
+
"learning_rate": 9.090909090909092e-05,
|
| 148 |
+
"loss": 0.5469,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.14432989690721648,
|
| 153 |
+
"grad_norm": 0.22017175307824505,
|
| 154 |
+
"learning_rate": 9.545454545454546e-05,
|
| 155 |
+
"loss": 0.5226,
|
| 156 |
+
"step": 21
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.15120274914089346,
|
| 160 |
+
"grad_norm": 0.1887039063728806,
|
| 161 |
+
"learning_rate": 0.0001,
|
| 162 |
+
"loss": 0.4873,
|
| 163 |
+
"step": 22
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.15807560137457044,
|
| 167 |
+
"grad_norm": 0.1793474824943474,
|
| 168 |
+
"learning_rate": 0.00010454545454545455,
|
| 169 |
+
"loss": 0.4978,
|
| 170 |
+
"step": 23
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.16494845360824742,
|
| 174 |
+
"grad_norm": 0.23697858977104094,
|
| 175 |
+
"learning_rate": 0.00010909090909090909,
|
| 176 |
+
"loss": 0.4814,
|
| 177 |
+
"step": 24
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.1718213058419244,
|
| 181 |
+
"grad_norm": 0.17636901545890651,
|
| 182 |
+
"learning_rate": 0.00011363636363636365,
|
| 183 |
+
"loss": 0.5103,
|
| 184 |
+
"step": 25
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.17869415807560138,
|
| 188 |
+
"grad_norm": 0.16142348168311232,
|
| 189 |
+
"learning_rate": 0.0001181818181818182,
|
| 190 |
+
"loss": 0.4773,
|
| 191 |
+
"step": 26
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.18556701030927836,
|
| 195 |
+
"grad_norm": 0.1425385061824693,
|
| 196 |
+
"learning_rate": 0.00012272727272727272,
|
| 197 |
+
"loss": 0.492,
|
| 198 |
+
"step": 27
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.19243986254295534,
|
| 202 |
+
"grad_norm": 0.1265888154640839,
|
| 203 |
+
"learning_rate": 0.00012727272727272728,
|
| 204 |
+
"loss": 0.4854,
|
| 205 |
+
"step": 28
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.19931271477663232,
|
| 209 |
+
"grad_norm": 0.11449596495934233,
|
| 210 |
+
"learning_rate": 0.0001318181818181818,
|
| 211 |
+
"loss": 0.4615,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.20618556701030927,
|
| 216 |
+
"grad_norm": 0.12461604075605497,
|
| 217 |
+
"learning_rate": 0.00013636363636363637,
|
| 218 |
+
"loss": 0.4703,
|
| 219 |
+
"step": 30
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.21305841924398625,
|
| 223 |
+
"grad_norm": 0.12169420338653658,
|
| 224 |
+
"learning_rate": 0.00014090909090909093,
|
| 225 |
+
"loss": 0.4433,
|
| 226 |
+
"step": 31
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.21993127147766323,
|
| 230 |
+
"grad_norm": 0.10581158171701577,
|
| 231 |
+
"learning_rate": 0.00014545454545454546,
|
| 232 |
+
"loss": 0.4397,
|
| 233 |
+
"step": 32
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.2268041237113402,
|
| 237 |
+
"grad_norm": 0.11163612722377497,
|
| 238 |
+
"learning_rate": 0.00015000000000000001,
|
| 239 |
+
"loss": 0.4605,
|
| 240 |
+
"step": 33
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.23367697594501718,
|
| 244 |
+
"grad_norm": 0.11309459248751377,
|
| 245 |
+
"learning_rate": 0.00015454545454545454,
|
| 246 |
+
"loss": 0.4495,
|
| 247 |
+
"step": 34
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.24054982817869416,
|
| 251 |
+
"grad_norm": 0.11367215299047806,
|
| 252 |
+
"learning_rate": 0.0001590909090909091,
|
| 253 |
+
"loss": 0.4308,
|
| 254 |
+
"step": 35
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.24742268041237114,
|
| 258 |
+
"grad_norm": 0.10809464118167192,
|
| 259 |
+
"learning_rate": 0.00016363636363636366,
|
| 260 |
+
"loss": 0.4368,
|
| 261 |
+
"step": 36
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.2542955326460481,
|
| 265 |
+
"grad_norm": 0.10502574406388546,
|
| 266 |
+
"learning_rate": 0.0001681818181818182,
|
| 267 |
+
"loss": 0.4196,
|
| 268 |
+
"step": 37
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.2611683848797251,
|
| 272 |
+
"grad_norm": 0.10320182075757336,
|
| 273 |
+
"learning_rate": 0.00017272727272727275,
|
| 274 |
+
"loss": 0.42,
|
| 275 |
+
"step": 38
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.26804123711340205,
|
| 279 |
+
"grad_norm": 0.11138418873828733,
|
| 280 |
+
"learning_rate": 0.00017727272727272728,
|
| 281 |
+
"loss": 0.4514,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.27491408934707906,
|
| 286 |
+
"grad_norm": 0.11111081000023773,
|
| 287 |
+
"learning_rate": 0.00018181818181818183,
|
| 288 |
+
"loss": 0.4162,
|
| 289 |
+
"step": 40
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.281786941580756,
|
| 293 |
+
"grad_norm": 0.10773385531295475,
|
| 294 |
+
"learning_rate": 0.00018636363636363636,
|
| 295 |
+
"loss": 0.4058,
|
| 296 |
+
"step": 41
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.28865979381443296,
|
| 300 |
+
"grad_norm": 0.10631032605354059,
|
| 301 |
+
"learning_rate": 0.00019090909090909092,
|
| 302 |
+
"loss": 0.4029,
|
| 303 |
+
"step": 42
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.29553264604810997,
|
| 307 |
+
"grad_norm": 0.10836378456827221,
|
| 308 |
+
"learning_rate": 0.00019545454545454548,
|
| 309 |
+
"loss": 0.4096,
|
| 310 |
+
"step": 43
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.3024054982817869,
|
| 314 |
+
"grad_norm": 0.10880423378093391,
|
| 315 |
+
"learning_rate": 0.0002,
|
| 316 |
+
"loss": 0.4196,
|
| 317 |
+
"step": 44
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.30927835051546393,
|
| 321 |
+
"grad_norm": 0.1066915228080529,
|
| 322 |
+
"learning_rate": 0.00019999677214588312,
|
| 323 |
+
"loss": 0.4049,
|
| 324 |
+
"step": 45
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.3161512027491409,
|
| 328 |
+
"grad_norm": 0.1009132049783852,
|
| 329 |
+
"learning_rate": 0.00019998708879191335,
|
| 330 |
+
"loss": 0.4015,
|
| 331 |
+
"step": 46
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.3230240549828179,
|
| 335 |
+
"grad_norm": 0.10423644166095698,
|
| 336 |
+
"learning_rate": 0.00019997095056321971,
|
| 337 |
+
"loss": 0.3989,
|
| 338 |
+
"step": 47
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.32989690721649484,
|
| 342 |
+
"grad_norm": 0.10222126638135792,
|
| 343 |
+
"learning_rate": 0.00019994835850163924,
|
| 344 |
+
"loss": 0.3911,
|
| 345 |
+
"step": 48
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.33676975945017185,
|
| 349 |
+
"grad_norm": 0.09539411299428839,
|
| 350 |
+
"learning_rate": 0.00019991931406564944,
|
| 351 |
+
"loss": 0.3861,
|
| 352 |
+
"step": 49
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.3436426116838488,
|
| 356 |
+
"grad_norm": 0.10040146458411044,
|
| 357 |
+
"learning_rate": 0.00019988381913027442,
|
| 358 |
+
"loss": 0.3932,
|
| 359 |
+
"step": 50
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.35051546391752575,
|
| 363 |
+
"grad_norm": 0.10573179360993629,
|
| 364 |
+
"learning_rate": 0.00019984187598696363,
|
| 365 |
+
"loss": 0.3936,
|
| 366 |
+
"step": 51
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.35738831615120276,
|
| 370 |
+
"grad_norm": 0.10564577769908695,
|
| 371 |
+
"learning_rate": 0.00019979348734344398,
|
| 372 |
+
"loss": 0.4071,
|
| 373 |
+
"step": 52
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.3642611683848797,
|
| 377 |
+
"grad_norm": 0.10695511977009023,
|
| 378 |
+
"learning_rate": 0.00019973865632354516,
|
| 379 |
+
"loss": 0.3882,
|
| 380 |
+
"step": 53
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.3711340206185567,
|
| 384 |
+
"grad_norm": 0.10064589165597464,
|
| 385 |
+
"learning_rate": 0.0001996773864669978,
|
| 386 |
+
"loss": 0.3785,
|
| 387 |
+
"step": 54
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.37800687285223367,
|
| 391 |
+
"grad_norm": 0.09791457914129995,
|
| 392 |
+
"learning_rate": 0.00019960968172920516,
|
| 393 |
+
"loss": 0.3811,
|
| 394 |
+
"step": 55
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.3848797250859107,
|
| 398 |
+
"grad_norm": 0.1118566682283669,
|
| 399 |
+
"learning_rate": 0.00019953554648098748,
|
| 400 |
+
"loss": 0.3938,
|
| 401 |
+
"step": 56
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.3917525773195876,
|
| 405 |
+
"grad_norm": 0.10214824698553887,
|
| 406 |
+
"learning_rate": 0.0001994549855083001,
|
| 407 |
+
"loss": 0.3925,
|
| 408 |
+
"step": 57
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.39862542955326463,
|
| 412 |
+
"grad_norm": 0.1014103487175917,
|
| 413 |
+
"learning_rate": 0.0001993680040119244,
|
| 414 |
+
"loss": 0.3802,
|
| 415 |
+
"step": 58
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.4054982817869416,
|
| 419 |
+
"grad_norm": 0.0993499098682824,
|
| 420 |
+
"learning_rate": 0.00019927460760713197,
|
| 421 |
+
"loss": 0.381,
|
| 422 |
+
"step": 59
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.41237113402061853,
|
| 426 |
+
"grad_norm": 0.10206073076536111,
|
| 427 |
+
"learning_rate": 0.00019917480232332224,
|
| 428 |
+
"loss": 0.3657,
|
| 429 |
+
"step": 60
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.41924398625429554,
|
| 433 |
+
"grad_norm": 0.09828556302343387,
|
| 434 |
+
"learning_rate": 0.00019906859460363307,
|
| 435 |
+
"loss": 0.3761,
|
| 436 |
+
"step": 61
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.4261168384879725,
|
| 440 |
+
"grad_norm": 0.10139688359355604,
|
| 441 |
+
"learning_rate": 0.00019895599130452505,
|
| 442 |
+
"loss": 0.3749,
|
| 443 |
+
"step": 62
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.4329896907216495,
|
| 447 |
+
"grad_norm": 0.10711125411772042,
|
| 448 |
+
"learning_rate": 0.0001988369996953386,
|
| 449 |
+
"loss": 0.3706,
|
| 450 |
+
"step": 63
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.43986254295532645,
|
| 454 |
+
"grad_norm": 0.09724017597982272,
|
| 455 |
+
"learning_rate": 0.00019871162745782478,
|
| 456 |
+
"loss": 0.3705,
|
| 457 |
+
"step": 64
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.44673539518900346,
|
| 461 |
+
"grad_norm": 0.1012279008931452,
|
| 462 |
+
"learning_rate": 0.00019857988268564953,
|
| 463 |
+
"loss": 0.3712,
|
| 464 |
+
"step": 65
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.4536082474226804,
|
| 468 |
+
"grad_norm": 0.09806011793458508,
|
| 469 |
+
"learning_rate": 0.0001984417738838709,
|
| 470 |
+
"loss": 0.355,
|
| 471 |
+
"step": 66
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.46048109965635736,
|
| 475 |
+
"grad_norm": 0.10463912089089605,
|
| 476 |
+
"learning_rate": 0.0001982973099683902,
|
| 477 |
+
"loss": 0.3812,
|
| 478 |
+
"step": 67
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.46735395189003437,
|
| 482 |
+
"grad_norm": 0.10616651630870087,
|
| 483 |
+
"learning_rate": 0.0001981465002653763,
|
| 484 |
+
"loss": 0.3671,
|
| 485 |
+
"step": 68
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.4742268041237113,
|
| 489 |
+
"grad_norm": 0.11658375430444518,
|
| 490 |
+
"learning_rate": 0.00019798935451066361,
|
| 491 |
+
"loss": 0.3822,
|
| 492 |
+
"step": 69
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.48109965635738833,
|
| 496 |
+
"grad_norm": 0.10686011983880458,
|
| 497 |
+
"learning_rate": 0.0001978258828491236,
|
| 498 |
+
"loss": 0.3564,
|
| 499 |
+
"step": 70
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.4879725085910653,
|
| 503 |
+
"grad_norm": 0.10160970818733071,
|
| 504 |
+
"learning_rate": 0.00019765609583400977,
|
| 505 |
+
"loss": 0.3656,
|
| 506 |
+
"step": 71
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.4948453608247423,
|
| 510 |
+
"grad_norm": 0.10845572916815471,
|
| 511 |
+
"learning_rate": 0.0001974800044262764,
|
| 512 |
+
"loss": 0.3623,
|
| 513 |
+
"step": 72
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.5017182130584192,
|
| 517 |
+
"grad_norm": 0.10898867605088586,
|
| 518 |
+
"learning_rate": 0.00019729761999387103,
|
| 519 |
+
"loss": 0.3689,
|
| 520 |
+
"step": 73
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.5085910652920962,
|
| 524 |
+
"grad_norm": 0.10030807679917116,
|
| 525 |
+
"learning_rate": 0.00019710895431100046,
|
| 526 |
+
"loss": 0.3743,
|
| 527 |
+
"step": 74
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.5154639175257731,
|
| 531 |
+
"grad_norm": 0.10640353541210322,
|
| 532 |
+
"learning_rate": 0.00019691401955737072,
|
| 533 |
+
"loss": 0.3664,
|
| 534 |
+
"step": 75
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.5223367697594502,
|
| 538 |
+
"grad_norm": 0.10268077724904813,
|
| 539 |
+
"learning_rate": 0.00019671282831740076,
|
| 540 |
+
"loss": 0.3446,
|
| 541 |
+
"step": 76
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.5292096219931272,
|
| 545 |
+
"grad_norm": 0.1026968446802754,
|
| 546 |
+
"learning_rate": 0.00019650539357941003,
|
| 547 |
+
"loss": 0.3524,
|
| 548 |
+
"step": 77
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.5360824742268041,
|
| 552 |
+
"grad_norm": 0.10265145337121352,
|
| 553 |
+
"learning_rate": 0.00019629172873477995,
|
| 554 |
+
"loss": 0.3615,
|
| 555 |
+
"step": 78
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.5429553264604811,
|
| 559 |
+
"grad_norm": 0.10234137978544197,
|
| 560 |
+
"learning_rate": 0.00019607184757708951,
|
| 561 |
+
"loss": 0.348,
|
| 562 |
+
"step": 79
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.5498281786941581,
|
| 566 |
+
"grad_norm": 0.10414003831964998,
|
| 567 |
+
"learning_rate": 0.00019584576430122473,
|
| 568 |
+
"loss": 0.3455,
|
| 569 |
+
"step": 80
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.5567010309278351,
|
| 573 |
+
"grad_norm": 0.10371223386491027,
|
| 574 |
+
"learning_rate": 0.00019561349350246226,
|
| 575 |
+
"loss": 0.3378,
|
| 576 |
+
"step": 81
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.563573883161512,
|
| 580 |
+
"grad_norm": 0.10200261115662425,
|
| 581 |
+
"learning_rate": 0.00019537505017552716,
|
| 582 |
+
"loss": 0.3424,
|
| 583 |
+
"step": 82
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.570446735395189,
|
| 587 |
+
"grad_norm": 0.10679690224389146,
|
| 588 |
+
"learning_rate": 0.00019513044971362494,
|
| 589 |
+
"loss": 0.3381,
|
| 590 |
+
"step": 83
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.5773195876288659,
|
| 594 |
+
"grad_norm": 0.11747209193015294,
|
| 595 |
+
"learning_rate": 0.00019487970790744774,
|
| 596 |
+
"loss": 0.3608,
|
| 597 |
+
"step": 84
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.584192439862543,
|
| 601 |
+
"grad_norm": 0.10455385108750616,
|
| 602 |
+
"learning_rate": 0.000194622840944155,
|
| 603 |
+
"loss": 0.3513,
|
| 604 |
+
"step": 85
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.5910652920962199,
|
| 608 |
+
"grad_norm": 0.11103687968369377,
|
| 609 |
+
"learning_rate": 0.00019435986540632843,
|
| 610 |
+
"loss": 0.3322,
|
| 611 |
+
"step": 86
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.5979381443298969,
|
| 615 |
+
"grad_norm": 0.1044628007637211,
|
| 616 |
+
"learning_rate": 0.00019409079827090145,
|
| 617 |
+
"loss": 0.3452,
|
| 618 |
+
"step": 87
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.6048109965635738,
|
| 622 |
+
"grad_norm": 0.1062512518470538,
|
| 623 |
+
"learning_rate": 0.00019381565690806328,
|
| 624 |
+
"loss": 0.3295,
|
| 625 |
+
"step": 88
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.6116838487972509,
|
| 629 |
+
"grad_norm": 0.10737195453497204,
|
| 630 |
+
"learning_rate": 0.00019353445908013755,
|
| 631 |
+
"loss": 0.3322,
|
| 632 |
+
"step": 89
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.6185567010309279,
|
| 636 |
+
"grad_norm": 0.10702988631259788,
|
| 637 |
+
"learning_rate": 0.00019324722294043558,
|
| 638 |
+
"loss": 0.3436,
|
| 639 |
+
"step": 90
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.6254295532646048,
|
| 643 |
+
"grad_norm": 0.10765269601948493,
|
| 644 |
+
"learning_rate": 0.00019295396703208453,
|
| 645 |
+
"loss": 0.3628,
|
| 646 |
+
"step": 91
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.6323024054982818,
|
| 650 |
+
"grad_norm": 0.1002631772076248,
|
| 651 |
+
"learning_rate": 0.00019265471028683014,
|
| 652 |
+
"loss": 0.337,
|
| 653 |
+
"step": 92
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.6391752577319587,
|
| 657 |
+
"grad_norm": 0.10607703170572558,
|
| 658 |
+
"learning_rate": 0.00019234947202381486,
|
| 659 |
+
"loss": 0.3312,
|
| 660 |
+
"step": 93
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.6460481099656358,
|
| 664 |
+
"grad_norm": 0.10229636934864904,
|
| 665 |
+
"learning_rate": 0.00019203827194833026,
|
| 666 |
+
"loss": 0.3613,
|
| 667 |
+
"step": 94
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.6529209621993127,
|
| 671 |
+
"grad_norm": 0.10916337452983911,
|
| 672 |
+
"learning_rate": 0.00019172113015054532,
|
| 673 |
+
"loss": 0.3402,
|
| 674 |
+
"step": 95
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.6597938144329897,
|
| 678 |
+
"grad_norm": 0.10094159423939311,
|
| 679 |
+
"learning_rate": 0.00019139806710420914,
|
| 680 |
+
"loss": 0.3308,
|
| 681 |
+
"step": 96
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.6666666666666666,
|
| 685 |
+
"grad_norm": 0.10420707581199606,
|
| 686 |
+
"learning_rate": 0.00019106910366532942,
|
| 687 |
+
"loss": 0.3297,
|
| 688 |
+
"step": 97
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.6735395189003437,
|
| 692 |
+
"grad_norm": 0.11033214518450266,
|
| 693 |
+
"learning_rate": 0.000190734261070826,
|
| 694 |
+
"loss": 0.3368,
|
| 695 |
+
"step": 98
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.6804123711340206,
|
| 699 |
+
"grad_norm": 0.10648307368306249,
|
| 700 |
+
"learning_rate": 0.00019039356093715975,
|
| 701 |
+
"loss": 0.3436,
|
| 702 |
+
"step": 99
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.6872852233676976,
|
| 706 |
+
"grad_norm": 0.1023047486718038,
|
| 707 |
+
"learning_rate": 0.00019004702525893732,
|
| 708 |
+
"loss": 0.3389,
|
| 709 |
+
"step": 100
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.6941580756013745,
|
| 713 |
+
"grad_norm": 0.10884134971992637,
|
| 714 |
+
"learning_rate": 0.000189694676407491,
|
| 715 |
+
"loss": 0.3343,
|
| 716 |
+
"step": 101
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.7010309278350515,
|
| 720 |
+
"grad_norm": 0.10624545430750702,
|
| 721 |
+
"learning_rate": 0.0001893365371294346,
|
| 722 |
+
"loss": 0.334,
|
| 723 |
+
"step": 102
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.7079037800687286,
|
| 727 |
+
"grad_norm": 0.10418100794965049,
|
| 728 |
+
"learning_rate": 0.00018897263054519498,
|
| 729 |
+
"loss": 0.3332,
|
| 730 |
+
"step": 103
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.7147766323024055,
|
| 734 |
+
"grad_norm": 0.10442261089390925,
|
| 735 |
+
"learning_rate": 0.00018860298014751944,
|
| 736 |
+
"loss": 0.3231,
|
| 737 |
+
"step": 104
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.7216494845360825,
|
| 741 |
+
"grad_norm": 0.11510206196708461,
|
| 742 |
+
"learning_rate": 0.0001882276097999592,
|
| 743 |
+
"loss": 0.3331,
|
| 744 |
+
"step": 105
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.7285223367697594,
|
| 748 |
+
"grad_norm": 0.10447026389489032,
|
| 749 |
+
"learning_rate": 0.00018784654373532866,
|
| 750 |
+
"loss": 0.3454,
|
| 751 |
+
"step": 106
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.7353951890034365,
|
| 755 |
+
"grad_norm": 0.10590015702307645,
|
| 756 |
+
"learning_rate": 0.00018745980655414114,
|
| 757 |
+
"loss": 0.3283,
|
| 758 |
+
"step": 107
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.7422680412371134,
|
| 762 |
+
"grad_norm": 0.1019688644295514,
|
| 763 |
+
"learning_rate": 0.00018706742322302064,
|
| 764 |
+
"loss": 0.3368,
|
| 765 |
+
"step": 108
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.7491408934707904,
|
| 769 |
+
"grad_norm": 0.11013306581811545,
|
| 770 |
+
"learning_rate": 0.00018666941907309026,
|
| 771 |
+
"loss": 0.3371,
|
| 772 |
+
"step": 109
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.7560137457044673,
|
| 776 |
+
"grad_norm": 0.09950125567962896,
|
| 777 |
+
"learning_rate": 0.0001862658197983366,
|
| 778 |
+
"loss": 0.3206,
|
| 779 |
+
"step": 110
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.7628865979381443,
|
| 783 |
+
"grad_norm": 0.09961853222606298,
|
| 784 |
+
"learning_rate": 0.0001858566514539513,
|
| 785 |
+
"loss": 0.3245,
|
| 786 |
+
"step": 111
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.7697594501718213,
|
| 790 |
+
"grad_norm": 0.10822873754218582,
|
| 791 |
+
"learning_rate": 0.00018544194045464886,
|
| 792 |
+
"loss": 0.336,
|
| 793 |
+
"step": 112
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.7766323024054983,
|
| 797 |
+
"grad_norm": 0.10256746162360054,
|
| 798 |
+
"learning_rate": 0.00018502171357296144,
|
| 799 |
+
"loss": 0.3279,
|
| 800 |
+
"step": 113
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.7835051546391752,
|
| 804 |
+
"grad_norm": 0.09750927062180448,
|
| 805 |
+
"learning_rate": 0.0001845959979375104,
|
| 806 |
+
"loss": 0.3388,
|
| 807 |
+
"step": 114
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.7903780068728522,
|
| 811 |
+
"grad_norm": 0.09793726368683904,
|
| 812 |
+
"learning_rate": 0.00018416482103125506,
|
| 813 |
+
"loss": 0.3303,
|
| 814 |
+
"step": 115
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.7972508591065293,
|
| 818 |
+
"grad_norm": 0.10598098121402388,
|
| 819 |
+
"learning_rate": 0.0001837282106897185,
|
| 820 |
+
"loss": 0.3216,
|
| 821 |
+
"step": 116
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.8041237113402062,
|
| 825 |
+
"grad_norm": 0.10184730892936761,
|
| 826 |
+
"learning_rate": 0.00018328619509919044,
|
| 827 |
+
"loss": 0.3341,
|
| 828 |
+
"step": 117
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.8109965635738832,
|
| 832 |
+
"grad_norm": 0.10057188453324177,
|
| 833 |
+
"learning_rate": 0.0001828388027949078,
|
| 834 |
+
"loss": 0.3377,
|
| 835 |
+
"step": 118
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.8178694158075601,
|
| 839 |
+
"grad_norm": 0.10702981460753135,
|
| 840 |
+
"learning_rate": 0.00018238606265921238,
|
| 841 |
+
"loss": 0.3408,
|
| 842 |
+
"step": 119
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.8247422680412371,
|
| 846 |
+
"grad_norm": 0.10861206391221787,
|
| 847 |
+
"learning_rate": 0.00018192800391968642,
|
| 848 |
+
"loss": 0.3413,
|
| 849 |
+
"step": 120
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.8316151202749141,
|
| 853 |
+
"grad_norm": 0.10665429392095056,
|
| 854 |
+
"learning_rate": 0.00018146465614726567,
|
| 855 |
+
"loss": 0.3234,
|
| 856 |
+
"step": 121
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.8384879725085911,
|
| 860 |
+
"grad_norm": 0.10180559686021746,
|
| 861 |
+
"learning_rate": 0.00018099604925433043,
|
| 862 |
+
"loss": 0.3263,
|
| 863 |
+
"step": 122
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.845360824742268,
|
| 867 |
+
"grad_norm": 0.11025637152512564,
|
| 868 |
+
"learning_rate": 0.00018052221349277442,
|
| 869 |
+
"loss": 0.3295,
|
| 870 |
+
"step": 123
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.852233676975945,
|
| 874 |
+
"grad_norm": 0.10266709615387361,
|
| 875 |
+
"learning_rate": 0.00018004317945205197,
|
| 876 |
+
"loss": 0.3275,
|
| 877 |
+
"step": 124
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.8591065292096219,
|
| 881 |
+
"grad_norm": 0.10117796525027944,
|
| 882 |
+
"learning_rate": 0.0001795589780572031,
|
| 883 |
+
"loss": 0.3541,
|
| 884 |
+
"step": 125
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.865979381443299,
|
| 888 |
+
"grad_norm": 0.10526178050004703,
|
| 889 |
+
"learning_rate": 0.00017906964056685706,
|
| 890 |
+
"loss": 0.3312,
|
| 891 |
+
"step": 126
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.872852233676976,
|
| 895 |
+
"grad_norm": 0.1063992770321544,
|
| 896 |
+
"learning_rate": 0.00017857519857121458,
|
| 897 |
+
"loss": 0.3312,
|
| 898 |
+
"step": 127
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.8797250859106529,
|
| 902 |
+
"grad_norm": 0.10996856445056626,
|
| 903 |
+
"learning_rate": 0.00017807568399000822,
|
| 904 |
+
"loss": 0.3527,
|
| 905 |
+
"step": 128
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.8865979381443299,
|
| 909 |
+
"grad_norm": 0.09753881684843836,
|
| 910 |
+
"learning_rate": 0.000177571129070442,
|
| 911 |
+
"loss": 0.3117,
|
| 912 |
+
"step": 129
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.8934707903780069,
|
| 916 |
+
"grad_norm": 0.10008693889485847,
|
| 917 |
+
"learning_rate": 0.0001770615663851093,
|
| 918 |
+
"loss": 0.3188,
|
| 919 |
+
"step": 130
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.9003436426116839,
|
| 923 |
+
"grad_norm": 0.10344071862676044,
|
| 924 |
+
"learning_rate": 0.0001765470288298905,
|
| 925 |
+
"loss": 0.3252,
|
| 926 |
+
"step": 131
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.9072164948453608,
|
| 930 |
+
"grad_norm": 0.09310502474075653,
|
| 931 |
+
"learning_rate": 0.0001760275496218288,
|
| 932 |
+
"loss": 0.3058,
|
| 933 |
+
"step": 132
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.9140893470790378,
|
| 937 |
+
"grad_norm": 0.10039282316295665,
|
| 938 |
+
"learning_rate": 0.0001755031622969862,
|
| 939 |
+
"loss": 0.3468,
|
| 940 |
+
"step": 133
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.9209621993127147,
|
| 944 |
+
"grad_norm": 0.09871302822261965,
|
| 945 |
+
"learning_rate": 0.00017497390070827848,
|
| 946 |
+
"loss": 0.3241,
|
| 947 |
+
"step": 134
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.9278350515463918,
|
| 951 |
+
"grad_norm": 0.10019109474446979,
|
| 952 |
+
"learning_rate": 0.00017443979902328956,
|
| 953 |
+
"loss": 0.3247,
|
| 954 |
+
"step": 135
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.9347079037800687,
|
| 958 |
+
"grad_norm": 0.09742365798894928,
|
| 959 |
+
"learning_rate": 0.00017390089172206592,
|
| 960 |
+
"loss": 0.3268,
|
| 961 |
+
"step": 136
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.9415807560137457,
|
| 965 |
+
"grad_norm": 0.09613939400204266,
|
| 966 |
+
"learning_rate": 0.00017335721359489057,
|
| 967 |
+
"loss": 0.3017,
|
| 968 |
+
"step": 137
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.9484536082474226,
|
| 972 |
+
"grad_norm": 0.11015286314690642,
|
| 973 |
+
"learning_rate": 0.00017280879974003707,
|
| 974 |
+
"loss": 0.3369,
|
| 975 |
+
"step": 138
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.9553264604810997,
|
| 979 |
+
"grad_norm": 0.09848269873542065,
|
| 980 |
+
"learning_rate": 0.0001722556855615039,
|
| 981 |
+
"loss": 0.3215,
|
| 982 |
+
"step": 139
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.9621993127147767,
|
| 986 |
+
"grad_norm": 0.10356185439767568,
|
| 987 |
+
"learning_rate": 0.00017169790676672858,
|
| 988 |
+
"loss": 0.3209,
|
| 989 |
+
"step": 140
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.9690721649484536,
|
| 993 |
+
"grad_norm": 0.10200587363931662,
|
| 994 |
+
"learning_rate": 0.0001711354993642827,
|
| 995 |
+
"loss": 0.3156,
|
| 996 |
+
"step": 141
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.9759450171821306,
|
| 1000 |
+
"grad_norm": 0.10631936681232042,
|
| 1001 |
+
"learning_rate": 0.0001705684996615472,
|
| 1002 |
+
"loss": 0.3139,
|
| 1003 |
+
"step": 142
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.9828178694158075,
|
| 1007 |
+
"grad_norm": 0.09891246233686207,
|
| 1008 |
+
"learning_rate": 0.0001699969442623686,
|
| 1009 |
+
"loss": 0.3252,
|
| 1010 |
+
"step": 143
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.9896907216494846,
|
| 1014 |
+
"grad_norm": 0.09631426655006751,
|
| 1015 |
+
"learning_rate": 0.00016942087006469592,
|
| 1016 |
+
"loss": 0.3092,
|
| 1017 |
+
"step": 144
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.9965635738831615,
|
| 1021 |
+
"grad_norm": 0.1007256123012449,
|
| 1022 |
+
"learning_rate": 0.00016884031425819853,
|
| 1023 |
+
"loss": 0.3214,
|
| 1024 |
+
"step": 145
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.9965635738831615,
|
| 1028 |
+
"eval_loss": 0.31896448135375977,
|
| 1029 |
+
"eval_runtime": 31.3575,
|
| 1030 |
+
"eval_samples_per_second": 31.157,
|
| 1031 |
+
"eval_steps_per_second": 0.989,
|
| 1032 |
+
"step": 145
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 1.0034364261168385,
|
| 1036 |
+
"grad_norm": 0.09742513564214555,
|
| 1037 |
+
"learning_rate": 0.00016825531432186543,
|
| 1038 |
+
"loss": 0.3131,
|
| 1039 |
+
"step": 146
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.0103092783505154,
|
| 1043 |
+
"grad_norm": 0.1000843163672815,
|
| 1044 |
+
"learning_rate": 0.00016766590802158566,
|
| 1045 |
+
"loss": 0.308,
|
| 1046 |
+
"step": 147
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 1.0171821305841924,
|
| 1050 |
+
"grad_norm": 0.1018182338551042,
|
| 1051 |
+
"learning_rate": 0.0001670721334077103,
|
| 1052 |
+
"loss": 0.3077,
|
| 1053 |
+
"step": 148
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 1.0240549828178693,
|
| 1057 |
+
"grad_norm": 0.10928234134401811,
|
| 1058 |
+
"learning_rate": 0.00016647402881259598,
|
| 1059 |
+
"loss": 0.3137,
|
| 1060 |
+
"step": 149
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 1.0309278350515463,
|
| 1064 |
+
"grad_norm": 0.10482761572794695,
|
| 1065 |
+
"learning_rate": 0.00016587163284813032,
|
| 1066 |
+
"loss": 0.2969,
|
| 1067 |
+
"step": 150
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 1.0378006872852235,
|
| 1071 |
+
"grad_norm": 0.10755010011949057,
|
| 1072 |
+
"learning_rate": 0.00016526498440323914,
|
| 1073 |
+
"loss": 0.3027,
|
| 1074 |
+
"step": 151
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 1.0446735395189004,
|
| 1078 |
+
"grad_norm": 0.09961868507020075,
|
| 1079 |
+
"learning_rate": 0.0001646541226413761,
|
| 1080 |
+
"loss": 0.2918,
|
| 1081 |
+
"step": 152
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 1.0515463917525774,
|
| 1085 |
+
"grad_norm": 0.10275553359018949,
|
| 1086 |
+
"learning_rate": 0.00016403908699799425,
|
| 1087 |
+
"loss": 0.2935,
|
| 1088 |
+
"step": 153
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 1.0584192439862543,
|
| 1092 |
+
"grad_norm": 0.10314122954609171,
|
| 1093 |
+
"learning_rate": 0.00016341991717800023,
|
| 1094 |
+
"loss": 0.3107,
|
| 1095 |
+
"step": 154
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 1.0652920962199313,
|
| 1099 |
+
"grad_norm": 0.10201304020136209,
|
| 1100 |
+
"learning_rate": 0.00016279665315319114,
|
| 1101 |
+
"loss": 0.2925,
|
| 1102 |
+
"step": 155
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 1.0721649484536082,
|
| 1106 |
+
"grad_norm": 0.11295490175270977,
|
| 1107 |
+
"learning_rate": 0.0001621693351596739,
|
| 1108 |
+
"loss": 0.3165,
|
| 1109 |
+
"step": 156
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.0790378006872852,
|
| 1113 |
+
"grad_norm": 0.11187236691946645,
|
| 1114 |
+
"learning_rate": 0.00016153800369526788,
|
| 1115 |
+
"loss": 0.3008,
|
| 1116 |
+
"step": 157
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 1.0859106529209621,
|
| 1120 |
+
"grad_norm": 0.10649522892041895,
|
| 1121 |
+
"learning_rate": 0.0001609026995168904,
|
| 1122 |
+
"loss": 0.3072,
|
| 1123 |
+
"step": 158
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.0927835051546393,
|
| 1127 |
+
"grad_norm": 0.10138792178292377,
|
| 1128 |
+
"learning_rate": 0.00016026346363792567,
|
| 1129 |
+
"loss": 0.2968,
|
| 1130 |
+
"step": 159
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 1.0996563573883162,
|
| 1134 |
+
"grad_norm": 0.10490079005967201,
|
| 1135 |
+
"learning_rate": 0.00015962033732557686,
|
| 1136 |
+
"loss": 0.3074,
|
| 1137 |
+
"step": 160
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 1.1065292096219932,
|
| 1141 |
+
"grad_norm": 0.10339230680715779,
|
| 1142 |
+
"learning_rate": 0.00015897336209820239,
|
| 1143 |
+
"loss": 0.2904,
|
| 1144 |
+
"step": 161
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 1.1134020618556701,
|
| 1148 |
+
"grad_norm": 0.11480233498020057,
|
| 1149 |
+
"learning_rate": 0.00015832257972263523,
|
| 1150 |
+
"loss": 0.31,
|
| 1151 |
+
"step": 162
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 1.120274914089347,
|
| 1155 |
+
"grad_norm": 0.10620125747115913,
|
| 1156 |
+
"learning_rate": 0.00015766803221148673,
|
| 1157 |
+
"loss": 0.3041,
|
| 1158 |
+
"step": 163
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 1.127147766323024,
|
| 1162 |
+
"grad_norm": 0.10787426446881714,
|
| 1163 |
+
"learning_rate": 0.0001570097618204345,
|
| 1164 |
+
"loss": 0.3142,
|
| 1165 |
+
"step": 164
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.134020618556701,
|
| 1169 |
+
"grad_norm": 0.10375711742292576,
|
| 1170 |
+
"learning_rate": 0.00015634781104549442,
|
| 1171 |
+
"loss": 0.2916,
|
| 1172 |
+
"step": 165
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 1.140893470790378,
|
| 1176 |
+
"grad_norm": 0.1085251854270091,
|
| 1177 |
+
"learning_rate": 0.00015568222262027717,
|
| 1178 |
+
"loss": 0.2969,
|
| 1179 |
+
"step": 166
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.147766323024055,
|
| 1183 |
+
"grad_norm": 0.10620158303898176,
|
| 1184 |
+
"learning_rate": 0.00015501303951322943,
|
| 1185 |
+
"loss": 0.3069,
|
| 1186 |
+
"step": 167
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.1546391752577319,
|
| 1190 |
+
"grad_norm": 0.10221091343876472,
|
| 1191 |
+
"learning_rate": 0.00015434030492486023,
|
| 1192 |
+
"loss": 0.3015,
|
| 1193 |
+
"step": 168
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.161512027491409,
|
| 1197 |
+
"grad_norm": 0.11025720771873336,
|
| 1198 |
+
"learning_rate": 0.00015366406228495172,
|
| 1199 |
+
"loss": 0.2923,
|
| 1200 |
+
"step": 169
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.168384879725086,
|
| 1204 |
+
"grad_norm": 0.11019089511672239,
|
| 1205 |
+
"learning_rate": 0.00015298435524975572,
|
| 1206 |
+
"loss": 0.3088,
|
| 1207 |
+
"step": 170
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.175257731958763,
|
| 1211 |
+
"grad_norm": 0.1066893326930603,
|
| 1212 |
+
"learning_rate": 0.00015230122769917527,
|
| 1213 |
+
"loss": 0.3006,
|
| 1214 |
+
"step": 171
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 1.1821305841924399,
|
| 1218 |
+
"grad_norm": 0.11028825839961288,
|
| 1219 |
+
"learning_rate": 0.00015161472373393186,
|
| 1220 |
+
"loss": 0.3028,
|
| 1221 |
+
"step": 172
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.1890034364261168,
|
| 1225 |
+
"grad_norm": 0.10136590795578818,
|
| 1226 |
+
"learning_rate": 0.00015092488767271857,
|
| 1227 |
+
"loss": 0.2789,
|
| 1228 |
+
"step": 173
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.1958762886597938,
|
| 1232 |
+
"grad_norm": 0.1032678246834342,
|
| 1233 |
+
"learning_rate": 0.00015023176404933874,
|
| 1234 |
+
"loss": 0.2975,
|
| 1235 |
+
"step": 174
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.2027491408934707,
|
| 1239 |
+
"grad_norm": 0.11056648590020815,
|
| 1240 |
+
"learning_rate": 0.00014953539760983122,
|
| 1241 |
+
"loss": 0.3013,
|
| 1242 |
+
"step": 175
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.2096219931271477,
|
| 1246 |
+
"grad_norm": 0.09978579785299047,
|
| 1247 |
+
"learning_rate": 0.0001488358333095816,
|
| 1248 |
+
"loss": 0.2955,
|
| 1249 |
+
"step": 176
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.2164948453608249,
|
| 1253 |
+
"grad_norm": 0.10699681452783022,
|
| 1254 |
+
"learning_rate": 0.00014813311631041995,
|
| 1255 |
+
"loss": 0.2993,
|
| 1256 |
+
"step": 177
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.2233676975945018,
|
| 1260 |
+
"grad_norm": 0.10408363973129214,
|
| 1261 |
+
"learning_rate": 0.00014742729197770552,
|
| 1262 |
+
"loss": 0.2954,
|
| 1263 |
+
"step": 178
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.2302405498281788,
|
| 1267 |
+
"grad_norm": 0.1016249945819849,
|
| 1268 |
+
"learning_rate": 0.00014671840587739783,
|
| 1269 |
+
"loss": 0.2904,
|
| 1270 |
+
"step": 179
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 1.2371134020618557,
|
| 1274 |
+
"grad_norm": 0.10519072387735487,
|
| 1275 |
+
"learning_rate": 0.00014600650377311522,
|
| 1276 |
+
"loss": 0.2916,
|
| 1277 |
+
"step": 180
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.2439862542955327,
|
| 1281 |
+
"grad_norm": 0.1058597383518849,
|
| 1282 |
+
"learning_rate": 0.0001452916316231805,
|
| 1283 |
+
"loss": 0.301,
|
| 1284 |
+
"step": 181
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.2508591065292096,
|
| 1288 |
+
"grad_norm": 0.10402401073301179,
|
| 1289 |
+
"learning_rate": 0.00014457383557765386,
|
| 1290 |
+
"loss": 0.2875,
|
| 1291 |
+
"step": 182
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.2577319587628866,
|
| 1295 |
+
"grad_norm": 0.11052274366476848,
|
| 1296 |
+
"learning_rate": 0.00014385316197535372,
|
| 1297 |
+
"loss": 0.3033,
|
| 1298 |
+
"step": 183
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.2646048109965635,
|
| 1302 |
+
"grad_norm": 0.11220133909982184,
|
| 1303 |
+
"learning_rate": 0.00014312965734086518,
|
| 1304 |
+
"loss": 0.3048,
|
| 1305 |
+
"step": 184
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.2714776632302405,
|
| 1309 |
+
"grad_norm": 0.10111031103821645,
|
| 1310 |
+
"learning_rate": 0.0001424033683815365,
|
| 1311 |
+
"loss": 0.2883,
|
| 1312 |
+
"step": 185
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 1.2783505154639174,
|
| 1316 |
+
"grad_norm": 0.108844384730101,
|
| 1317 |
+
"learning_rate": 0.00014167434198446383,
|
| 1318 |
+
"loss": 0.3156,
|
| 1319 |
+
"step": 186
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.2852233676975944,
|
| 1323 |
+
"grad_norm": 0.11772254803070373,
|
| 1324 |
+
"learning_rate": 0.00014094262521346427,
|
| 1325 |
+
"loss": 0.2894,
|
| 1326 |
+
"step": 187
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 1.2920962199312716,
|
| 1330 |
+
"grad_norm": 0.10084875638276648,
|
| 1331 |
+
"learning_rate": 0.00014020826530603776,
|
| 1332 |
+
"loss": 0.2941,
|
| 1333 |
+
"step": 188
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.2989690721649485,
|
| 1337 |
+
"grad_norm": 0.10560762521845199,
|
| 1338 |
+
"learning_rate": 0.00013947130967031717,
|
| 1339 |
+
"loss": 0.2974,
|
| 1340 |
+
"step": 189
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.3058419243986255,
|
| 1344 |
+
"grad_norm": 0.1112230814008033,
|
| 1345 |
+
"learning_rate": 0.00013873180588200827,
|
| 1346 |
+
"loss": 0.3051,
|
| 1347 |
+
"step": 190
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 1.3127147766323024,
|
| 1351 |
+
"grad_norm": 0.10899405854558633,
|
| 1352 |
+
"learning_rate": 0.00013798980168131794,
|
| 1353 |
+
"loss": 0.3167,
|
| 1354 |
+
"step": 191
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 1.3195876288659794,
|
| 1358 |
+
"grad_norm": 0.10459030666604682,
|
| 1359 |
+
"learning_rate": 0.00013724534496987247,
|
| 1360 |
+
"loss": 0.2944,
|
| 1361 |
+
"step": 192
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 1.3264604810996563,
|
| 1365 |
+
"grad_norm": 0.10751960916175171,
|
| 1366 |
+
"learning_rate": 0.00013649848380762513,
|
| 1367 |
+
"loss": 0.2749,
|
| 1368 |
+
"step": 193
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 1.3333333333333333,
|
| 1372 |
+
"grad_norm": 0.10402996897559527,
|
| 1373 |
+
"learning_rate": 0.0001357492664097534,
|
| 1374 |
+
"loss": 0.2972,
|
| 1375 |
+
"step": 194
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.3402061855670104,
|
| 1379 |
+
"grad_norm": 0.10280900537193306,
|
| 1380 |
+
"learning_rate": 0.00013499774114354655,
|
| 1381 |
+
"loss": 0.3093,
|
| 1382 |
+
"step": 195
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.3470790378006874,
|
| 1386 |
+
"grad_norm": 0.10262166065058195,
|
| 1387 |
+
"learning_rate": 0.0001342439565252831,
|
| 1388 |
+
"loss": 0.3023,
|
| 1389 |
+
"step": 196
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.3539518900343643,
|
| 1393 |
+
"grad_norm": 0.11035322484519891,
|
| 1394 |
+
"learning_rate": 0.00013348796121709862,
|
| 1395 |
+
"loss": 0.283,
|
| 1396 |
+
"step": 197
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 1.3608247422680413,
|
| 1400 |
+
"grad_norm": 0.10183546777840774,
|
| 1401 |
+
"learning_rate": 0.0001327298040238446,
|
| 1402 |
+
"loss": 0.2895,
|
| 1403 |
+
"step": 198
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.3676975945017182,
|
| 1407 |
+
"grad_norm": 0.09846452794645016,
|
| 1408 |
+
"learning_rate": 0.00013196953388993726,
|
| 1409 |
+
"loss": 0.2775,
|
| 1410 |
+
"step": 199
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 1.3745704467353952,
|
| 1414 |
+
"grad_norm": 0.10082546961742879,
|
| 1415 |
+
"learning_rate": 0.00013120719989619833,
|
| 1416 |
+
"loss": 0.2927,
|
| 1417 |
+
"step": 200
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.3814432989690721,
|
| 1421 |
+
"grad_norm": 0.10338270125162533,
|
| 1422 |
+
"learning_rate": 0.00013044285125668614,
|
| 1423 |
+
"loss": 0.3022,
|
| 1424 |
+
"step": 201
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.388316151202749,
|
| 1428 |
+
"grad_norm": 0.10489686085507173,
|
| 1429 |
+
"learning_rate": 0.0001296765373155188,
|
| 1430 |
+
"loss": 0.2886,
|
| 1431 |
+
"step": 202
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 1.395189003436426,
|
| 1435 |
+
"grad_norm": 0.10826965142841156,
|
| 1436 |
+
"learning_rate": 0.00012890830754368855,
|
| 1437 |
+
"loss": 0.2839,
|
| 1438 |
+
"step": 203
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.402061855670103,
|
| 1442 |
+
"grad_norm": 0.1095657642687174,
|
| 1443 |
+
"learning_rate": 0.0001281382115358679,
|
| 1444 |
+
"loss": 0.2974,
|
| 1445 |
+
"step": 204
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 1.40893470790378,
|
| 1449 |
+
"grad_norm": 0.10290443785837475,
|
| 1450 |
+
"learning_rate": 0.0001273662990072083,
|
| 1451 |
+
"loss": 0.2916,
|
| 1452 |
+
"step": 205
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 1.4158075601374571,
|
| 1456 |
+
"grad_norm": 0.10488572942462458,
|
| 1457 |
+
"learning_rate": 0.00012659261979013043,
|
| 1458 |
+
"loss": 0.2896,
|
| 1459 |
+
"step": 206
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.422680412371134,
|
| 1463 |
+
"grad_norm": 0.0991511111540952,
|
| 1464 |
+
"learning_rate": 0.00012581722383110718,
|
| 1465 |
+
"loss": 0.2815,
|
| 1466 |
+
"step": 207
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.429553264604811,
|
| 1470 |
+
"grad_norm": 0.11084616014817153,
|
| 1471 |
+
"learning_rate": 0.00012504016118743935,
|
| 1472 |
+
"loss": 0.2857,
|
| 1473 |
+
"step": 208
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 1.436426116838488,
|
| 1477 |
+
"grad_norm": 0.10755359121093663,
|
| 1478 |
+
"learning_rate": 0.00012426148202402404,
|
| 1479 |
+
"loss": 0.2784,
|
| 1480 |
+
"step": 209
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.443298969072165,
|
| 1484 |
+
"grad_norm": 0.10341376660828379,
|
| 1485 |
+
"learning_rate": 0.00012348123661011601,
|
| 1486 |
+
"loss": 0.2947,
|
| 1487 |
+
"step": 210
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 1.4501718213058419,
|
| 1491 |
+
"grad_norm": 0.10964970839444545,
|
| 1492 |
+
"learning_rate": 0.00012269947531608276,
|
| 1493 |
+
"loss": 0.3056,
|
| 1494 |
+
"step": 211
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.4570446735395188,
|
| 1498 |
+
"grad_norm": 0.10241977399698814,
|
| 1499 |
+
"learning_rate": 0.00012191624861015254,
|
| 1500 |
+
"loss": 0.3102,
|
| 1501 |
+
"step": 212
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.463917525773196,
|
| 1505 |
+
"grad_norm": 0.10337164620385511,
|
| 1506 |
+
"learning_rate": 0.00012113160705515625,
|
| 1507 |
+
"loss": 0.3122,
|
| 1508 |
+
"step": 213
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.470790378006873,
|
| 1512 |
+
"grad_norm": 0.10002116602053045,
|
| 1513 |
+
"learning_rate": 0.0001203456013052634,
|
| 1514 |
+
"loss": 0.2905,
|
| 1515 |
+
"step": 214
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 1.47766323024055,
|
| 1519 |
+
"grad_norm": 0.1005828345016487,
|
| 1520 |
+
"learning_rate": 0.00011955828210271187,
|
| 1521 |
+
"loss": 0.2897,
|
| 1522 |
+
"step": 215
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.4845360824742269,
|
| 1526 |
+
"grad_norm": 0.10315190703705018,
|
| 1527 |
+
"learning_rate": 0.00011876970027453222,
|
| 1528 |
+
"loss": 0.2896,
|
| 1529 |
+
"step": 216
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 1.4914089347079038,
|
| 1533 |
+
"grad_norm": 0.10209266642811851,
|
| 1534 |
+
"learning_rate": 0.00011797990672926652,
|
| 1535 |
+
"loss": 0.2862,
|
| 1536 |
+
"step": 217
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 1.4982817869415808,
|
| 1540 |
+
"grad_norm": 0.10350382211989749,
|
| 1541 |
+
"learning_rate": 0.00011718895245368167,
|
| 1542 |
+
"loss": 0.2993,
|
| 1543 |
+
"step": 218
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.5051546391752577,
|
| 1547 |
+
"grad_norm": 0.09959069681120952,
|
| 1548 |
+
"learning_rate": 0.00011639688850947799,
|
| 1549 |
+
"loss": 0.279,
|
| 1550 |
+
"step": 219
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 1.5120274914089347,
|
| 1554 |
+
"grad_norm": 0.1017699606831777,
|
| 1555 |
+
"learning_rate": 0.00011560376602999272,
|
| 1556 |
+
"loss": 0.2924,
|
| 1557 |
+
"step": 220
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.5189003436426116,
|
| 1561 |
+
"grad_norm": 0.10199010748060407,
|
| 1562 |
+
"learning_rate": 0.00011480963621689905,
|
| 1563 |
+
"loss": 0.3007,
|
| 1564 |
+
"step": 221
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.5257731958762886,
|
| 1568 |
+
"grad_norm": 0.10438569087355797,
|
| 1569 |
+
"learning_rate": 0.00011401455033690076,
|
| 1570 |
+
"loss": 0.2828,
|
| 1571 |
+
"step": 222
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.5326460481099655,
|
| 1575 |
+
"grad_norm": 0.106590652554821,
|
| 1576 |
+
"learning_rate": 0.00011321855971842243,
|
| 1577 |
+
"loss": 0.299,
|
| 1578 |
+
"step": 223
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 1.5395189003436425,
|
| 1582 |
+
"grad_norm": 0.10067806688032917,
|
| 1583 |
+
"learning_rate": 0.00011242171574829599,
|
| 1584 |
+
"loss": 0.2897,
|
| 1585 |
+
"step": 224
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.5463917525773194,
|
| 1589 |
+
"grad_norm": 0.11078489209589217,
|
| 1590 |
+
"learning_rate": 0.00011162406986844323,
|
| 1591 |
+
"loss": 0.2791,
|
| 1592 |
+
"step": 225
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 1.5532646048109966,
|
| 1596 |
+
"grad_norm": 0.10748644329087202,
|
| 1597 |
+
"learning_rate": 0.00011082567357255484,
|
| 1598 |
+
"loss": 0.3004,
|
| 1599 |
+
"step": 226
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.5601374570446735,
|
| 1603 |
+
"grad_norm": 0.09232201373607851,
|
| 1604 |
+
"learning_rate": 0.00011002657840276627,
|
| 1605 |
+
"loss": 0.2647,
|
| 1606 |
+
"step": 227
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.5670103092783505,
|
| 1610 |
+
"grad_norm": 0.10419820524514652,
|
| 1611 |
+
"learning_rate": 0.00010922683594633021,
|
| 1612 |
+
"loss": 0.2897,
|
| 1613 |
+
"step": 228
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.5738831615120275,
|
| 1617 |
+
"grad_norm": 0.11387366232726741,
|
| 1618 |
+
"learning_rate": 0.00010842649783228624,
|
| 1619 |
+
"loss": 0.3071,
|
| 1620 |
+
"step": 229
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 1.5807560137457046,
|
| 1624 |
+
"grad_norm": 0.1112209148078402,
|
| 1625 |
+
"learning_rate": 0.00010762561572812788,
|
| 1626 |
+
"loss": 0.2791,
|
| 1627 |
+
"step": 230
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.5876288659793816,
|
| 1631 |
+
"grad_norm": 0.1027329268180551,
|
| 1632 |
+
"learning_rate": 0.0001068242413364671,
|
| 1633 |
+
"loss": 0.2962,
|
| 1634 |
+
"step": 231
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 1.5945017182130585,
|
| 1638 |
+
"grad_norm": 0.10377377307338428,
|
| 1639 |
+
"learning_rate": 0.00010602242639169648,
|
| 1640 |
+
"loss": 0.2905,
|
| 1641 |
+
"step": 232
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.6013745704467355,
|
| 1645 |
+
"grad_norm": 0.10333591660725541,
|
| 1646 |
+
"learning_rate": 0.0001052202226566494,
|
| 1647 |
+
"loss": 0.294,
|
| 1648 |
+
"step": 233
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.6082474226804124,
|
| 1652 |
+
"grad_norm": 0.10653830908113968,
|
| 1653 |
+
"learning_rate": 0.00010441768191925847,
|
| 1654 |
+
"loss": 0.3023,
|
| 1655 |
+
"step": 234
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.6151202749140894,
|
| 1659 |
+
"grad_norm": 0.10203788732177198,
|
| 1660 |
+
"learning_rate": 0.00010361485598921212,
|
| 1661 |
+
"loss": 0.2961,
|
| 1662 |
+
"step": 235
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.6219931271477663,
|
| 1666 |
+
"grad_norm": 0.105467450842672,
|
| 1667 |
+
"learning_rate": 0.00010281179669461005,
|
| 1668 |
+
"loss": 0.2992,
|
| 1669 |
+
"step": 236
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.6288659793814433,
|
| 1673 |
+
"grad_norm": 0.10232600415297333,
|
| 1674 |
+
"learning_rate": 0.00010200855587861724,
|
| 1675 |
+
"loss": 0.2894,
|
| 1676 |
+
"step": 237
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 1.6357388316151202,
|
| 1680 |
+
"grad_norm": 0.1000078228085237,
|
| 1681 |
+
"learning_rate": 0.0001012051853961172,
|
| 1682 |
+
"loss": 0.2924,
|
| 1683 |
+
"step": 238
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 1.6426116838487972,
|
| 1687 |
+
"grad_norm": 0.10443130272760445,
|
| 1688 |
+
"learning_rate": 0.00010040173711036431,
|
| 1689 |
+
"loss": 0.2987,
|
| 1690 |
+
"step": 239
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.6494845360824741,
|
| 1694 |
+
"grad_norm": 0.10490297286528327,
|
| 1695 |
+
"learning_rate": 9.959826288963571e-05,
|
| 1696 |
+
"loss": 0.2859,
|
| 1697 |
+
"step": 240
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.656357388316151,
|
| 1701 |
+
"grad_norm": 0.10350833328466466,
|
| 1702 |
+
"learning_rate": 9.879481460388282e-05,
|
| 1703 |
+
"loss": 0.2931,
|
| 1704 |
+
"step": 241
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 1.663230240549828,
|
| 1708 |
+
"grad_norm": 0.10225492666499335,
|
| 1709 |
+
"learning_rate": 9.799144412138275e-05,
|
| 1710 |
+
"loss": 0.2878,
|
| 1711 |
+
"step": 242
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 1.670103092783505,
|
| 1715 |
+
"grad_norm": 0.10508397146202163,
|
| 1716 |
+
"learning_rate": 9.718820330538998e-05,
|
| 1717 |
+
"loss": 0.3004,
|
| 1718 |
+
"step": 243
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 1.6769759450171822,
|
| 1722 |
+
"grad_norm": 0.10283137687490299,
|
| 1723 |
+
"learning_rate": 9.638514401078788e-05,
|
| 1724 |
+
"loss": 0.2714,
|
| 1725 |
+
"step": 244
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 1.6838487972508591,
|
| 1729 |
+
"grad_norm": 0.10397776989480932,
|
| 1730 |
+
"learning_rate": 9.558231808074156e-05,
|
| 1731 |
+
"loss": 0.2771,
|
| 1732 |
+
"step": 245
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 1.690721649484536,
|
| 1736 |
+
"grad_norm": 0.10092959577494873,
|
| 1737 |
+
"learning_rate": 9.477977734335061e-05,
|
| 1738 |
+
"loss": 0.2899,
|
| 1739 |
+
"step": 246
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.697594501718213,
|
| 1743 |
+
"grad_norm": 0.09948236816567041,
|
| 1744 |
+
"learning_rate": 9.397757360830353e-05,
|
| 1745 |
+
"loss": 0.2805,
|
| 1746 |
+
"step": 247
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 1.7044673539518902,
|
| 1750 |
+
"grad_norm": 0.10097655488635103,
|
| 1751 |
+
"learning_rate": 9.317575866353292e-05,
|
| 1752 |
+
"loss": 0.296,
|
| 1753 |
+
"step": 248
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.7113402061855671,
|
| 1757 |
+
"grad_norm": 0.09936217630195925,
|
| 1758 |
+
"learning_rate": 9.23743842718721e-05,
|
| 1759 |
+
"loss": 0.2973,
|
| 1760 |
+
"step": 249
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 1.718213058419244,
|
| 1764 |
+
"grad_norm": 0.09913343698008316,
|
| 1765 |
+
"learning_rate": 9.157350216771378e-05,
|
| 1766 |
+
"loss": 0.2899,
|
| 1767 |
+
"step": 250
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 1.725085910652921,
|
| 1771 |
+
"grad_norm": 0.09873019613022356,
|
| 1772 |
+
"learning_rate": 9.077316405366981e-05,
|
| 1773 |
+
"loss": 0.2827,
|
| 1774 |
+
"step": 251
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 1.731958762886598,
|
| 1778 |
+
"grad_norm": 0.09795703366595764,
|
| 1779 |
+
"learning_rate": 8.997342159723371e-05,
|
| 1780 |
+
"loss": 0.2833,
|
| 1781 |
+
"step": 252
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.738831615120275,
|
| 1785 |
+
"grad_norm": 0.0971884033549596,
|
| 1786 |
+
"learning_rate": 8.917432642744518e-05,
|
| 1787 |
+
"loss": 0.2813,
|
| 1788 |
+
"step": 253
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 1.745704467353952,
|
| 1792 |
+
"grad_norm": 0.10306201837901838,
|
| 1793 |
+
"learning_rate": 8.83759301315568e-05,
|
| 1794 |
+
"loss": 0.2924,
|
| 1795 |
+
"step": 254
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 1.7525773195876289,
|
| 1799 |
+
"grad_norm": 0.10374979571562502,
|
| 1800 |
+
"learning_rate": 8.757828425170404e-05,
|
| 1801 |
+
"loss": 0.2838,
|
| 1802 |
+
"step": 255
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 1.7594501718213058,
|
| 1806 |
+
"grad_norm": 0.1048687957670358,
|
| 1807 |
+
"learning_rate": 8.678144028157759e-05,
|
| 1808 |
+
"loss": 0.2886,
|
| 1809 |
+
"step": 256
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 1.7663230240549828,
|
| 1813 |
+
"grad_norm": 0.10256152550776314,
|
| 1814 |
+
"learning_rate": 8.598544966309925e-05,
|
| 1815 |
+
"loss": 0.2879,
|
| 1816 |
+
"step": 257
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 1.7731958762886597,
|
| 1820 |
+
"grad_norm": 0.10130883507858145,
|
| 1821 |
+
"learning_rate": 8.519036378310096e-05,
|
| 1822 |
+
"loss": 0.2909,
|
| 1823 |
+
"step": 258
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 1.7800687285223367,
|
| 1827 |
+
"grad_norm": 0.10467386983378474,
|
| 1828 |
+
"learning_rate": 8.43962339700073e-05,
|
| 1829 |
+
"loss": 0.2828,
|
| 1830 |
+
"step": 259
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 1.7869415807560136,
|
| 1834 |
+
"grad_norm": 0.09820302182298171,
|
| 1835 |
+
"learning_rate": 8.360311149052205e-05,
|
| 1836 |
+
"loss": 0.2844,
|
| 1837 |
+
"step": 260
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 1.7938144329896906,
|
| 1841 |
+
"grad_norm": 0.10102934299035139,
|
| 1842 |
+
"learning_rate": 8.281104754631835e-05,
|
| 1843 |
+
"loss": 0.2898,
|
| 1844 |
+
"step": 261
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 1.8006872852233677,
|
| 1848 |
+
"grad_norm": 0.0958354586875355,
|
| 1849 |
+
"learning_rate": 8.20200932707335e-05,
|
| 1850 |
+
"loss": 0.28,
|
| 1851 |
+
"step": 262
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 1.8075601374570447,
|
| 1855 |
+
"grad_norm": 0.09957495231488761,
|
| 1856 |
+
"learning_rate": 8.123029972546781e-05,
|
| 1857 |
+
"loss": 0.284,
|
| 1858 |
+
"step": 263
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 1.8144329896907216,
|
| 1862 |
+
"grad_norm": 0.10693182349883178,
|
| 1863 |
+
"learning_rate": 8.044171789728816e-05,
|
| 1864 |
+
"loss": 0.2892,
|
| 1865 |
+
"step": 264
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 1.8213058419243986,
|
| 1869 |
+
"grad_norm": 0.10646064359160178,
|
| 1870 |
+
"learning_rate": 7.965439869473664e-05,
|
| 1871 |
+
"loss": 0.2958,
|
| 1872 |
+
"step": 265
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 1.8281786941580758,
|
| 1876 |
+
"grad_norm": 0.10085406622127646,
|
| 1877 |
+
"learning_rate": 7.886839294484377e-05,
|
| 1878 |
+
"loss": 0.2818,
|
| 1879 |
+
"step": 266
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 1.8350515463917527,
|
| 1883 |
+
"grad_norm": 0.10500151086048896,
|
| 1884 |
+
"learning_rate": 7.808375138984745e-05,
|
| 1885 |
+
"loss": 0.2733,
|
| 1886 |
+
"step": 267
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 1.8419243986254297,
|
| 1890 |
+
"grad_norm": 0.1002635668475348,
|
| 1891 |
+
"learning_rate": 7.730052468391725e-05,
|
| 1892 |
+
"loss": 0.285,
|
| 1893 |
+
"step": 268
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 1.8487972508591066,
|
| 1897 |
+
"grad_norm": 0.09642728830623505,
|
| 1898 |
+
"learning_rate": 7.6518763389884e-05,
|
| 1899 |
+
"loss": 0.2728,
|
| 1900 |
+
"step": 269
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 1.8556701030927836,
|
| 1904 |
+
"grad_norm": 0.10268576835416232,
|
| 1905 |
+
"learning_rate": 7.573851797597602e-05,
|
| 1906 |
+
"loss": 0.2795,
|
| 1907 |
+
"step": 270
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 1.8625429553264605,
|
| 1911 |
+
"grad_norm": 0.10795408269996427,
|
| 1912 |
+
"learning_rate": 7.495983881256067e-05,
|
| 1913 |
+
"loss": 0.295,
|
| 1914 |
+
"step": 271
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 1.8694158075601375,
|
| 1918 |
+
"grad_norm": 0.10121463339523644,
|
| 1919 |
+
"learning_rate": 7.418277616889282e-05,
|
| 1920 |
+
"loss": 0.2787,
|
| 1921 |
+
"step": 272
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 1.8762886597938144,
|
| 1925 |
+
"grad_norm": 0.0982041211411212,
|
| 1926 |
+
"learning_rate": 7.340738020986961e-05,
|
| 1927 |
+
"loss": 0.2913,
|
| 1928 |
+
"step": 273
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 1.8831615120274914,
|
| 1932 |
+
"grad_norm": 0.09689663299295423,
|
| 1933 |
+
"learning_rate": 7.263370099279172e-05,
|
| 1934 |
+
"loss": 0.2678,
|
| 1935 |
+
"step": 274
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.8900343642611683,
|
| 1939 |
+
"grad_norm": 0.09890680768599139,
|
| 1940 |
+
"learning_rate": 7.186178846413214e-05,
|
| 1941 |
+
"loss": 0.2622,
|
| 1942 |
+
"step": 275
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 1.8969072164948453,
|
| 1946 |
+
"grad_norm": 0.10218531402065698,
|
| 1947 |
+
"learning_rate": 7.109169245631149e-05,
|
| 1948 |
+
"loss": 0.2948,
|
| 1949 |
+
"step": 276
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 1.9037800687285222,
|
| 1953 |
+
"grad_norm": 0.09970764344361568,
|
| 1954 |
+
"learning_rate": 7.032346268448118e-05,
|
| 1955 |
+
"loss": 0.2936,
|
| 1956 |
+
"step": 277
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 1.9106529209621992,
|
| 1960 |
+
"grad_norm": 0.10646182283668124,
|
| 1961 |
+
"learning_rate": 6.955714874331387e-05,
|
| 1962 |
+
"loss": 0.2841,
|
| 1963 |
+
"step": 278
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 1.9175257731958761,
|
| 1967 |
+
"grad_norm": 0.1005308262146528,
|
| 1968 |
+
"learning_rate": 6.87928001038017e-05,
|
| 1969 |
+
"loss": 0.2741,
|
| 1970 |
+
"step": 279
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 1.9243986254295533,
|
| 1974 |
+
"grad_norm": 0.09801297561347891,
|
| 1975 |
+
"learning_rate": 6.803046611006278e-05,
|
| 1976 |
+
"loss": 0.2812,
|
| 1977 |
+
"step": 280
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 1.9312714776632303,
|
| 1981 |
+
"grad_norm": 0.10072370231430908,
|
| 1982 |
+
"learning_rate": 6.727019597615545e-05,
|
| 1983 |
+
"loss": 0.2861,
|
| 1984 |
+
"step": 281
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 1.9381443298969072,
|
| 1988 |
+
"grad_norm": 0.10002866089004976,
|
| 1989 |
+
"learning_rate": 6.651203878290139e-05,
|
| 1990 |
+
"loss": 0.2747,
|
| 1991 |
+
"step": 282
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 1.9450171821305842,
|
| 1995 |
+
"grad_norm": 0.09520289893088804,
|
| 1996 |
+
"learning_rate": 6.575604347471695e-05,
|
| 1997 |
+
"loss": 0.275,
|
| 1998 |
+
"step": 283
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 1.9518900343642611,
|
| 2002 |
+
"grad_norm": 0.10328917600068767,
|
| 2003 |
+
"learning_rate": 6.500225885645346e-05,
|
| 2004 |
+
"loss": 0.2836,
|
| 2005 |
+
"step": 284
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 1.9587628865979383,
|
| 2009 |
+
"grad_norm": 0.09800249344342209,
|
| 2010 |
+
"learning_rate": 6.425073359024663e-05,
|
| 2011 |
+
"loss": 0.2904,
|
| 2012 |
+
"step": 285
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 1.9656357388316152,
|
| 2016 |
+
"grad_norm": 0.09952065394630734,
|
| 2017 |
+
"learning_rate": 6.350151619237488e-05,
|
| 2018 |
+
"loss": 0.2804,
|
| 2019 |
+
"step": 286
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 1.9725085910652922,
|
| 2023 |
+
"grad_norm": 0.10493099048201263,
|
| 2024 |
+
"learning_rate": 6.275465503012751e-05,
|
| 2025 |
+
"loss": 0.2818,
|
| 2026 |
+
"step": 287
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.9793814432989691,
|
| 2030 |
+
"grad_norm": 0.09722127539298636,
|
| 2031 |
+
"learning_rate": 6.201019831868208e-05,
|
| 2032 |
+
"loss": 0.2839,
|
| 2033 |
+
"step": 288
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 1.986254295532646,
|
| 2037 |
+
"grad_norm": 0.09609058787958422,
|
| 2038 |
+
"learning_rate": 6.126819411799175e-05,
|
| 2039 |
+
"loss": 0.2725,
|
| 2040 |
+
"step": 289
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 1.993127147766323,
|
| 2044 |
+
"grad_norm": 0.09821072360835045,
|
| 2045 |
+
"learning_rate": 6.052869032968285e-05,
|
| 2046 |
+
"loss": 0.2811,
|
| 2047 |
+
"step": 290
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 2.0,
|
| 2051 |
+
"grad_norm": 0.10288780904807039,
|
| 2052 |
+
"learning_rate": 5.979173469396227e-05,
|
| 2053 |
+
"loss": 0.2761,
|
| 2054 |
+
"step": 291
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 2.0,
|
| 2058 |
+
"eval_loss": 0.2949652373790741,
|
| 2059 |
+
"eval_runtime": 27.5209,
|
| 2060 |
+
"eval_samples_per_second": 35.5,
|
| 2061 |
+
"eval_steps_per_second": 1.126,
|
| 2062 |
+
"step": 291
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 2.006872852233677,
|
| 2066 |
+
"grad_norm": 0.097538336458536,
|
| 2067 |
+
"learning_rate": 5.905737478653572e-05,
|
| 2068 |
+
"loss": 0.2679,
|
| 2069 |
+
"step": 292
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 2.013745704467354,
|
| 2073 |
+
"grad_norm": 0.09988881705573123,
|
| 2074 |
+
"learning_rate": 5.83256580155362e-05,
|
| 2075 |
+
"loss": 0.2525,
|
| 2076 |
+
"step": 293
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 2.020618556701031,
|
| 2080 |
+
"grad_norm": 0.0964482303899947,
|
| 2081 |
+
"learning_rate": 5.7596631618463514e-05,
|
| 2082 |
+
"loss": 0.2425,
|
| 2083 |
+
"step": 294
|
| 2084 |
+
},
|
| 2085 |
+
{
|
| 2086 |
+
"epoch": 2.027491408934708,
|
| 2087 |
+
"grad_norm": 0.10864263032849888,
|
| 2088 |
+
"learning_rate": 5.687034265913485e-05,
|
| 2089 |
+
"loss": 0.27,
|
| 2090 |
+
"step": 295
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 2.0343642611683848,
|
| 2094 |
+
"grad_norm": 0.11734835455675438,
|
| 2095 |
+
"learning_rate": 5.614683802464631e-05,
|
| 2096 |
+
"loss": 0.2624,
|
| 2097 |
+
"step": 296
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 2.0412371134020617,
|
| 2101 |
+
"grad_norm": 0.11289186335789302,
|
| 2102 |
+
"learning_rate": 5.542616442234618e-05,
|
| 2103 |
+
"loss": 0.259,
|
| 2104 |
+
"step": 297
|
| 2105 |
+
},
|
| 2106 |
+
{
|
| 2107 |
+
"epoch": 2.0481099656357387,
|
| 2108 |
+
"grad_norm": 0.11224803307996767,
|
| 2109 |
+
"learning_rate": 5.470836837681954e-05,
|
| 2110 |
+
"loss": 0.2607,
|
| 2111 |
+
"step": 298
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 2.0549828178694156,
|
| 2115 |
+
"grad_norm": 0.10867092848715432,
|
| 2116 |
+
"learning_rate": 5.399349622688479e-05,
|
| 2117 |
+
"loss": 0.2703,
|
| 2118 |
+
"step": 299
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 2.0618556701030926,
|
| 2122 |
+
"grad_norm": 0.10624834548230039,
|
| 2123 |
+
"learning_rate": 5.32815941226022e-05,
|
| 2124 |
+
"loss": 0.2654,
|
| 2125 |
+
"step": 300
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 2.06872852233677,
|
| 2129 |
+
"grad_norm": 0.10758062626317398,
|
| 2130 |
+
"learning_rate": 5.2572708022294504e-05,
|
| 2131 |
+
"loss": 0.2698,
|
| 2132 |
+
"step": 301
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 2.075601374570447,
|
| 2136 |
+
"grad_norm": 0.1071415029446744,
|
| 2137 |
+
"learning_rate": 5.1866883689580056e-05,
|
| 2138 |
+
"loss": 0.2632,
|
| 2139 |
+
"step": 302
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 2.082474226804124,
|
| 2143 |
+
"grad_norm": 0.10125455614344606,
|
| 2144 |
+
"learning_rate": 5.116416669041843e-05,
|
| 2145 |
+
"loss": 0.2538,
|
| 2146 |
+
"step": 303
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 2.089347079037801,
|
| 2150 |
+
"grad_norm": 0.10522800373863611,
|
| 2151 |
+
"learning_rate": 5.046460239016879e-05,
|
| 2152 |
+
"loss": 0.2693,
|
| 2153 |
+
"step": 304
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 2.0962199312714778,
|
| 2157 |
+
"grad_norm": 0.10672396619233848,
|
| 2158 |
+
"learning_rate": 4.976823595066128e-05,
|
| 2159 |
+
"loss": 0.2575,
|
| 2160 |
+
"step": 305
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 2.1030927835051547,
|
| 2164 |
+
"grad_norm": 0.10982835933289055,
|
| 2165 |
+
"learning_rate": 4.907511232728145e-05,
|
| 2166 |
+
"loss": 0.2684,
|
| 2167 |
+
"step": 306
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 2.1099656357388317,
|
| 2171 |
+
"grad_norm": 0.10787287180280963,
|
| 2172 |
+
"learning_rate": 4.8385276266068146e-05,
|
| 2173 |
+
"loss": 0.258,
|
| 2174 |
+
"step": 307
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 2.1168384879725086,
|
| 2178 |
+
"grad_norm": 0.11168112585189982,
|
| 2179 |
+
"learning_rate": 4.7698772300824756e-05,
|
| 2180 |
+
"loss": 0.2665,
|
| 2181 |
+
"step": 308
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 2.1237113402061856,
|
| 2185 |
+
"grad_norm": 0.10887273811279173,
|
| 2186 |
+
"learning_rate": 4.7015644750244306e-05,
|
| 2187 |
+
"loss": 0.2609,
|
| 2188 |
+
"step": 309
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 2.1305841924398625,
|
| 2192 |
+
"grad_norm": 0.10561230206084926,
|
| 2193 |
+
"learning_rate": 4.6335937715048306e-05,
|
| 2194 |
+
"loss": 0.2569,
|
| 2195 |
+
"step": 310
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 2.1374570446735395,
|
| 2199 |
+
"grad_norm": 0.1085500386167761,
|
| 2200 |
+
"learning_rate": 4.565969507513981e-05,
|
| 2201 |
+
"loss": 0.2636,
|
| 2202 |
+
"step": 311
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 2.1443298969072164,
|
| 2206 |
+
"grad_norm": 0.1012240249829896,
|
| 2207 |
+
"learning_rate": 4.498696048677059e-05,
|
| 2208 |
+
"loss": 0.2547,
|
| 2209 |
+
"step": 312
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 2.1512027491408934,
|
| 2213 |
+
"grad_norm": 0.1030596147725605,
|
| 2214 |
+
"learning_rate": 4.4317777379722866e-05,
|
| 2215 |
+
"loss": 0.2531,
|
| 2216 |
+
"step": 313
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 2.1580756013745703,
|
| 2220 |
+
"grad_norm": 0.10613673774187961,
|
| 2221 |
+
"learning_rate": 4.365218895450558e-05,
|
| 2222 |
+
"loss": 0.2608,
|
| 2223 |
+
"step": 314
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 2.1649484536082473,
|
| 2227 |
+
"grad_norm": 0.10496672072120719,
|
| 2228 |
+
"learning_rate": 4.29902381795655e-05,
|
| 2229 |
+
"loss": 0.2711,
|
| 2230 |
+
"step": 315
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 2.1718213058419242,
|
| 2234 |
+
"grad_norm": 0.1017243256571039,
|
| 2235 |
+
"learning_rate": 4.2331967788513295e-05,
|
| 2236 |
+
"loss": 0.2506,
|
| 2237 |
+
"step": 316
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 2.178694158075601,
|
| 2241 |
+
"grad_norm": 0.1083774959415722,
|
| 2242 |
+
"learning_rate": 4.167742027736482e-05,
|
| 2243 |
+
"loss": 0.2706,
|
| 2244 |
+
"step": 317
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 2.1855670103092786,
|
| 2248 |
+
"grad_norm": 0.10375985752479969,
|
| 2249 |
+
"learning_rate": 4.102663790179764e-05,
|
| 2250 |
+
"loss": 0.2446,
|
| 2251 |
+
"step": 318
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 2.1924398625429555,
|
| 2255 |
+
"grad_norm": 0.10866264016891114,
|
| 2256 |
+
"learning_rate": 4.037966267442315e-05,
|
| 2257 |
+
"loss": 0.2782,
|
| 2258 |
+
"step": 319
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 2.1993127147766325,
|
| 2262 |
+
"grad_norm": 0.10829606781249333,
|
| 2263 |
+
"learning_rate": 3.973653636207437e-05,
|
| 2264 |
+
"loss": 0.2628,
|
| 2265 |
+
"step": 320
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 2.2061855670103094,
|
| 2269 |
+
"grad_norm": 0.11273324743023601,
|
| 2270 |
+
"learning_rate": 3.909730048310962e-05,
|
| 2271 |
+
"loss": 0.2661,
|
| 2272 |
+
"step": 321
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 2.2130584192439864,
|
| 2276 |
+
"grad_norm": 0.10783971891161509,
|
| 2277 |
+
"learning_rate": 3.846199630473216e-05,
|
| 2278 |
+
"loss": 0.257,
|
| 2279 |
+
"step": 322
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 2.2199312714776633,
|
| 2283 |
+
"grad_norm": 0.10744882559462787,
|
| 2284 |
+
"learning_rate": 3.7830664840326145e-05,
|
| 2285 |
+
"loss": 0.2431,
|
| 2286 |
+
"step": 323
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 2.2268041237113403,
|
| 2290 |
+
"grad_norm": 0.10847966130290622,
|
| 2291 |
+
"learning_rate": 3.720334684680889e-05,
|
| 2292 |
+
"loss": 0.2645,
|
| 2293 |
+
"step": 324
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 2.2336769759450172,
|
| 2297 |
+
"grad_norm": 0.10687099898167607,
|
| 2298 |
+
"learning_rate": 3.6580082821999786e-05,
|
| 2299 |
+
"loss": 0.2558,
|
| 2300 |
+
"step": 325
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 2.240549828178694,
|
| 2304 |
+
"grad_norm": 0.10744277727601985,
|
| 2305 |
+
"learning_rate": 3.596091300200578e-05,
|
| 2306 |
+
"loss": 0.2648,
|
| 2307 |
+
"step": 326
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 2.247422680412371,
|
| 2311 |
+
"grad_norm": 0.1075527766804997,
|
| 2312 |
+
"learning_rate": 3.534587735862391e-05,
|
| 2313 |
+
"loss": 0.2662,
|
| 2314 |
+
"step": 327
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 2.254295532646048,
|
| 2318 |
+
"grad_norm": 0.10693314058206296,
|
| 2319 |
+
"learning_rate": 3.473501559676088e-05,
|
| 2320 |
+
"loss": 0.2691,
|
| 2321 |
+
"step": 328
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 2.261168384879725,
|
| 2325 |
+
"grad_norm": 0.10977366989631175,
|
| 2326 |
+
"learning_rate": 3.4128367151869714e-05,
|
| 2327 |
+
"loss": 0.2721,
|
| 2328 |
+
"step": 329
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 2.268041237113402,
|
| 2332 |
+
"grad_norm": 0.10791580217848652,
|
| 2333 |
+
"learning_rate": 3.352597118740404e-05,
|
| 2334 |
+
"loss": 0.2574,
|
| 2335 |
+
"step": 330
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 2.274914089347079,
|
| 2339 |
+
"grad_norm": 0.1117741340523688,
|
| 2340 |
+
"learning_rate": 3.292786659228973e-05,
|
| 2341 |
+
"loss": 0.2701,
|
| 2342 |
+
"step": 331
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 2.281786941580756,
|
| 2346 |
+
"grad_norm": 0.10832840646361684,
|
| 2347 |
+
"learning_rate": 3.233409197841437e-05,
|
| 2348 |
+
"loss": 0.2533,
|
| 2349 |
+
"step": 332
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 2.288659793814433,
|
| 2353 |
+
"grad_norm": 0.11021992972158987,
|
| 2354 |
+
"learning_rate": 3.174468567813461e-05,
|
| 2355 |
+
"loss": 0.2506,
|
| 2356 |
+
"step": 333
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 2.29553264604811,
|
| 2360 |
+
"grad_norm": 0.11244737426877117,
|
| 2361 |
+
"learning_rate": 3.115968574180149e-05,
|
| 2362 |
+
"loss": 0.2741,
|
| 2363 |
+
"step": 334
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 2.3024054982817868,
|
| 2367 |
+
"grad_norm": 0.10439400399730109,
|
| 2368 |
+
"learning_rate": 3.0579129935304066e-05,
|
| 2369 |
+
"loss": 0.2423,
|
| 2370 |
+
"step": 335
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 2.3092783505154637,
|
| 2374 |
+
"grad_norm": 0.11154784823612464,
|
| 2375 |
+
"learning_rate": 3.0003055737631403e-05,
|
| 2376 |
+
"loss": 0.2606,
|
| 2377 |
+
"step": 336
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 2.3161512027491407,
|
| 2381 |
+
"grad_norm": 0.11618867786187727,
|
| 2382 |
+
"learning_rate": 2.9431500338452832e-05,
|
| 2383 |
+
"loss": 0.2705,
|
| 2384 |
+
"step": 337
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 2.323024054982818,
|
| 2388 |
+
"grad_norm": 0.10953255344278685,
|
| 2389 |
+
"learning_rate": 2.886450063571735e-05,
|
| 2390 |
+
"loss": 0.2597,
|
| 2391 |
+
"step": 338
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 2.329896907216495,
|
| 2395 |
+
"grad_norm": 0.10670783300766909,
|
| 2396 |
+
"learning_rate": 2.8302093233271453e-05,
|
| 2397 |
+
"loss": 0.2451,
|
| 2398 |
+
"step": 339
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 2.336769759450172,
|
| 2402 |
+
"grad_norm": 0.10795627839159605,
|
| 2403 |
+
"learning_rate": 2.7744314438496088e-05,
|
| 2404 |
+
"loss": 0.2571,
|
| 2405 |
+
"step": 340
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.343642611683849,
|
| 2409 |
+
"grad_norm": 0.10876660642511277,
|
| 2410 |
+
"learning_rate": 2.7191200259962934e-05,
|
| 2411 |
+
"loss": 0.2646,
|
| 2412 |
+
"step": 341
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 2.350515463917526,
|
| 2416 |
+
"grad_norm": 0.10862804831060525,
|
| 2417 |
+
"learning_rate": 2.6642786405109475e-05,
|
| 2418 |
+
"loss": 0.2679,
|
| 2419 |
+
"step": 342
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 2.357388316151203,
|
| 2423 |
+
"grad_norm": 0.1068062333293375,
|
| 2424 |
+
"learning_rate": 2.6099108277934103e-05,
|
| 2425 |
+
"loss": 0.2676,
|
| 2426 |
+
"step": 343
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 2.3642611683848798,
|
| 2430 |
+
"grad_norm": 0.10420269958200333,
|
| 2431 |
+
"learning_rate": 2.556020097671046e-05,
|
| 2432 |
+
"loss": 0.2669,
|
| 2433 |
+
"step": 344
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 2.3711340206185567,
|
| 2437 |
+
"grad_norm": 0.10951228936587228,
|
| 2438 |
+
"learning_rate": 2.5026099291721516e-05,
|
| 2439 |
+
"loss": 0.2519,
|
| 2440 |
+
"step": 345
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 2.3780068728522337,
|
| 2444 |
+
"grad_norm": 0.11129513461474132,
|
| 2445 |
+
"learning_rate": 2.449683770301382e-05,
|
| 2446 |
+
"loss": 0.2571,
|
| 2447 |
+
"step": 346
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.3848797250859106,
|
| 2451 |
+
"grad_norm": 0.10763529170641503,
|
| 2452 |
+
"learning_rate": 2.397245037817125e-05,
|
| 2453 |
+
"loss": 0.2567,
|
| 2454 |
+
"step": 347
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 2.3917525773195876,
|
| 2458 |
+
"grad_norm": 0.10496278836242602,
|
| 2459 |
+
"learning_rate": 2.345297117010954e-05,
|
| 2460 |
+
"loss": 0.2459,
|
| 2461 |
+
"step": 348
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 2.3986254295532645,
|
| 2465 |
+
"grad_norm": 0.11015993294501161,
|
| 2466 |
+
"learning_rate": 2.2938433614890697e-05,
|
| 2467 |
+
"loss": 0.2702,
|
| 2468 |
+
"step": 349
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 2.4054982817869415,
|
| 2472 |
+
"grad_norm": 0.10738471004067172,
|
| 2473 |
+
"learning_rate": 2.242887092955801e-05,
|
| 2474 |
+
"loss": 0.2585,
|
| 2475 |
+
"step": 350
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 2.4123711340206184,
|
| 2479 |
+
"grad_norm": 0.11218254036958039,
|
| 2480 |
+
"learning_rate": 2.1924316009991787e-05,
|
| 2481 |
+
"loss": 0.2607,
|
| 2482 |
+
"step": 351
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 2.4192439862542954,
|
| 2486 |
+
"grad_norm": 0.10844975577221563,
|
| 2487 |
+
"learning_rate": 2.1424801428785447e-05,
|
| 2488 |
+
"loss": 0.2561,
|
| 2489 |
+
"step": 352
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 2.4261168384879723,
|
| 2493 |
+
"grad_norm": 0.10599271281183413,
|
| 2494 |
+
"learning_rate": 2.0930359433142932e-05,
|
| 2495 |
+
"loss": 0.2501,
|
| 2496 |
+
"step": 353
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 2.4329896907216497,
|
| 2500 |
+
"grad_norm": 0.11017719350663994,
|
| 2501 |
+
"learning_rate": 2.0441021942796944e-05,
|
| 2502 |
+
"loss": 0.2618,
|
| 2503 |
+
"step": 354
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 2.4398625429553267,
|
| 2507 |
+
"grad_norm": 0.10668690539385935,
|
| 2508 |
+
"learning_rate": 1.995682054794803e-05,
|
| 2509 |
+
"loss": 0.264,
|
| 2510 |
+
"step": 355
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 2.4467353951890036,
|
| 2514 |
+
"grad_norm": 0.10450005811019492,
|
| 2515 |
+
"learning_rate": 1.9477786507225616e-05,
|
| 2516 |
+
"loss": 0.2553,
|
| 2517 |
+
"step": 356
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 2.4536082474226806,
|
| 2521 |
+
"grad_norm": 0.11220780349903246,
|
| 2522 |
+
"learning_rate": 1.900395074566962e-05,
|
| 2523 |
+
"loss": 0.2654,
|
| 2524 |
+
"step": 357
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 2.4604810996563575,
|
| 2528 |
+
"grad_norm": 0.108477409950911,
|
| 2529 |
+
"learning_rate": 1.8535343852734332e-05,
|
| 2530 |
+
"loss": 0.2569,
|
| 2531 |
+
"step": 358
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 2.4673539518900345,
|
| 2535 |
+
"grad_norm": 0.10460682785615502,
|
| 2536 |
+
"learning_rate": 1.8071996080313602e-05,
|
| 2537 |
+
"loss": 0.2535,
|
| 2538 |
+
"step": 359
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 2.4742268041237114,
|
| 2542 |
+
"grad_norm": 0.1076833252048245,
|
| 2543 |
+
"learning_rate": 1.76139373407876e-05,
|
| 2544 |
+
"loss": 0.2651,
|
| 2545 |
+
"step": 360
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 2.4810996563573884,
|
| 2549 |
+
"grad_norm": 0.10826152402956611,
|
| 2550 |
+
"learning_rate": 1.7161197205092216e-05,
|
| 2551 |
+
"loss": 0.2676,
|
| 2552 |
+
"step": 361
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 2.4879725085910653,
|
| 2556 |
+
"grad_norm": 0.10750093502253717,
|
| 2557 |
+
"learning_rate": 1.6713804900809582e-05,
|
| 2558 |
+
"loss": 0.2559,
|
| 2559 |
+
"step": 362
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 2.4948453608247423,
|
| 2563 |
+
"grad_norm": 0.1038001972682153,
|
| 2564 |
+
"learning_rate": 1.6271789310281517e-05,
|
| 2565 |
+
"loss": 0.2423,
|
| 2566 |
+
"step": 363
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 2.5017182130584192,
|
| 2570 |
+
"grad_norm": 0.10672842152395061,
|
| 2571 |
+
"learning_rate": 1.583517896874498e-05,
|
| 2572 |
+
"loss": 0.2593,
|
| 2573 |
+
"step": 364
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 2.508591065292096,
|
| 2577 |
+
"grad_norm": 0.10761120219562789,
|
| 2578 |
+
"learning_rate": 1.540400206248963e-05,
|
| 2579 |
+
"loss": 0.2595,
|
| 2580 |
+
"step": 365
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 2.515463917525773,
|
| 2584 |
+
"grad_norm": 0.10816060382518869,
|
| 2585 |
+
"learning_rate": 1.4978286427038601e-05,
|
| 2586 |
+
"loss": 0.2539,
|
| 2587 |
+
"step": 366
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 2.52233676975945,
|
| 2591 |
+
"grad_norm": 0.10455085413886135,
|
| 2592 |
+
"learning_rate": 1.4558059545351143e-05,
|
| 2593 |
+
"loss": 0.2564,
|
| 2594 |
+
"step": 367
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 2.529209621993127,
|
| 2598 |
+
"grad_norm": 0.10969390724809498,
|
| 2599 |
+
"learning_rate": 1.4143348546048707e-05,
|
| 2600 |
+
"loss": 0.2552,
|
| 2601 |
+
"step": 368
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 2.536082474226804,
|
| 2605 |
+
"grad_norm": 0.11070585821467833,
|
| 2606 |
+
"learning_rate": 1.3734180201663439e-05,
|
| 2607 |
+
"loss": 0.2561,
|
| 2608 |
+
"step": 369
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 2.542955326460481,
|
| 2612 |
+
"grad_norm": 0.11065650967955783,
|
| 2613 |
+
"learning_rate": 1.3330580926909763e-05,
|
| 2614 |
+
"loss": 0.2552,
|
| 2615 |
+
"step": 370
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 2.549828178694158,
|
| 2619 |
+
"grad_norm": 0.10753458918108932,
|
| 2620 |
+
"learning_rate": 1.2932576776979377e-05,
|
| 2621 |
+
"loss": 0.2599,
|
| 2622 |
+
"step": 371
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 2.556701030927835,
|
| 2626 |
+
"grad_norm": 0.11215812047440889,
|
| 2627 |
+
"learning_rate": 1.2540193445858883e-05,
|
| 2628 |
+
"loss": 0.26,
|
| 2629 |
+
"step": 372
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 2.563573883161512,
|
| 2633 |
+
"grad_norm": 0.11029227457181179,
|
| 2634 |
+
"learning_rate": 1.2153456264671337e-05,
|
| 2635 |
+
"loss": 0.2529,
|
| 2636 |
+
"step": 373
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 2.5704467353951888,
|
| 2640 |
+
"grad_norm": 0.1095036155268631,
|
| 2641 |
+
"learning_rate": 1.1772390200040817e-05,
|
| 2642 |
+
"loss": 0.2503,
|
| 2643 |
+
"step": 374
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 2.5773195876288657,
|
| 2647 |
+
"grad_norm": 0.10487179029197675,
|
| 2648 |
+
"learning_rate": 1.139701985248055e-05,
|
| 2649 |
+
"loss": 0.2421,
|
| 2650 |
+
"step": 375
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 2.584192439862543,
|
| 2654 |
+
"grad_norm": 0.10299175954283599,
|
| 2655 |
+
"learning_rate": 1.1027369454805058e-05,
|
| 2656 |
+
"loss": 0.2502,
|
| 2657 |
+
"step": 376
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 2.59106529209622,
|
| 2661 |
+
"grad_norm": 0.10855354801856847,
|
| 2662 |
+
"learning_rate": 1.0663462870565411e-05,
|
| 2663 |
+
"loss": 0.2608,
|
| 2664 |
+
"step": 377
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 2.597938144329897,
|
| 2668 |
+
"grad_norm": 0.10315061729781284,
|
| 2669 |
+
"learning_rate": 1.0305323592509009e-05,
|
| 2670 |
+
"loss": 0.2528,
|
| 2671 |
+
"step": 378
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 2.604810996563574,
|
| 2675 |
+
"grad_norm": 0.10626039326093263,
|
| 2676 |
+
"learning_rate": 9.952974741062703e-06,
|
| 2677 |
+
"loss": 0.264,
|
| 2678 |
+
"step": 379
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 2.611683848797251,
|
| 2682 |
+
"grad_norm": 0.10745271675212126,
|
| 2683 |
+
"learning_rate": 9.606439062840256e-06,
|
| 2684 |
+
"loss": 0.2594,
|
| 2685 |
+
"step": 380
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 2.618556701030928,
|
| 2689 |
+
"grad_norm": 0.10974406713413064,
|
| 2690 |
+
"learning_rate": 9.265738929174051e-06,
|
| 2691 |
+
"loss": 0.2582,
|
| 2692 |
+
"step": 381
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 2.625429553264605,
|
| 2696 |
+
"grad_norm": 0.10537263197622973,
|
| 2697 |
+
"learning_rate": 8.93089633467058e-06,
|
| 2698 |
+
"loss": 0.2513,
|
| 2699 |
+
"step": 382
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 2.6323024054982818,
|
| 2703 |
+
"grad_norm": 0.10396472858049301,
|
| 2704 |
+
"learning_rate": 8.601932895790877e-06,
|
| 2705 |
+
"loss": 0.2583,
|
| 2706 |
+
"step": 383
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 2.6391752577319587,
|
| 2710 |
+
"grad_norm": 0.10276278086397593,
|
| 2711 |
+
"learning_rate": 8.278869849454718e-06,
|
| 2712 |
+
"loss": 0.2529,
|
| 2713 |
+
"step": 384
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 2.6460481099656357,
|
| 2717 |
+
"grad_norm": 0.10899885081403418,
|
| 2718 |
+
"learning_rate": 7.961728051669737e-06,
|
| 2719 |
+
"loss": 0.2637,
|
| 2720 |
+
"step": 385
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 2.6529209621993126,
|
| 2724 |
+
"grad_norm": 0.10582666731745977,
|
| 2725 |
+
"learning_rate": 7.650527976185173e-06,
|
| 2726 |
+
"loss": 0.2535,
|
| 2727 |
+
"step": 386
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 2.6597938144329896,
|
| 2731 |
+
"grad_norm": 0.10725397582281598,
|
| 2732 |
+
"learning_rate": 7.3452897131698564e-06,
|
| 2733 |
+
"loss": 0.2603,
|
| 2734 |
+
"step": 387
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 2.6666666666666665,
|
| 2738 |
+
"grad_norm": 0.10749509280278831,
|
| 2739 |
+
"learning_rate": 7.046032967915483e-06,
|
| 2740 |
+
"loss": 0.2489,
|
| 2741 |
+
"step": 388
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 2.673539518900344,
|
| 2745 |
+
"grad_norm": 0.1087201328812968,
|
| 2746 |
+
"learning_rate": 6.75277705956443e-06,
|
| 2747 |
+
"loss": 0.2604,
|
| 2748 |
+
"step": 389
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 2.680412371134021,
|
| 2752 |
+
"grad_norm": 0.10783575017518213,
|
| 2753 |
+
"learning_rate": 6.465540919862456e-06,
|
| 2754 |
+
"loss": 0.2557,
|
| 2755 |
+
"step": 390
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 2.687285223367698,
|
| 2759 |
+
"grad_norm": 0.10758942906584705,
|
| 2760 |
+
"learning_rate": 6.184343091936751e-06,
|
| 2761 |
+
"loss": 0.2526,
|
| 2762 |
+
"step": 391
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 2.6941580756013748,
|
| 2766 |
+
"grad_norm": 0.10528296367950161,
|
| 2767 |
+
"learning_rate": 5.909201729098579e-06,
|
| 2768 |
+
"loss": 0.2487,
|
| 2769 |
+
"step": 392
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 2.7010309278350517,
|
| 2773 |
+
"grad_norm": 0.11256201944946768,
|
| 2774 |
+
"learning_rate": 5.640134593671598e-06,
|
| 2775 |
+
"loss": 0.2714,
|
| 2776 |
+
"step": 393
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 2.7079037800687287,
|
| 2780 |
+
"grad_norm": 0.10863241059875908,
|
| 2781 |
+
"learning_rate": 5.3771590558450265e-06,
|
| 2782 |
+
"loss": 0.2574,
|
| 2783 |
+
"step": 394
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 2.7147766323024056,
|
| 2787 |
+
"grad_norm": 0.10549349184272533,
|
| 2788 |
+
"learning_rate": 5.12029209255227e-06,
|
| 2789 |
+
"loss": 0.2554,
|
| 2790 |
+
"step": 395
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 2.7216494845360826,
|
| 2794 |
+
"grad_norm": 0.10847174602209737,
|
| 2795 |
+
"learning_rate": 4.869550286375091e-06,
|
| 2796 |
+
"loss": 0.268,
|
| 2797 |
+
"step": 396
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 2.7285223367697595,
|
| 2801 |
+
"grad_norm": 0.10488356613015705,
|
| 2802 |
+
"learning_rate": 4.624949824472858e-06,
|
| 2803 |
+
"loss": 0.2671,
|
| 2804 |
+
"step": 397
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 2.7353951890034365,
|
| 2808 |
+
"grad_norm": 0.10475860935600055,
|
| 2809 |
+
"learning_rate": 4.386506497537757e-06,
|
| 2810 |
+
"loss": 0.2514,
|
| 2811 |
+
"step": 398
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 2.7422680412371134,
|
| 2815 |
+
"grad_norm": 0.10597713121898167,
|
| 2816 |
+
"learning_rate": 4.154235698775277e-06,
|
| 2817 |
+
"loss": 0.2597,
|
| 2818 |
+
"step": 399
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 2.7491408934707904,
|
| 2822 |
+
"grad_norm": 0.10595800588015507,
|
| 2823 |
+
"learning_rate": 3.928152422910491e-06,
|
| 2824 |
+
"loss": 0.2602,
|
| 2825 |
+
"step": 400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 2.7560137457044673,
|
| 2829 |
+
"grad_norm": 0.11154644583398933,
|
| 2830 |
+
"learning_rate": 3.7082712652200867e-06,
|
| 2831 |
+
"loss": 0.264,
|
| 2832 |
+
"step": 401
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 2.7628865979381443,
|
| 2836 |
+
"grad_norm": 0.10694150624625907,
|
| 2837 |
+
"learning_rate": 3.4946064205899965e-06,
|
| 2838 |
+
"loss": 0.25,
|
| 2839 |
+
"step": 402
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 2.7697594501718212,
|
| 2843 |
+
"grad_norm": 0.10318772901644747,
|
| 2844 |
+
"learning_rate": 3.287171682599255e-06,
|
| 2845 |
+
"loss": 0.2493,
|
| 2846 |
+
"step": 403
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 2.776632302405498,
|
| 2850 |
+
"grad_norm": 0.1061914310952031,
|
| 2851 |
+
"learning_rate": 3.085980442629288e-06,
|
| 2852 |
+
"loss": 0.2591,
|
| 2853 |
+
"step": 404
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 2.783505154639175,
|
| 2857 |
+
"grad_norm": 0.10446042630271005,
|
| 2858 |
+
"learning_rate": 2.8910456889995498e-06,
|
| 2859 |
+
"loss": 0.254,
|
| 2860 |
+
"step": 405
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"epoch": 2.790378006872852,
|
| 2864 |
+
"grad_norm": 0.10518845803888338,
|
| 2865 |
+
"learning_rate": 2.7023800061289907e-06,
|
| 2866 |
+
"loss": 0.2533,
|
| 2867 |
+
"step": 406
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 2.797250859106529,
|
| 2871 |
+
"grad_norm": 0.10867396165520546,
|
| 2872 |
+
"learning_rate": 2.5199955737236104e-06,
|
| 2873 |
+
"loss": 0.2724,
|
| 2874 |
+
"step": 407
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 2.804123711340206,
|
| 2878 |
+
"grad_norm": 0.10694114528768675,
|
| 2879 |
+
"learning_rate": 2.3439041659902407e-06,
|
| 2880 |
+
"loss": 0.2564,
|
| 2881 |
+
"step": 408
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 2.810996563573883,
|
| 2885 |
+
"grad_norm": 0.10518421192748285,
|
| 2886 |
+
"learning_rate": 2.174117150876398e-06,
|
| 2887 |
+
"loss": 0.2609,
|
| 2888 |
+
"step": 409
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 2.81786941580756,
|
| 2892 |
+
"grad_norm": 0.10729647762632918,
|
| 2893 |
+
"learning_rate": 2.010645489336382e-06,
|
| 2894 |
+
"loss": 0.2587,
|
| 2895 |
+
"step": 410
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 2.824742268041237,
|
| 2899 |
+
"grad_norm": 0.10754314349393188,
|
| 2900 |
+
"learning_rate": 1.8534997346237093e-06,
|
| 2901 |
+
"loss": 0.2632,
|
| 2902 |
+
"step": 411
|
| 2903 |
+
},
|
| 2904 |
+
{
|
| 2905 |
+
"epoch": 2.8316151202749142,
|
| 2906 |
+
"grad_norm": 0.10802289125290783,
|
| 2907 |
+
"learning_rate": 1.7026900316098215e-06,
|
| 2908 |
+
"loss": 0.2566,
|
| 2909 |
+
"step": 412
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 2.838487972508591,
|
| 2913 |
+
"grad_norm": 0.10931075718242897,
|
| 2914 |
+
"learning_rate": 1.5582261161291245e-06,
|
| 2915 |
+
"loss": 0.2626,
|
| 2916 |
+
"step": 413
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 2.845360824742268,
|
| 2920 |
+
"grad_norm": 0.10501355229769045,
|
| 2921 |
+
"learning_rate": 1.4201173143504888e-06,
|
| 2922 |
+
"loss": 0.2621,
|
| 2923 |
+
"step": 414
|
| 2924 |
+
},
|
| 2925 |
+
{
|
| 2926 |
+
"epoch": 2.852233676975945,
|
| 2927 |
+
"grad_norm": 0.10753436065183695,
|
| 2928 |
+
"learning_rate": 1.2883725421752201e-06,
|
| 2929 |
+
"loss": 0.2475,
|
| 2930 |
+
"step": 415
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 2.859106529209622,
|
| 2934 |
+
"grad_norm": 0.10463768922989454,
|
| 2935 |
+
"learning_rate": 1.1630003046614323e-06,
|
| 2936 |
+
"loss": 0.2409,
|
| 2937 |
+
"step": 416
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 2.865979381443299,
|
| 2941 |
+
"grad_norm": 0.10543898222201153,
|
| 2942 |
+
"learning_rate": 1.0440086954749517e-06,
|
| 2943 |
+
"loss": 0.2505,
|
| 2944 |
+
"step": 417
|
| 2945 |
+
},
|
| 2946 |
+
{
|
| 2947 |
+
"epoch": 2.872852233676976,
|
| 2948 |
+
"grad_norm": 0.10892369128650758,
|
| 2949 |
+
"learning_rate": 9.314053963669245e-07,
|
| 2950 |
+
"loss": 0.2456,
|
| 2951 |
+
"step": 418
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 2.879725085910653,
|
| 2955 |
+
"grad_norm": 0.1062255349453702,
|
| 2956 |
+
"learning_rate": 8.251976766777913e-07,
|
| 2957 |
+
"loss": 0.2649,
|
| 2958 |
+
"step": 419
|
| 2959 |
+
},
|
| 2960 |
+
{
|
| 2961 |
+
"epoch": 2.88659793814433,
|
| 2962 |
+
"grad_norm": 0.10869941184854195,
|
| 2963 |
+
"learning_rate": 7.253923928680406e-07,
|
| 2964 |
+
"loss": 0.2559,
|
| 2965 |
+
"step": 420
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 2.893470790378007,
|
| 2969 |
+
"grad_norm": 0.1109190039810706,
|
| 2970 |
+
"learning_rate": 6.319959880756177e-07,
|
| 2971 |
+
"loss": 0.2529,
|
| 2972 |
+
"step": 421
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 2.9003436426116838,
|
| 2976 |
+
"grad_norm": 0.10656654333503314,
|
| 2977 |
+
"learning_rate": 5.450144916999134e-07,
|
| 2978 |
+
"loss": 0.2566,
|
| 2979 |
+
"step": 422
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 2.9072164948453607,
|
| 2983 |
+
"grad_norm": 0.10681708087523223,
|
| 2984 |
+
"learning_rate": 4.644535190125421e-07,
|
| 2985 |
+
"loss": 0.2526,
|
| 2986 |
+
"step": 423
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 2.9140893470790377,
|
| 2990 |
+
"grad_norm": 0.11101267026333346,
|
| 2991 |
+
"learning_rate": 3.903182707948649e-07,
|
| 2992 |
+
"loss": 0.2545,
|
| 2993 |
+
"step": 424
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 2.9209621993127146,
|
| 2997 |
+
"grad_norm": 0.10449251320574354,
|
| 2998 |
+
"learning_rate": 3.2261353300219176e-07,
|
| 2999 |
+
"loss": 0.2397,
|
| 3000 |
+
"step": 425
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 2.927835051546392,
|
| 3004 |
+
"grad_norm": 0.10614426653162999,
|
| 3005 |
+
"learning_rate": 2.613436764548505e-07,
|
| 3006 |
+
"loss": 0.2451,
|
| 3007 |
+
"step": 426
|
| 3008 |
+
},
|
| 3009 |
+
{
|
| 3010 |
+
"epoch": 2.934707903780069,
|
| 3011 |
+
"grad_norm": 0.10344408619271417,
|
| 3012 |
+
"learning_rate": 2.0651265655603492e-07,
|
| 3013 |
+
"loss": 0.231,
|
| 3014 |
+
"step": 427
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 2.941580756013746,
|
| 3018 |
+
"grad_norm": 0.10751538816500814,
|
| 3019 |
+
"learning_rate": 1.5812401303639813e-07,
|
| 3020 |
+
"loss": 0.253,
|
| 3021 |
+
"step": 428
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 2.948453608247423,
|
| 3025 |
+
"grad_norm": 0.10902510633946422,
|
| 3026 |
+
"learning_rate": 1.1618086972559062e-07,
|
| 3027 |
+
"loss": 0.2534,
|
| 3028 |
+
"step": 429
|
| 3029 |
+
},
|
| 3030 |
+
{
|
| 3031 |
+
"epoch": 2.9553264604811,
|
| 3032 |
+
"grad_norm": 0.10961740281781805,
|
| 3033 |
+
"learning_rate": 8.068593435055505e-08,
|
| 3034 |
+
"loss": 0.2611,
|
| 3035 |
+
"step": 430
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 2.9621993127147768,
|
| 3039 |
+
"grad_norm": 0.10667412761426893,
|
| 3040 |
+
"learning_rate": 5.164149836077714e-08,
|
| 3041 |
+
"loss": 0.2566,
|
| 3042 |
+
"step": 431
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 2.9690721649484537,
|
| 3046 |
+
"grad_norm": 0.10469145515742652,
|
| 3047 |
+
"learning_rate": 2.9049436780281825e-08,
|
| 3048 |
+
"loss": 0.245,
|
| 3049 |
+
"step": 432
|
| 3050 |
+
},
|
| 3051 |
+
{
|
| 3052 |
+
"epoch": 2.9759450171821307,
|
| 3053 |
+
"grad_norm": 0.10866002746514136,
|
| 3054 |
+
"learning_rate": 1.2911208086663351e-08,
|
| 3055 |
+
"loss": 0.258,
|
| 3056 |
+
"step": 433
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 2.9828178694158076,
|
| 3060 |
+
"grad_norm": 0.10779386109678486,
|
| 3061 |
+
"learning_rate": 3.2278541168717646e-09,
|
| 3062 |
+
"loss": 0.2668,
|
| 3063 |
+
"step": 434
|
| 3064 |
+
},
|
| 3065 |
+
{
|
| 3066 |
+
"epoch": 2.9896907216494846,
|
| 3067 |
+
"grad_norm": 0.11009984251752782,
|
| 3068 |
+
"learning_rate": 0.0,
|
| 3069 |
+
"loss": 0.2592,
|
| 3070 |
+
"step": 435
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"epoch": 2.9896907216494846,
|
| 3074 |
+
"eval_loss": 0.29278308153152466,
|
| 3075 |
+
"eval_runtime": 27.5656,
|
| 3076 |
+
"eval_samples_per_second": 35.443,
|
| 3077 |
+
"eval_steps_per_second": 1.125,
|
| 3078 |
+
"step": 435
|
| 3079 |
+
},
|
| 3080 |
+
{
|
| 3081 |
+
"epoch": 2.9896907216494846,
|
| 3082 |
+
"step": 435,
|
| 3083 |
+
"total_flos": 1.330526714897367e+17,
|
| 3084 |
+
"train_loss": 0.3336494640714821,
|
| 3085 |
+
"train_runtime": 4477.4048,
|
| 3086 |
+
"train_samples_per_second": 12.436,
|
| 3087 |
+
"train_steps_per_second": 0.097
|
| 3088 |
+
}
|
| 3089 |
+
],
|
| 3090 |
+
"logging_steps": 1,
|
| 3091 |
+
"max_steps": 435,
|
| 3092 |
+
"num_input_tokens_seen": 0,
|
| 3093 |
+
"num_train_epochs": 3,
|
| 3094 |
+
"save_steps": 500,
|
| 3095 |
+
"stateful_callbacks": {
|
| 3096 |
+
"TrainerControl": {
|
| 3097 |
+
"args": {
|
| 3098 |
+
"should_epoch_stop": false,
|
| 3099 |
+
"should_evaluate": false,
|
| 3100 |
+
"should_log": false,
|
| 3101 |
+
"should_save": true,
|
| 3102 |
+
"should_training_stop": true
|
| 3103 |
+
},
|
| 3104 |
+
"attributes": {}
|
| 3105 |
+
}
|
| 3106 |
+
},
|
| 3107 |
+
"total_flos": 1.330526714897367e+17,
|
| 3108 |
+
"train_batch_size": 8,
|
| 3109 |
+
"trial_name": null,
|
| 3110 |
+
"trial_params": null
|
| 3111 |
+
}
|