Model save
Browse files- README.md +69 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +3321 -0
README.md
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: barc0/cot-400k-barc-llama3.1-8b-ins-fft-transduction_lr1e-5_epoch3
|
| 3 |
+
library_name: peft
|
| 4 |
+
license: llama3.1
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- sft
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: cot-trainset-ft-transduction-v2-lora-train
|
| 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 |
+
# cot-trainset-ft-transduction-v2-lora-train
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [barc0/cot-400k-barc-llama3.1-8b-ins-fft-transduction_lr1e-5_epoch3](https://huggingface.co/barc0/cot-400k-barc-llama3.1-8b-ins-fft-transduction_lr1e-5_epoch3) on an unknown dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.2107
|
| 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: 2
|
| 42 |
+
- eval_batch_size: 2
|
| 43 |
+
- seed: 42
|
| 44 |
+
- distributed_type: multi-GPU
|
| 45 |
+
- num_devices: 4
|
| 46 |
+
- gradient_accumulation_steps: 2
|
| 47 |
+
- total_train_batch_size: 16
|
| 48 |
+
- total_eval_batch_size: 8
|
| 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 |
+
| 0.1316 | 1.0 | 155 | 0.1494 |
|
| 59 |
+
| 0.073 | 2.0 | 310 | 0.1669 |
|
| 60 |
+
| 0.0305 | 3.0 | 465 | 0.2107 |
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
### Framework versions
|
| 64 |
+
|
| 65 |
+
- PEFT 0.12.0
|
| 66 |
+
- Transformers 4.45.0.dev0
|
| 67 |
+
- Pytorch 2.4.0+cu121
|
| 68 |
+
- Datasets 2.21.0
|
| 69 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 261374226563072.0,
|
| 4 |
+
"train_loss": 0.09038178783751304,
|
| 5 |
+
"train_runtime": 3542.1658,
|
| 6 |
+
"train_samples": 2476,
|
| 7 |
+
"train_samples_per_second": 2.097,
|
| 8 |
+
"train_steps_per_second": 0.131
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 261374226563072.0,
|
| 4 |
+
"train_loss": 0.09038178783751304,
|
| 5 |
+
"train_runtime": 3542.1658,
|
| 6 |
+
"train_samples": 2476,
|
| 7 |
+
"train_samples_per_second": 2.097,
|
| 8 |
+
"train_steps_per_second": 0.131
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3321 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 465,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0064516129032258064,
|
| 13 |
+
"grad_norm": 0.15169227525554954,
|
| 14 |
+
"learning_rate": 4.255319148936171e-06,
|
| 15 |
+
"loss": 0.1803,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.012903225806451613,
|
| 20 |
+
"grad_norm": 0.1108587617034759,
|
| 21 |
+
"learning_rate": 8.510638297872341e-06,
|
| 22 |
+
"loss": 0.1441,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01935483870967742,
|
| 27 |
+
"grad_norm": 0.08345105115839851,
|
| 28 |
+
"learning_rate": 1.2765957446808511e-05,
|
| 29 |
+
"loss": 0.1118,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.025806451612903226,
|
| 34 |
+
"grad_norm": 0.1426234633379846,
|
| 35 |
+
"learning_rate": 1.7021276595744682e-05,
|
| 36 |
+
"loss": 0.1313,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03225806451612903,
|
| 41 |
+
"grad_norm": 0.14217530777409554,
|
| 42 |
+
"learning_rate": 2.1276595744680852e-05,
|
| 43 |
+
"loss": 0.1195,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.03870967741935484,
|
| 48 |
+
"grad_norm": 0.1459141077043439,
|
| 49 |
+
"learning_rate": 2.5531914893617022e-05,
|
| 50 |
+
"loss": 0.1453,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.04516129032258064,
|
| 55 |
+
"grad_norm": 0.10195797727709278,
|
| 56 |
+
"learning_rate": 2.9787234042553192e-05,
|
| 57 |
+
"loss": 0.1119,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.05161290322580645,
|
| 62 |
+
"grad_norm": 0.13553076939195616,
|
| 63 |
+
"learning_rate": 3.4042553191489365e-05,
|
| 64 |
+
"loss": 0.1791,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.05806451612903226,
|
| 69 |
+
"grad_norm": 0.15732711391586654,
|
| 70 |
+
"learning_rate": 3.829787234042553e-05,
|
| 71 |
+
"loss": 0.1282,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.06451612903225806,
|
| 76 |
+
"grad_norm": 0.1192108991071351,
|
| 77 |
+
"learning_rate": 4.2553191489361704e-05,
|
| 78 |
+
"loss": 0.1288,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.07096774193548387,
|
| 83 |
+
"grad_norm": 0.1534053229135478,
|
| 84 |
+
"learning_rate": 4.680851063829788e-05,
|
| 85 |
+
"loss": 0.15,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.07741935483870968,
|
| 90 |
+
"grad_norm": 0.11808236646609899,
|
| 91 |
+
"learning_rate": 5.1063829787234044e-05,
|
| 92 |
+
"loss": 0.1098,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.08387096774193549,
|
| 97 |
+
"grad_norm": 0.1127350100020763,
|
| 98 |
+
"learning_rate": 5.531914893617022e-05,
|
| 99 |
+
"loss": 0.1214,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.09032258064516129,
|
| 104 |
+
"grad_norm": 0.13802666747672132,
|
| 105 |
+
"learning_rate": 5.9574468085106384e-05,
|
| 106 |
+
"loss": 0.1555,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.0967741935483871,
|
| 111 |
+
"grad_norm": 0.19187355993638458,
|
| 112 |
+
"learning_rate": 6.382978723404256e-05,
|
| 113 |
+
"loss": 0.1474,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.1032258064516129,
|
| 118 |
+
"grad_norm": 0.16430813087649648,
|
| 119 |
+
"learning_rate": 6.808510638297873e-05,
|
| 120 |
+
"loss": 0.1461,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.10967741935483871,
|
| 125 |
+
"grad_norm": 0.13789502094335715,
|
| 126 |
+
"learning_rate": 7.23404255319149e-05,
|
| 127 |
+
"loss": 0.1226,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.11612903225806452,
|
| 132 |
+
"grad_norm": 0.173980091579583,
|
| 133 |
+
"learning_rate": 7.659574468085106e-05,
|
| 134 |
+
"loss": 0.1558,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.12258064516129032,
|
| 139 |
+
"grad_norm": 0.1418687450951512,
|
| 140 |
+
"learning_rate": 8.085106382978723e-05,
|
| 141 |
+
"loss": 0.1379,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.12903225806451613,
|
| 146 |
+
"grad_norm": 0.1765388069700401,
|
| 147 |
+
"learning_rate": 8.510638297872341e-05,
|
| 148 |
+
"loss": 0.149,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.13548387096774195,
|
| 153 |
+
"grad_norm": 0.17659794510341198,
|
| 154 |
+
"learning_rate": 8.936170212765958e-05,
|
| 155 |
+
"loss": 0.1101,
|
| 156 |
+
"step": 21
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.14193548387096774,
|
| 160 |
+
"grad_norm": 0.22633201272355616,
|
| 161 |
+
"learning_rate": 9.361702127659576e-05,
|
| 162 |
+
"loss": 0.1656,
|
| 163 |
+
"step": 22
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.14838709677419354,
|
| 167 |
+
"grad_norm": 0.16004849032165383,
|
| 168 |
+
"learning_rate": 9.787234042553192e-05,
|
| 169 |
+
"loss": 0.1616,
|
| 170 |
+
"step": 23
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.15483870967741936,
|
| 174 |
+
"grad_norm": 0.16981049690586422,
|
| 175 |
+
"learning_rate": 0.00010212765957446809,
|
| 176 |
+
"loss": 0.1321,
|
| 177 |
+
"step": 24
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.16129032258064516,
|
| 181 |
+
"grad_norm": 0.13581353274905067,
|
| 182 |
+
"learning_rate": 0.00010638297872340425,
|
| 183 |
+
"loss": 0.1122,
|
| 184 |
+
"step": 25
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.16774193548387098,
|
| 188 |
+
"grad_norm": 0.2089185781750498,
|
| 189 |
+
"learning_rate": 0.00011063829787234043,
|
| 190 |
+
"loss": 0.1527,
|
| 191 |
+
"step": 26
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.17419354838709677,
|
| 195 |
+
"grad_norm": 0.16125655225492164,
|
| 196 |
+
"learning_rate": 0.00011489361702127661,
|
| 197 |
+
"loss": 0.1263,
|
| 198 |
+
"step": 27
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.18064516129032257,
|
| 202 |
+
"grad_norm": 0.13753328983503424,
|
| 203 |
+
"learning_rate": 0.00011914893617021277,
|
| 204 |
+
"loss": 0.1224,
|
| 205 |
+
"step": 28
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.1870967741935484,
|
| 209 |
+
"grad_norm": 0.19089590529877826,
|
| 210 |
+
"learning_rate": 0.00012340425531914893,
|
| 211 |
+
"loss": 0.168,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.1935483870967742,
|
| 216 |
+
"grad_norm": 0.15419681175807715,
|
| 217 |
+
"learning_rate": 0.00012765957446808513,
|
| 218 |
+
"loss": 0.138,
|
| 219 |
+
"step": 30
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.2,
|
| 223 |
+
"grad_norm": 0.18178665024271073,
|
| 224 |
+
"learning_rate": 0.00013191489361702127,
|
| 225 |
+
"loss": 0.1955,
|
| 226 |
+
"step": 31
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.2064516129032258,
|
| 230 |
+
"grad_norm": 0.11760523051627117,
|
| 231 |
+
"learning_rate": 0.00013617021276595746,
|
| 232 |
+
"loss": 0.1367,
|
| 233 |
+
"step": 32
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.2129032258064516,
|
| 237 |
+
"grad_norm": 0.10853469531105876,
|
| 238 |
+
"learning_rate": 0.00014042553191489363,
|
| 239 |
+
"loss": 0.1208,
|
| 240 |
+
"step": 33
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.21935483870967742,
|
| 244 |
+
"grad_norm": 0.12890024443339382,
|
| 245 |
+
"learning_rate": 0.0001446808510638298,
|
| 246 |
+
"loss": 0.142,
|
| 247 |
+
"step": 34
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.22580645161290322,
|
| 251 |
+
"grad_norm": 0.12692879603118554,
|
| 252 |
+
"learning_rate": 0.00014893617021276596,
|
| 253 |
+
"loss": 0.1268,
|
| 254 |
+
"step": 35
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.23225806451612904,
|
| 258 |
+
"grad_norm": 0.1529612097417899,
|
| 259 |
+
"learning_rate": 0.00015319148936170213,
|
| 260 |
+
"loss": 0.1194,
|
| 261 |
+
"step": 36
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.23870967741935484,
|
| 265 |
+
"grad_norm": 0.129432114467059,
|
| 266 |
+
"learning_rate": 0.00015744680851063832,
|
| 267 |
+
"loss": 0.128,
|
| 268 |
+
"step": 37
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.24516129032258063,
|
| 272 |
+
"grad_norm": 0.11343138573627701,
|
| 273 |
+
"learning_rate": 0.00016170212765957446,
|
| 274 |
+
"loss": 0.102,
|
| 275 |
+
"step": 38
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.25161290322580643,
|
| 279 |
+
"grad_norm": 0.16577469112721435,
|
| 280 |
+
"learning_rate": 0.00016595744680851065,
|
| 281 |
+
"loss": 0.1691,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.25806451612903225,
|
| 286 |
+
"grad_norm": 0.13614466586953358,
|
| 287 |
+
"learning_rate": 0.00017021276595744682,
|
| 288 |
+
"loss": 0.1348,
|
| 289 |
+
"step": 40
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.2645161290322581,
|
| 293 |
+
"grad_norm": 0.1524812917659128,
|
| 294 |
+
"learning_rate": 0.00017446808510638298,
|
| 295 |
+
"loss": 0.1281,
|
| 296 |
+
"step": 41
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.2709677419354839,
|
| 300 |
+
"grad_norm": 0.14519269708506746,
|
| 301 |
+
"learning_rate": 0.00017872340425531915,
|
| 302 |
+
"loss": 0.1349,
|
| 303 |
+
"step": 42
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.27741935483870966,
|
| 307 |
+
"grad_norm": 0.18300481897670345,
|
| 308 |
+
"learning_rate": 0.00018297872340425532,
|
| 309 |
+
"loss": 0.1694,
|
| 310 |
+
"step": 43
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.2838709677419355,
|
| 314 |
+
"grad_norm": 0.11929331561330575,
|
| 315 |
+
"learning_rate": 0.0001872340425531915,
|
| 316 |
+
"loss": 0.1331,
|
| 317 |
+
"step": 44
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.2903225806451613,
|
| 321 |
+
"grad_norm": 0.12604932986068976,
|
| 322 |
+
"learning_rate": 0.00019148936170212768,
|
| 323 |
+
"loss": 0.0949,
|
| 324 |
+
"step": 45
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.2967741935483871,
|
| 328 |
+
"grad_norm": 0.14226790630565908,
|
| 329 |
+
"learning_rate": 0.00019574468085106384,
|
| 330 |
+
"loss": 0.1477,
|
| 331 |
+
"step": 46
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.3032258064516129,
|
| 335 |
+
"grad_norm": 0.1465980952326029,
|
| 336 |
+
"learning_rate": 0.0002,
|
| 337 |
+
"loss": 0.1408,
|
| 338 |
+
"step": 47
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3096774193548387,
|
| 342 |
+
"grad_norm": 0.14535685898764863,
|
| 343 |
+
"learning_rate": 0.0001999971756719333,
|
| 344 |
+
"loss": 0.1474,
|
| 345 |
+
"step": 48
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.3161290322580645,
|
| 349 |
+
"grad_norm": 0.1559112597851861,
|
| 350 |
+
"learning_rate": 0.00019998870284726968,
|
| 351 |
+
"loss": 0.1568,
|
| 352 |
+
"step": 49
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.3225806451612903,
|
| 356 |
+
"grad_norm": 0.14392375991483491,
|
| 357 |
+
"learning_rate": 0.00019997458200460993,
|
| 358 |
+
"loss": 0.1526,
|
| 359 |
+
"step": 50
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.32903225806451614,
|
| 363 |
+
"grad_norm": 0.18920444771524633,
|
| 364 |
+
"learning_rate": 0.00019995481394159188,
|
| 365 |
+
"loss": 0.1384,
|
| 366 |
+
"step": 51
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.33548387096774196,
|
| 370 |
+
"grad_norm": 0.17943388616343298,
|
| 371 |
+
"learning_rate": 0.0001999293997748454,
|
| 372 |
+
"loss": 0.1555,
|
| 373 |
+
"step": 52
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.3419354838709677,
|
| 377 |
+
"grad_norm": 0.16752646291327727,
|
| 378 |
+
"learning_rate": 0.00019989834093992945,
|
| 379 |
+
"loss": 0.1634,
|
| 380 |
+
"step": 53
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.34838709677419355,
|
| 384 |
+
"grad_norm": 0.1484740777328073,
|
| 385 |
+
"learning_rate": 0.00019986163919125075,
|
| 386 |
+
"loss": 0.124,
|
| 387 |
+
"step": 54
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.3548387096774194,
|
| 391 |
+
"grad_norm": 0.15268630472434244,
|
| 392 |
+
"learning_rate": 0.00019981929660196492,
|
| 393 |
+
"loss": 0.1228,
|
| 394 |
+
"step": 55
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.36129032258064514,
|
| 398 |
+
"grad_norm": 0.19583236181584973,
|
| 399 |
+
"learning_rate": 0.0001997713155638592,
|
| 400 |
+
"loss": 0.1227,
|
| 401 |
+
"step": 56
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.36774193548387096,
|
| 405 |
+
"grad_norm": 0.1469999186659826,
|
| 406 |
+
"learning_rate": 0.00019971769878721743,
|
| 407 |
+
"loss": 0.1188,
|
| 408 |
+
"step": 57
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.3741935483870968,
|
| 412 |
+
"grad_norm": 0.10845193424043034,
|
| 413 |
+
"learning_rate": 0.000199658449300667,
|
| 414 |
+
"loss": 0.1177,
|
| 415 |
+
"step": 58
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.38064516129032255,
|
| 419 |
+
"grad_norm": 0.1845188187089657,
|
| 420 |
+
"learning_rate": 0.00019959357045100764,
|
| 421 |
+
"loss": 0.1726,
|
| 422 |
+
"step": 59
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.3870967741935484,
|
| 426 |
+
"grad_norm": 0.14769124665651473,
|
| 427 |
+
"learning_rate": 0.00019952306590302247,
|
| 428 |
+
"loss": 0.1634,
|
| 429 |
+
"step": 60
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.3935483870967742,
|
| 433 |
+
"grad_norm": 0.16408468392163889,
|
| 434 |
+
"learning_rate": 0.00019944693963927092,
|
| 435 |
+
"loss": 0.1584,
|
| 436 |
+
"step": 61
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.4,
|
| 440 |
+
"grad_norm": 0.15156206723802879,
|
| 441 |
+
"learning_rate": 0.00019936519595986394,
|
| 442 |
+
"loss": 0.1454,
|
| 443 |
+
"step": 62
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.4064516129032258,
|
| 447 |
+
"grad_norm": 0.15835599927161748,
|
| 448 |
+
"learning_rate": 0.00019927783948222084,
|
| 449 |
+
"loss": 0.1621,
|
| 450 |
+
"step": 63
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.4129032258064516,
|
| 454 |
+
"grad_norm": 0.1333411665662589,
|
| 455 |
+
"learning_rate": 0.00019918487514080865,
|
| 456 |
+
"loss": 0.1293,
|
| 457 |
+
"step": 64
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.41935483870967744,
|
| 461 |
+
"grad_norm": 0.1589874051481937,
|
| 462 |
+
"learning_rate": 0.00019908630818686338,
|
| 463 |
+
"loss": 0.1391,
|
| 464 |
+
"step": 65
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.4258064516129032,
|
| 468 |
+
"grad_norm": 0.12399538770065353,
|
| 469 |
+
"learning_rate": 0.0001989821441880933,
|
| 470 |
+
"loss": 0.1208,
|
| 471 |
+
"step": 66
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.432258064516129,
|
| 475 |
+
"grad_norm": 0.16949673743147955,
|
| 476 |
+
"learning_rate": 0.00019887238902836448,
|
| 477 |
+
"loss": 0.1139,
|
| 478 |
+
"step": 67
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.43870967741935485,
|
| 482 |
+
"grad_norm": 0.1875430265393267,
|
| 483 |
+
"learning_rate": 0.00019875704890736853,
|
| 484 |
+
"loss": 0.1856,
|
| 485 |
+
"step": 68
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.44516129032258067,
|
| 489 |
+
"grad_norm": 0.14927152386929934,
|
| 490 |
+
"learning_rate": 0.00019863613034027224,
|
| 491 |
+
"loss": 0.1516,
|
| 492 |
+
"step": 69
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.45161290322580644,
|
| 496 |
+
"grad_norm": 0.17501653378304205,
|
| 497 |
+
"learning_rate": 0.0001985096401573497,
|
| 498 |
+
"loss": 0.161,
|
| 499 |
+
"step": 70
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.45806451612903226,
|
| 503 |
+
"grad_norm": 0.17582762670350804,
|
| 504 |
+
"learning_rate": 0.00019837758550359636,
|
| 505 |
+
"loss": 0.1564,
|
| 506 |
+
"step": 71
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.4645161290322581,
|
| 510 |
+
"grad_norm": 0.15692049884405931,
|
| 511 |
+
"learning_rate": 0.0001982399738383255,
|
| 512 |
+
"loss": 0.1282,
|
| 513 |
+
"step": 72
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.47096774193548385,
|
| 517 |
+
"grad_norm": 0.14436392088728167,
|
| 518 |
+
"learning_rate": 0.00019809681293474693,
|
| 519 |
+
"loss": 0.1299,
|
| 520 |
+
"step": 73
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.4774193548387097,
|
| 524 |
+
"grad_norm": 0.19048463762976417,
|
| 525 |
+
"learning_rate": 0.0001979481108795278,
|
| 526 |
+
"loss": 0.1983,
|
| 527 |
+
"step": 74
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.4838709677419355,
|
| 531 |
+
"grad_norm": 0.13475205452089994,
|
| 532 |
+
"learning_rate": 0.00019779387607233586,
|
| 533 |
+
"loss": 0.156,
|
| 534 |
+
"step": 75
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.49032258064516127,
|
| 538 |
+
"grad_norm": 0.14145398172929924,
|
| 539 |
+
"learning_rate": 0.00019763411722536502,
|
| 540 |
+
"loss": 0.1355,
|
| 541 |
+
"step": 76
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.4967741935483871,
|
| 545 |
+
"grad_norm": 0.14400838677113517,
|
| 546 |
+
"learning_rate": 0.00019746884336284317,
|
| 547 |
+
"loss": 0.1371,
|
| 548 |
+
"step": 77
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.5032258064516129,
|
| 552 |
+
"grad_norm": 0.13421206031331137,
|
| 553 |
+
"learning_rate": 0.00019729806382052248,
|
| 554 |
+
"loss": 0.1156,
|
| 555 |
+
"step": 78
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.5096774193548387,
|
| 559 |
+
"grad_norm": 0.11613915473665105,
|
| 560 |
+
"learning_rate": 0.00019712178824515212,
|
| 561 |
+
"loss": 0.1293,
|
| 562 |
+
"step": 79
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.5161290322580645,
|
| 566 |
+
"grad_norm": 0.12472786830602234,
|
| 567 |
+
"learning_rate": 0.00019694002659393305,
|
| 568 |
+
"loss": 0.1189,
|
| 569 |
+
"step": 80
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.5225806451612903,
|
| 573 |
+
"grad_norm": 0.11050712097688373,
|
| 574 |
+
"learning_rate": 0.00019675278913395606,
|
| 575 |
+
"loss": 0.12,
|
| 576 |
+
"step": 81
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.5290322580645161,
|
| 580 |
+
"grad_norm": 0.1303276052231771,
|
| 581 |
+
"learning_rate": 0.0001965600864416213,
|
| 582 |
+
"loss": 0.1428,
|
| 583 |
+
"step": 82
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.535483870967742,
|
| 587 |
+
"grad_norm": 0.13774570876008593,
|
| 588 |
+
"learning_rate": 0.00019636192940204134,
|
| 589 |
+
"loss": 0.1438,
|
| 590 |
+
"step": 83
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.5419354838709678,
|
| 594 |
+
"grad_norm": 0.14810394982940484,
|
| 595 |
+
"learning_rate": 0.00019615832920842586,
|
| 596 |
+
"loss": 0.1404,
|
| 597 |
+
"step": 84
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.5483870967741935,
|
| 601 |
+
"grad_norm": 0.144275852100491,
|
| 602 |
+
"learning_rate": 0.00019594929736144976,
|
| 603 |
+
"loss": 0.1357,
|
| 604 |
+
"step": 85
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.5548387096774193,
|
| 608 |
+
"grad_norm": 0.16879600090790034,
|
| 609 |
+
"learning_rate": 0.0001957348456686032,
|
| 610 |
+
"loss": 0.1578,
|
| 611 |
+
"step": 86
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.5612903225806452,
|
| 615 |
+
"grad_norm": 0.1588074767274021,
|
| 616 |
+
"learning_rate": 0.00019551498624352496,
|
| 617 |
+
"loss": 0.1228,
|
| 618 |
+
"step": 87
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.567741935483871,
|
| 622 |
+
"grad_norm": 0.15067916918276408,
|
| 623 |
+
"learning_rate": 0.00019528973150531787,
|
| 624 |
+
"loss": 0.1323,
|
| 625 |
+
"step": 88
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.5741935483870968,
|
| 629 |
+
"grad_norm": 0.16266341913656662,
|
| 630 |
+
"learning_rate": 0.00019505909417784754,
|
| 631 |
+
"loss": 0.13,
|
| 632 |
+
"step": 89
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.5806451612903226,
|
| 636 |
+
"grad_norm": 0.121529537729675,
|
| 637 |
+
"learning_rate": 0.00019482308728902356,
|
| 638 |
+
"loss": 0.1067,
|
| 639 |
+
"step": 90
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.5870967741935483,
|
| 643 |
+
"grad_norm": 0.1740468182649888,
|
| 644 |
+
"learning_rate": 0.00019458172417006347,
|
| 645 |
+
"loss": 0.1513,
|
| 646 |
+
"step": 91
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.5935483870967742,
|
| 650 |
+
"grad_norm": 0.14062538318374462,
|
| 651 |
+
"learning_rate": 0.00019433501845473995,
|
| 652 |
+
"loss": 0.1438,
|
| 653 |
+
"step": 92
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.6,
|
| 657 |
+
"grad_norm": 0.17690034130801896,
|
| 658 |
+
"learning_rate": 0.00019408298407861042,
|
| 659 |
+
"loss": 0.1356,
|
| 660 |
+
"step": 93
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.6064516129032258,
|
| 664 |
+
"grad_norm": 0.18789482750619546,
|
| 665 |
+
"learning_rate": 0.00019382563527823026,
|
| 666 |
+
"loss": 0.1758,
|
| 667 |
+
"step": 94
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.6129032258064516,
|
| 671 |
+
"grad_norm": 0.16993067663839118,
|
| 672 |
+
"learning_rate": 0.00019356298659034817,
|
| 673 |
+
"loss": 0.1599,
|
| 674 |
+
"step": 95
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.6193548387096774,
|
| 678 |
+
"grad_norm": 0.16495058136550117,
|
| 679 |
+
"learning_rate": 0.00019329505285108542,
|
| 680 |
+
"loss": 0.1283,
|
| 681 |
+
"step": 96
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.6258064516129033,
|
| 685 |
+
"grad_norm": 0.15497395377626808,
|
| 686 |
+
"learning_rate": 0.00019302184919509755,
|
| 687 |
+
"loss": 0.1493,
|
| 688 |
+
"step": 97
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.632258064516129,
|
| 692 |
+
"grad_norm": 0.13660921526912856,
|
| 693 |
+
"learning_rate": 0.00019274339105471971,
|
| 694 |
+
"loss": 0.1307,
|
| 695 |
+
"step": 98
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.6387096774193548,
|
| 699 |
+
"grad_norm": 0.18246231884688152,
|
| 700 |
+
"learning_rate": 0.00019245969415909465,
|
| 701 |
+
"loss": 0.1598,
|
| 702 |
+
"step": 99
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.6451612903225806,
|
| 706 |
+
"grad_norm": 0.11344749529118914,
|
| 707 |
+
"learning_rate": 0.00019217077453328449,
|
| 708 |
+
"loss": 0.1304,
|
| 709 |
+
"step": 100
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.6516129032258065,
|
| 713 |
+
"grad_norm": 0.11682283731326468,
|
| 714 |
+
"learning_rate": 0.0001918766484973654,
|
| 715 |
+
"loss": 0.0977,
|
| 716 |
+
"step": 101
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.6580645161290323,
|
| 720 |
+
"grad_norm": 0.14494532290577813,
|
| 721 |
+
"learning_rate": 0.00019157733266550575,
|
| 722 |
+
"loss": 0.1338,
|
| 723 |
+
"step": 102
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.6645161290322581,
|
| 727 |
+
"grad_norm": 0.12095275202026515,
|
| 728 |
+
"learning_rate": 0.0001912728439450276,
|
| 729 |
+
"loss": 0.1513,
|
| 730 |
+
"step": 103
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.6709677419354839,
|
| 734 |
+
"grad_norm": 0.13423702981009097,
|
| 735 |
+
"learning_rate": 0.00019096319953545185,
|
| 736 |
+
"loss": 0.1335,
|
| 737 |
+
"step": 104
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.6774193548387096,
|
| 741 |
+
"grad_norm": 0.1670132860946028,
|
| 742 |
+
"learning_rate": 0.0001906484169275263,
|
| 743 |
+
"loss": 0.1607,
|
| 744 |
+
"step": 105
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.6838709677419355,
|
| 748 |
+
"grad_norm": 0.14053854153152684,
|
| 749 |
+
"learning_rate": 0.00019032851390223812,
|
| 750 |
+
"loss": 0.1365,
|
| 751 |
+
"step": 106
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.6903225806451613,
|
| 755 |
+
"grad_norm": 0.1399807021991922,
|
| 756 |
+
"learning_rate": 0.00019000350852980909,
|
| 757 |
+
"loss": 0.1589,
|
| 758 |
+
"step": 107
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.6967741935483871,
|
| 762 |
+
"grad_norm": 0.15473299551506894,
|
| 763 |
+
"learning_rate": 0.00018967341916867518,
|
| 764 |
+
"loss": 0.166,
|
| 765 |
+
"step": 108
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.7032258064516129,
|
| 769 |
+
"grad_norm": 0.1536872117526864,
|
| 770 |
+
"learning_rate": 0.00018933826446444933,
|
| 771 |
+
"loss": 0.1657,
|
| 772 |
+
"step": 109
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.7096774193548387,
|
| 776 |
+
"grad_norm": 0.1282004975556196,
|
| 777 |
+
"learning_rate": 0.0001889980633488683,
|
| 778 |
+
"loss": 0.1212,
|
| 779 |
+
"step": 110
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.7161290322580646,
|
| 783 |
+
"grad_norm": 0.16458328388975405,
|
| 784 |
+
"learning_rate": 0.00018865283503872324,
|
| 785 |
+
"loss": 0.1655,
|
| 786 |
+
"step": 111
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.7225806451612903,
|
| 790 |
+
"grad_norm": 0.1505113828615181,
|
| 791 |
+
"learning_rate": 0.00018830259903477426,
|
| 792 |
+
"loss": 0.1571,
|
| 793 |
+
"step": 112
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.7290322580645161,
|
| 797 |
+
"grad_norm": 0.16575595088070239,
|
| 798 |
+
"learning_rate": 0.0001879473751206489,
|
| 799 |
+
"loss": 0.1504,
|
| 800 |
+
"step": 113
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.7354838709677419,
|
| 804 |
+
"grad_norm": 0.1484230902451611,
|
| 805 |
+
"learning_rate": 0.0001875871833617246,
|
| 806 |
+
"loss": 0.1498,
|
| 807 |
+
"step": 114
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.7419354838709677,
|
| 811 |
+
"grad_norm": 0.2059907492830938,
|
| 812 |
+
"learning_rate": 0.0001872220441039952,
|
| 813 |
+
"loss": 0.134,
|
| 814 |
+
"step": 115
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.7483870967741936,
|
| 818 |
+
"grad_norm": 0.1491671097000444,
|
| 819 |
+
"learning_rate": 0.0001868519779729218,
|
| 820 |
+
"loss": 0.1374,
|
| 821 |
+
"step": 116
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.7548387096774194,
|
| 825 |
+
"grad_norm": 0.1727675086328308,
|
| 826 |
+
"learning_rate": 0.0001864770058722676,
|
| 827 |
+
"loss": 0.1624,
|
| 828 |
+
"step": 117
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.7612903225806451,
|
| 832 |
+
"grad_norm": 0.16357318423846662,
|
| 833 |
+
"learning_rate": 0.00018609714898291718,
|
| 834 |
+
"loss": 0.1528,
|
| 835 |
+
"step": 118
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.7677419354838709,
|
| 839 |
+
"grad_norm": 0.1584052674932312,
|
| 840 |
+
"learning_rate": 0.00018571242876167996,
|
| 841 |
+
"loss": 0.1321,
|
| 842 |
+
"step": 119
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.7741935483870968,
|
| 846 |
+
"grad_norm": 0.21714396600094343,
|
| 847 |
+
"learning_rate": 0.0001853228669400784,
|
| 848 |
+
"loss": 0.1748,
|
| 849 |
+
"step": 120
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.7806451612903226,
|
| 853 |
+
"grad_norm": 0.13743651841776636,
|
| 854 |
+
"learning_rate": 0.00018492848552312014,
|
| 855 |
+
"loss": 0.1493,
|
| 856 |
+
"step": 121
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.7870967741935484,
|
| 860 |
+
"grad_norm": 0.1541126978032927,
|
| 861 |
+
"learning_rate": 0.00018452930678805536,
|
| 862 |
+
"loss": 0.1331,
|
| 863 |
+
"step": 122
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.7935483870967742,
|
| 867 |
+
"grad_norm": 0.1571822882709535,
|
| 868 |
+
"learning_rate": 0.00018412535328311814,
|
| 869 |
+
"loss": 0.1427,
|
| 870 |
+
"step": 123
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.8,
|
| 874 |
+
"grad_norm": 0.17825934340851243,
|
| 875 |
+
"learning_rate": 0.00018371664782625287,
|
| 876 |
+
"loss": 0.1871,
|
| 877 |
+
"step": 124
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.8064516129032258,
|
| 881 |
+
"grad_norm": 0.1745800846806893,
|
| 882 |
+
"learning_rate": 0.00018330321350382544,
|
| 883 |
+
"loss": 0.1672,
|
| 884 |
+
"step": 125
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.8129032258064516,
|
| 888 |
+
"grad_norm": 0.13970218689990957,
|
| 889 |
+
"learning_rate": 0.00018288507366931905,
|
| 890 |
+
"loss": 0.1715,
|
| 891 |
+
"step": 126
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.8193548387096774,
|
| 895 |
+
"grad_norm": 0.15100292169016535,
|
| 896 |
+
"learning_rate": 0.00018246225194201517,
|
| 897 |
+
"loss": 0.1411,
|
| 898 |
+
"step": 127
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.8258064516129032,
|
| 902 |
+
"grad_norm": 0.1538126122981586,
|
| 903 |
+
"learning_rate": 0.00018203477220565912,
|
| 904 |
+
"loss": 0.1516,
|
| 905 |
+
"step": 128
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.832258064516129,
|
| 909 |
+
"grad_norm": 0.15630735898296536,
|
| 910 |
+
"learning_rate": 0.00018160265860711134,
|
| 911 |
+
"loss": 0.1636,
|
| 912 |
+
"step": 129
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.8387096774193549,
|
| 916 |
+
"grad_norm": 0.1427718560771215,
|
| 917 |
+
"learning_rate": 0.00018116593555498307,
|
| 918 |
+
"loss": 0.1297,
|
| 919 |
+
"step": 130
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.8451612903225807,
|
| 923 |
+
"grad_norm": 0.11911885929622754,
|
| 924 |
+
"learning_rate": 0.0001807246277182578,
|
| 925 |
+
"loss": 0.1115,
|
| 926 |
+
"step": 131
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.8516129032258064,
|
| 930 |
+
"grad_norm": 0.13702745725576418,
|
| 931 |
+
"learning_rate": 0.0001802787600248977,
|
| 932 |
+
"loss": 0.157,
|
| 933 |
+
"step": 132
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.8580645161290322,
|
| 937 |
+
"grad_norm": 0.17220644805792673,
|
| 938 |
+
"learning_rate": 0.0001798283576604356,
|
| 939 |
+
"loss": 0.1561,
|
| 940 |
+
"step": 133
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.864516129032258,
|
| 944 |
+
"grad_norm": 0.196867226472361,
|
| 945 |
+
"learning_rate": 0.0001793734460665523,
|
| 946 |
+
"loss": 0.1657,
|
| 947 |
+
"step": 134
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.8709677419354839,
|
| 951 |
+
"grad_norm": 0.1759359784948508,
|
| 952 |
+
"learning_rate": 0.00017891405093963938,
|
| 953 |
+
"loss": 0.1909,
|
| 954 |
+
"step": 135
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.8774193548387097,
|
| 958 |
+
"grad_norm": 0.16063163300244113,
|
| 959 |
+
"learning_rate": 0.0001784501982293479,
|
| 960 |
+
"loss": 0.1552,
|
| 961 |
+
"step": 136
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.8838709677419355,
|
| 965 |
+
"grad_norm": 0.17034183209183734,
|
| 966 |
+
"learning_rate": 0.00017798191413712243,
|
| 967 |
+
"loss": 0.1502,
|
| 968 |
+
"step": 137
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.8903225806451613,
|
| 972 |
+
"grad_norm": 0.14694529779128243,
|
| 973 |
+
"learning_rate": 0.0001775092251147211,
|
| 974 |
+
"loss": 0.1277,
|
| 975 |
+
"step": 138
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.896774193548387,
|
| 979 |
+
"grad_norm": 0.13174927545002138,
|
| 980 |
+
"learning_rate": 0.0001770321578627213,
|
| 981 |
+
"loss": 0.1277,
|
| 982 |
+
"step": 139
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.9032258064516129,
|
| 986 |
+
"grad_norm": 0.15784079444219237,
|
| 987 |
+
"learning_rate": 0.00017655073932901168,
|
| 988 |
+
"loss": 0.1534,
|
| 989 |
+
"step": 140
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.9096774193548387,
|
| 993 |
+
"grad_norm": 0.19691370909219638,
|
| 994 |
+
"learning_rate": 0.0001760649967072697,
|
| 995 |
+
"loss": 0.1688,
|
| 996 |
+
"step": 141
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.9161290322580645,
|
| 1000 |
+
"grad_norm": 0.15697904965484202,
|
| 1001 |
+
"learning_rate": 0.00017557495743542585,
|
| 1002 |
+
"loss": 0.1523,
|
| 1003 |
+
"step": 142
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.9225806451612903,
|
| 1007 |
+
"grad_norm": 0.1759980765103477,
|
| 1008 |
+
"learning_rate": 0.00017508064919411344,
|
| 1009 |
+
"loss": 0.15,
|
| 1010 |
+
"step": 143
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.9290322580645162,
|
| 1014 |
+
"grad_norm": 0.19254834048997346,
|
| 1015 |
+
"learning_rate": 0.00017458209990510527,
|
| 1016 |
+
"loss": 0.1474,
|
| 1017 |
+
"step": 144
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.9354838709677419,
|
| 1021 |
+
"grad_norm": 0.19220369870461818,
|
| 1022 |
+
"learning_rate": 0.00017407933772973637,
|
| 1023 |
+
"loss": 0.1678,
|
| 1024 |
+
"step": 145
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.9419354838709677,
|
| 1028 |
+
"grad_norm": 0.14789056250556576,
|
| 1029 |
+
"learning_rate": 0.00017357239106731317,
|
| 1030 |
+
"loss": 0.1634,
|
| 1031 |
+
"step": 146
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.9483870967741935,
|
| 1035 |
+
"grad_norm": 0.15823815803270533,
|
| 1036 |
+
"learning_rate": 0.00017306128855350942,
|
| 1037 |
+
"loss": 0.1744,
|
| 1038 |
+
"step": 147
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.9548387096774194,
|
| 1042 |
+
"grad_norm": 0.159128130793647,
|
| 1043 |
+
"learning_rate": 0.0001725460590587486,
|
| 1044 |
+
"loss": 0.1732,
|
| 1045 |
+
"step": 148
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.9612903225806452,
|
| 1049 |
+
"grad_norm": 0.1420932941022579,
|
| 1050 |
+
"learning_rate": 0.00017202673168657318,
|
| 1051 |
+
"loss": 0.1193,
|
| 1052 |
+
"step": 149
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.967741935483871,
|
| 1056 |
+
"grad_norm": 0.1581041276537875,
|
| 1057 |
+
"learning_rate": 0.0001715033357720006,
|
| 1058 |
+
"loss": 0.157,
|
| 1059 |
+
"step": 150
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.9741935483870968,
|
| 1063 |
+
"grad_norm": 0.13409040520330398,
|
| 1064 |
+
"learning_rate": 0.00017097590087986633,
|
| 1065 |
+
"loss": 0.1187,
|
| 1066 |
+
"step": 151
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.9806451612903225,
|
| 1070 |
+
"grad_norm": 0.15236261967510367,
|
| 1071 |
+
"learning_rate": 0.00017044445680315372,
|
| 1072 |
+
"loss": 0.1541,
|
| 1073 |
+
"step": 152
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.9870967741935484,
|
| 1077 |
+
"grad_norm": 0.15028826750982388,
|
| 1078 |
+
"learning_rate": 0.00016990903356131124,
|
| 1079 |
+
"loss": 0.1462,
|
| 1080 |
+
"step": 153
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.9935483870967742,
|
| 1084 |
+
"grad_norm": 0.14170487741522195,
|
| 1085 |
+
"learning_rate": 0.00016936966139855663,
|
| 1086 |
+
"loss": 0.1275,
|
| 1087 |
+
"step": 154
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 1.0,
|
| 1091 |
+
"grad_norm": 0.14288119090775778,
|
| 1092 |
+
"learning_rate": 0.00016882637078216868,
|
| 1093 |
+
"loss": 0.1316,
|
| 1094 |
+
"step": 155
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 1.0,
|
| 1098 |
+
"eval_loss": 0.14944089949131012,
|
| 1099 |
+
"eval_runtime": 27.6083,
|
| 1100 |
+
"eval_samples_per_second": 4.745,
|
| 1101 |
+
"eval_steps_per_second": 0.616,
|
| 1102 |
+
"step": 155
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 1.0064516129032257,
|
| 1106 |
+
"grad_norm": 0.08453811084356862,
|
| 1107 |
+
"learning_rate": 0.0001682791924007661,
|
| 1108 |
+
"loss": 0.0732,
|
| 1109 |
+
"step": 156
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.0129032258064516,
|
| 1113 |
+
"grad_norm": 0.1163343033490188,
|
| 1114 |
+
"learning_rate": 0.00016772815716257412,
|
| 1115 |
+
"loss": 0.0955,
|
| 1116 |
+
"step": 157
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 1.0193548387096774,
|
| 1120 |
+
"grad_norm": 0.12774969124655083,
|
| 1121 |
+
"learning_rate": 0.0001671732961936785,
|
| 1122 |
+
"loss": 0.1154,
|
| 1123 |
+
"step": 158
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.0258064516129033,
|
| 1127 |
+
"grad_norm": 0.10149504758995384,
|
| 1128 |
+
"learning_rate": 0.00016661464083626734,
|
| 1129 |
+
"loss": 0.0834,
|
| 1130 |
+
"step": 159
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 1.032258064516129,
|
| 1134 |
+
"grad_norm": 0.1558767636437416,
|
| 1135 |
+
"learning_rate": 0.00016605222264686086,
|
| 1136 |
+
"loss": 0.0978,
|
| 1137 |
+
"step": 160
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 1.038709677419355,
|
| 1141 |
+
"grad_norm": 0.14800857594022712,
|
| 1142 |
+
"learning_rate": 0.00016548607339452853,
|
| 1143 |
+
"loss": 0.0783,
|
| 1144 |
+
"step": 161
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 1.0451612903225806,
|
| 1148 |
+
"grad_norm": 0.10427437820954576,
|
| 1149 |
+
"learning_rate": 0.00016491622505909482,
|
| 1150 |
+
"loss": 0.0714,
|
| 1151 |
+
"step": 162
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 1.0516129032258064,
|
| 1155 |
+
"grad_norm": 0.16509881361577539,
|
| 1156 |
+
"learning_rate": 0.00016434270982933273,
|
| 1157 |
+
"loss": 0.0971,
|
| 1158 |
+
"step": 163
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 1.0580645161290323,
|
| 1162 |
+
"grad_norm": 0.15528612822142446,
|
| 1163 |
+
"learning_rate": 0.0001637655601011454,
|
| 1164 |
+
"loss": 0.0806,
|
| 1165 |
+
"step": 164
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.064516129032258,
|
| 1169 |
+
"grad_norm": 0.1578175598943513,
|
| 1170 |
+
"learning_rate": 0.00016318480847573642,
|
| 1171 |
+
"loss": 0.0962,
|
| 1172 |
+
"step": 165
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 1.070967741935484,
|
| 1176 |
+
"grad_norm": 0.2591491854389689,
|
| 1177 |
+
"learning_rate": 0.00016260048775776804,
|
| 1178 |
+
"loss": 0.1107,
|
| 1179 |
+
"step": 166
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.0774193548387097,
|
| 1183 |
+
"grad_norm": 0.1557337965418426,
|
| 1184 |
+
"learning_rate": 0.00016201263095350833,
|
| 1185 |
+
"loss": 0.0707,
|
| 1186 |
+
"step": 167
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.0838709677419356,
|
| 1190 |
+
"grad_norm": 0.24212581528206514,
|
| 1191 |
+
"learning_rate": 0.0001614212712689668,
|
| 1192 |
+
"loss": 0.0964,
|
| 1193 |
+
"step": 168
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.0903225806451613,
|
| 1197 |
+
"grad_norm": 0.12600203218602474,
|
| 1198 |
+
"learning_rate": 0.00016082644210801844,
|
| 1199 |
+
"loss": 0.0649,
|
| 1200 |
+
"step": 169
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.096774193548387,
|
| 1204 |
+
"grad_norm": 0.14377456483205683,
|
| 1205 |
+
"learning_rate": 0.00016022817707051724,
|
| 1206 |
+
"loss": 0.0778,
|
| 1207 |
+
"step": 170
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.103225806451613,
|
| 1211 |
+
"grad_norm": 0.1828183442232092,
|
| 1212 |
+
"learning_rate": 0.00015962650995039783,
|
| 1213 |
+
"loss": 0.1068,
|
| 1214 |
+
"step": 171
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 1.1096774193548387,
|
| 1218 |
+
"grad_norm": 0.1695078607562437,
|
| 1219 |
+
"learning_rate": 0.00015902147473376694,
|
| 1220 |
+
"loss": 0.0938,
|
| 1221 |
+
"step": 172
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.1161290322580646,
|
| 1225 |
+
"grad_norm": 0.10525262692642562,
|
| 1226 |
+
"learning_rate": 0.00015841310559698343,
|
| 1227 |
+
"loss": 0.0645,
|
| 1228 |
+
"step": 173
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.1225806451612903,
|
| 1232 |
+
"grad_norm": 0.15379971536703851,
|
| 1233 |
+
"learning_rate": 0.0001578014369047279,
|
| 1234 |
+
"loss": 0.0752,
|
| 1235 |
+
"step": 174
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.129032258064516,
|
| 1239 |
+
"grad_norm": 0.1352511385337785,
|
| 1240 |
+
"learning_rate": 0.00015718650320806142,
|
| 1241 |
+
"loss": 0.0803,
|
| 1242 |
+
"step": 175
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.135483870967742,
|
| 1246 |
+
"grad_norm": 0.1708537982318491,
|
| 1247 |
+
"learning_rate": 0.00015656833924247398,
|
| 1248 |
+
"loss": 0.0908,
|
| 1249 |
+
"step": 176
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.1419354838709677,
|
| 1253 |
+
"grad_norm": 0.16847128771716718,
|
| 1254 |
+
"learning_rate": 0.00015594697992592232,
|
| 1255 |
+
"loss": 0.0697,
|
| 1256 |
+
"step": 177
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.1483870967741936,
|
| 1260 |
+
"grad_norm": 0.14044376651199897,
|
| 1261 |
+
"learning_rate": 0.00015532246035685756,
|
| 1262 |
+
"loss": 0.0707,
|
| 1263 |
+
"step": 178
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.1548387096774193,
|
| 1267 |
+
"grad_norm": 0.13752732543534765,
|
| 1268 |
+
"learning_rate": 0.00015469481581224272,
|
| 1269 |
+
"loss": 0.0769,
|
| 1270 |
+
"step": 179
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 1.1612903225806452,
|
| 1274 |
+
"grad_norm": 0.20084535405957962,
|
| 1275 |
+
"learning_rate": 0.00015406408174555976,
|
| 1276 |
+
"loss": 0.1114,
|
| 1277 |
+
"step": 180
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.167741935483871,
|
| 1281 |
+
"grad_norm": 0.16882912932677738,
|
| 1282 |
+
"learning_rate": 0.0001534302937848073,
|
| 1283 |
+
"loss": 0.0977,
|
| 1284 |
+
"step": 181
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.1741935483870969,
|
| 1288 |
+
"grad_norm": 0.16088758279122345,
|
| 1289 |
+
"learning_rate": 0.00015279348773048786,
|
| 1290 |
+
"loss": 0.0862,
|
| 1291 |
+
"step": 182
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.1806451612903226,
|
| 1295 |
+
"grad_norm": 0.1610420842518173,
|
| 1296 |
+
"learning_rate": 0.00015215369955358566,
|
| 1297 |
+
"loss": 0.0909,
|
| 1298 |
+
"step": 183
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.1870967741935483,
|
| 1302 |
+
"grad_norm": 0.1567858759713509,
|
| 1303 |
+
"learning_rate": 0.0001515109653935348,
|
| 1304 |
+
"loss": 0.0988,
|
| 1305 |
+
"step": 184
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.1935483870967742,
|
| 1309 |
+
"grad_norm": 0.12400211503217344,
|
| 1310 |
+
"learning_rate": 0.00015086532155617784,
|
| 1311 |
+
"loss": 0.0813,
|
| 1312 |
+
"step": 185
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 1.2,
|
| 1316 |
+
"grad_norm": 0.12785181430507606,
|
| 1317 |
+
"learning_rate": 0.00015021680451171498,
|
| 1318 |
+
"loss": 0.0802,
|
| 1319 |
+
"step": 186
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.206451612903226,
|
| 1323 |
+
"grad_norm": 0.14298258948499543,
|
| 1324 |
+
"learning_rate": 0.00014956545089264407,
|
| 1325 |
+
"loss": 0.0843,
|
| 1326 |
+
"step": 187
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 1.2129032258064516,
|
| 1330 |
+
"grad_norm": 0.20148810623427083,
|
| 1331 |
+
"learning_rate": 0.0001489112974916912,
|
| 1332 |
+
"loss": 0.0942,
|
| 1333 |
+
"step": 188
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.2193548387096773,
|
| 1337 |
+
"grad_norm": 0.14657915199625932,
|
| 1338 |
+
"learning_rate": 0.00014825438125973264,
|
| 1339 |
+
"loss": 0.0829,
|
| 1340 |
+
"step": 189
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.2258064516129032,
|
| 1344 |
+
"grad_norm": 0.16147059815020365,
|
| 1345 |
+
"learning_rate": 0.00014759473930370736,
|
| 1346 |
+
"loss": 0.0756,
|
| 1347 |
+
"step": 190
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 1.232258064516129,
|
| 1351 |
+
"grad_norm": 0.17956881851269735,
|
| 1352 |
+
"learning_rate": 0.0001469324088845212,
|
| 1353 |
+
"loss": 0.1007,
|
| 1354 |
+
"step": 191
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 1.238709677419355,
|
| 1358 |
+
"grad_norm": 0.1578318532174886,
|
| 1359 |
+
"learning_rate": 0.00014626742741494206,
|
| 1360 |
+
"loss": 0.0983,
|
| 1361 |
+
"step": 192
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 1.2451612903225806,
|
| 1365 |
+
"grad_norm": 0.1649792762044239,
|
| 1366 |
+
"learning_rate": 0.00014559983245748638,
|
| 1367 |
+
"loss": 0.0905,
|
| 1368 |
+
"step": 193
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 1.2516129032258063,
|
| 1372 |
+
"grad_norm": 0.1455653603561888,
|
| 1373 |
+
"learning_rate": 0.00014492966172229777,
|
| 1374 |
+
"loss": 0.0791,
|
| 1375 |
+
"step": 194
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.2580645161290323,
|
| 1379 |
+
"grad_norm": 0.1312841252501999,
|
| 1380 |
+
"learning_rate": 0.00014425695306501658,
|
| 1381 |
+
"loss": 0.0613,
|
| 1382 |
+
"step": 195
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.2645161290322582,
|
| 1386 |
+
"grad_norm": 0.1361347125426188,
|
| 1387 |
+
"learning_rate": 0.00014358174448464154,
|
| 1388 |
+
"loss": 0.0773,
|
| 1389 |
+
"step": 196
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.270967741935484,
|
| 1393 |
+
"grad_norm": 0.17551022044504175,
|
| 1394 |
+
"learning_rate": 0.00014290407412138366,
|
| 1395 |
+
"loss": 0.0985,
|
| 1396 |
+
"step": 197
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 1.2774193548387096,
|
| 1400 |
+
"grad_norm": 0.1456899710914213,
|
| 1401 |
+
"learning_rate": 0.00014222398025451135,
|
| 1402 |
+
"loss": 0.0827,
|
| 1403 |
+
"step": 198
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.2838709677419355,
|
| 1407 |
+
"grad_norm": 0.19482064326323745,
|
| 1408 |
+
"learning_rate": 0.00014154150130018866,
|
| 1409 |
+
"loss": 0.0974,
|
| 1410 |
+
"step": 199
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 1.2903225806451613,
|
| 1414 |
+
"grad_norm": 0.13544624920378107,
|
| 1415 |
+
"learning_rate": 0.0001408566758093048,
|
| 1416 |
+
"loss": 0.0742,
|
| 1417 |
+
"step": 200
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.2967741935483872,
|
| 1421 |
+
"grad_norm": 0.12685823273525554,
|
| 1422 |
+
"learning_rate": 0.00014016954246529696,
|
| 1423 |
+
"loss": 0.0677,
|
| 1424 |
+
"step": 201
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.303225806451613,
|
| 1428 |
+
"grad_norm": 0.15457862761329938,
|
| 1429 |
+
"learning_rate": 0.00013948014008196487,
|
| 1430 |
+
"loss": 0.0815,
|
| 1431 |
+
"step": 202
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 1.3096774193548386,
|
| 1435 |
+
"grad_norm": 0.1482256655702007,
|
| 1436 |
+
"learning_rate": 0.0001387885076012785,
|
| 1437 |
+
"loss": 0.0961,
|
| 1438 |
+
"step": 203
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.3161290322580645,
|
| 1442 |
+
"grad_norm": 0.20896278322687534,
|
| 1443 |
+
"learning_rate": 0.00013809468409117846,
|
| 1444 |
+
"loss": 0.1049,
|
| 1445 |
+
"step": 204
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 1.3225806451612903,
|
| 1449 |
+
"grad_norm": 0.175019480492322,
|
| 1450 |
+
"learning_rate": 0.00013739870874336898,
|
| 1451 |
+
"loss": 0.0884,
|
| 1452 |
+
"step": 205
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 1.3290322580645162,
|
| 1456 |
+
"grad_norm": 0.14363356864891735,
|
| 1457 |
+
"learning_rate": 0.00013670062087110422,
|
| 1458 |
+
"loss": 0.0864,
|
| 1459 |
+
"step": 206
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.335483870967742,
|
| 1463 |
+
"grad_norm": 0.1397081258858762,
|
| 1464 |
+
"learning_rate": 0.00013600045990696762,
|
| 1465 |
+
"loss": 0.0797,
|
| 1466 |
+
"step": 207
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.3419354838709676,
|
| 1470 |
+
"grad_norm": 0.1667674455506937,
|
| 1471 |
+
"learning_rate": 0.0001352982654006444,
|
| 1472 |
+
"loss": 0.0977,
|
| 1473 |
+
"step": 208
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 1.3483870967741935,
|
| 1477 |
+
"grad_norm": 0.1409898972781143,
|
| 1478 |
+
"learning_rate": 0.00013459407701668763,
|
| 1479 |
+
"loss": 0.0923,
|
| 1480 |
+
"step": 209
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.3548387096774195,
|
| 1484 |
+
"grad_norm": 0.16290520924324786,
|
| 1485 |
+
"learning_rate": 0.00013388793453227767,
|
| 1486 |
+
"loss": 0.1033,
|
| 1487 |
+
"step": 210
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 1.3612903225806452,
|
| 1491 |
+
"grad_norm": 0.14967784709530377,
|
| 1492 |
+
"learning_rate": 0.0001331798778349752,
|
| 1493 |
+
"loss": 0.0843,
|
| 1494 |
+
"step": 211
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.367741935483871,
|
| 1498 |
+
"grad_norm": 0.14370370779202904,
|
| 1499 |
+
"learning_rate": 0.00013246994692046836,
|
| 1500 |
+
"loss": 0.0845,
|
| 1501 |
+
"step": 212
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.3741935483870968,
|
| 1505 |
+
"grad_norm": 0.15288920332301245,
|
| 1506 |
+
"learning_rate": 0.00013175818189031327,
|
| 1507 |
+
"loss": 0.0962,
|
| 1508 |
+
"step": 213
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.3806451612903226,
|
| 1512 |
+
"grad_norm": 0.16907810212754584,
|
| 1513 |
+
"learning_rate": 0.00013104462294966896,
|
| 1514 |
+
"loss": 0.1033,
|
| 1515 |
+
"step": 214
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 1.3870967741935485,
|
| 1519 |
+
"grad_norm": 0.17023880837738012,
|
| 1520 |
+
"learning_rate": 0.00013032931040502627,
|
| 1521 |
+
"loss": 0.1155,
|
| 1522 |
+
"step": 215
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.3935483870967742,
|
| 1526 |
+
"grad_norm": 0.12189251069561345,
|
| 1527 |
+
"learning_rate": 0.00012961228466193116,
|
| 1528 |
+
"loss": 0.0692,
|
| 1529 |
+
"step": 216
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 1.4,
|
| 1533 |
+
"grad_norm": 0.18108512316647296,
|
| 1534 |
+
"learning_rate": 0.00012889358622270223,
|
| 1535 |
+
"loss": 0.086,
|
| 1536 |
+
"step": 217
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 1.4064516129032258,
|
| 1540 |
+
"grad_norm": 0.15804455793477432,
|
| 1541 |
+
"learning_rate": 0.00012817325568414297,
|
| 1542 |
+
"loss": 0.0751,
|
| 1543 |
+
"step": 218
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.4129032258064516,
|
| 1547 |
+
"grad_norm": 0.13668879316730062,
|
| 1548 |
+
"learning_rate": 0.00012745133373524853,
|
| 1549 |
+
"loss": 0.0786,
|
| 1550 |
+
"step": 219
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 1.4193548387096775,
|
| 1554 |
+
"grad_norm": 0.18610949095399393,
|
| 1555 |
+
"learning_rate": 0.0001267278611549073,
|
| 1556 |
+
"loss": 0.0732,
|
| 1557 |
+
"step": 220
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.4258064516129032,
|
| 1561 |
+
"grad_norm": 0.186875270696186,
|
| 1562 |
+
"learning_rate": 0.00012600287880959763,
|
| 1563 |
+
"loss": 0.1051,
|
| 1564 |
+
"step": 221
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.432258064516129,
|
| 1568 |
+
"grad_norm": 0.15558801012557805,
|
| 1569 |
+
"learning_rate": 0.0001252764276510792,
|
| 1570 |
+
"loss": 0.0879,
|
| 1571 |
+
"step": 222
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.4387096774193548,
|
| 1575 |
+
"grad_norm": 0.16188457902393685,
|
| 1576 |
+
"learning_rate": 0.00012454854871407994,
|
| 1577 |
+
"loss": 0.0887,
|
| 1578 |
+
"step": 223
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 1.4451612903225808,
|
| 1582 |
+
"grad_norm": 0.1332789968563102,
|
| 1583 |
+
"learning_rate": 0.00012381928311397806,
|
| 1584 |
+
"loss": 0.0819,
|
| 1585 |
+
"step": 224
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.4516129032258065,
|
| 1589 |
+
"grad_norm": 0.13104439295337186,
|
| 1590 |
+
"learning_rate": 0.0001230886720444796,
|
| 1591 |
+
"loss": 0.0992,
|
| 1592 |
+
"step": 225
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 1.4580645161290322,
|
| 1596 |
+
"grad_norm": 0.1741921210517873,
|
| 1597 |
+
"learning_rate": 0.00012235675677529158,
|
| 1598 |
+
"loss": 0.1049,
|
| 1599 |
+
"step": 226
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.4645161290322581,
|
| 1603 |
+
"grad_norm": 0.17686059237450052,
|
| 1604 |
+
"learning_rate": 0.00012162357864979072,
|
| 1605 |
+
"loss": 0.0942,
|
| 1606 |
+
"step": 227
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.4709677419354839,
|
| 1610 |
+
"grad_norm": 0.14208255193217198,
|
| 1611 |
+
"learning_rate": 0.00012088917908268821,
|
| 1612 |
+
"loss": 0.0932,
|
| 1613 |
+
"step": 228
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.4774193548387098,
|
| 1617 |
+
"grad_norm": 0.16207984639669018,
|
| 1618 |
+
"learning_rate": 0.00012015359955769021,
|
| 1619 |
+
"loss": 0.0952,
|
| 1620 |
+
"step": 229
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 1.4838709677419355,
|
| 1624 |
+
"grad_norm": 0.10567516069189406,
|
| 1625 |
+
"learning_rate": 0.00011941688162515467,
|
| 1626 |
+
"loss": 0.0698,
|
| 1627 |
+
"step": 230
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.4903225806451612,
|
| 1631 |
+
"grad_norm": 0.14315458863752037,
|
| 1632 |
+
"learning_rate": 0.00011867906689974428,
|
| 1633 |
+
"loss": 0.0706,
|
| 1634 |
+
"step": 231
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 1.4967741935483871,
|
| 1638 |
+
"grad_norm": 0.14644874259664967,
|
| 1639 |
+
"learning_rate": 0.00011794019705807584,
|
| 1640 |
+
"loss": 0.0954,
|
| 1641 |
+
"step": 232
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.5032258064516129,
|
| 1645 |
+
"grad_norm": 0.1522258926853921,
|
| 1646 |
+
"learning_rate": 0.00011720031383636585,
|
| 1647 |
+
"loss": 0.1026,
|
| 1648 |
+
"step": 233
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.5096774193548388,
|
| 1652 |
+
"grad_norm": 0.2389589012648026,
|
| 1653 |
+
"learning_rate": 0.00011645945902807341,
|
| 1654 |
+
"loss": 0.0981,
|
| 1655 |
+
"step": 234
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.5161290322580645,
|
| 1659 |
+
"grad_norm": 0.1539160607687386,
|
| 1660 |
+
"learning_rate": 0.00011571767448153901,
|
| 1661 |
+
"loss": 0.074,
|
| 1662 |
+
"step": 235
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.5225806451612902,
|
| 1666 |
+
"grad_norm": 0.17340052794011998,
|
| 1667 |
+
"learning_rate": 0.00011497500209762102,
|
| 1668 |
+
"loss": 0.0943,
|
| 1669 |
+
"step": 236
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.5290322580645161,
|
| 1673 |
+
"grad_norm": 0.11294207936842038,
|
| 1674 |
+
"learning_rate": 0.00011423148382732853,
|
| 1675 |
+
"loss": 0.0644,
|
| 1676 |
+
"step": 237
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 1.535483870967742,
|
| 1680 |
+
"grad_norm": 0.14391081206665055,
|
| 1681 |
+
"learning_rate": 0.00011348716166945195,
|
| 1682 |
+
"loss": 0.0774,
|
| 1683 |
+
"step": 238
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 1.5419354838709678,
|
| 1687 |
+
"grad_norm": 0.16298875474891092,
|
| 1688 |
+
"learning_rate": 0.0001127420776681905,
|
| 1689 |
+
"loss": 0.0786,
|
| 1690 |
+
"step": 239
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.5483870967741935,
|
| 1694 |
+
"grad_norm": 0.1325076119120941,
|
| 1695 |
+
"learning_rate": 0.00011199627391077732,
|
| 1696 |
+
"loss": 0.0889,
|
| 1697 |
+
"step": 240
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.5548387096774192,
|
| 1701 |
+
"grad_norm": 0.10971482413094012,
|
| 1702 |
+
"learning_rate": 0.00011124979252510208,
|
| 1703 |
+
"loss": 0.0737,
|
| 1704 |
+
"step": 241
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 1.5612903225806452,
|
| 1708 |
+
"grad_norm": 0.14614299703062,
|
| 1709 |
+
"learning_rate": 0.0001105026756773314,
|
| 1710 |
+
"loss": 0.0682,
|
| 1711 |
+
"step": 242
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 1.567741935483871,
|
| 1715 |
+
"grad_norm": 0.22690104820583093,
|
| 1716 |
+
"learning_rate": 0.00010975496556952682,
|
| 1717 |
+
"loss": 0.1094,
|
| 1718 |
+
"step": 243
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 1.5741935483870968,
|
| 1722 |
+
"grad_norm": 0.2561727438038473,
|
| 1723 |
+
"learning_rate": 0.00010900670443726135,
|
| 1724 |
+
"loss": 0.0851,
|
| 1725 |
+
"step": 244
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 1.5806451612903225,
|
| 1729 |
+
"grad_norm": 0.15371305701947427,
|
| 1730 |
+
"learning_rate": 0.00010825793454723325,
|
| 1731 |
+
"loss": 0.0923,
|
| 1732 |
+
"step": 245
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 1.5870967741935482,
|
| 1736 |
+
"grad_norm": 0.16267574566743875,
|
| 1737 |
+
"learning_rate": 0.00010750869819487883,
|
| 1738 |
+
"loss": 0.1036,
|
| 1739 |
+
"step": 246
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.5935483870967742,
|
| 1743 |
+
"grad_norm": 0.15591155698404394,
|
| 1744 |
+
"learning_rate": 0.00010675903770198333,
|
| 1745 |
+
"loss": 0.0893,
|
| 1746 |
+
"step": 247
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 1.6,
|
| 1750 |
+
"grad_norm": 0.14338972962339533,
|
| 1751 |
+
"learning_rate": 0.00010600899541429004,
|
| 1752 |
+
"loss": 0.0837,
|
| 1753 |
+
"step": 248
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.6064516129032258,
|
| 1757 |
+
"grad_norm": 0.12387607320751257,
|
| 1758 |
+
"learning_rate": 0.00010525861369910877,
|
| 1759 |
+
"loss": 0.0755,
|
| 1760 |
+
"step": 249
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 1.6129032258064515,
|
| 1764 |
+
"grad_norm": 0.16606169294386383,
|
| 1765 |
+
"learning_rate": 0.00010450793494292224,
|
| 1766 |
+
"loss": 0.1043,
|
| 1767 |
+
"step": 250
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 1.6193548387096774,
|
| 1771 |
+
"grad_norm": 0.1795920159350681,
|
| 1772 |
+
"learning_rate": 0.00010375700154899208,
|
| 1773 |
+
"loss": 0.1008,
|
| 1774 |
+
"step": 251
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 1.6258064516129034,
|
| 1778 |
+
"grad_norm": 0.13025097291519463,
|
| 1779 |
+
"learning_rate": 0.00010300585593496348,
|
| 1780 |
+
"loss": 0.0851,
|
| 1781 |
+
"step": 252
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.632258064516129,
|
| 1785 |
+
"grad_norm": 0.14349816154023654,
|
| 1786 |
+
"learning_rate": 0.00010225454053046921,
|
| 1787 |
+
"loss": 0.0807,
|
| 1788 |
+
"step": 253
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 1.6387096774193548,
|
| 1792 |
+
"grad_norm": 0.16695448888874226,
|
| 1793 |
+
"learning_rate": 0.00010150309777473306,
|
| 1794 |
+
"loss": 0.1117,
|
| 1795 |
+
"step": 254
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 1.6451612903225805,
|
| 1799 |
+
"grad_norm": 0.15743712528269815,
|
| 1800 |
+
"learning_rate": 0.0001007515701141722,
|
| 1801 |
+
"loss": 0.086,
|
| 1802 |
+
"step": 255
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 1.6516129032258065,
|
| 1806 |
+
"grad_norm": 0.16704335210894908,
|
| 1807 |
+
"learning_rate": 0.0001,
|
| 1808 |
+
"loss": 0.0886,
|
| 1809 |
+
"step": 256
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 1.6580645161290324,
|
| 1813 |
+
"grad_norm": 0.10245531520994122,
|
| 1814 |
+
"learning_rate": 9.924842988582782e-05,
|
| 1815 |
+
"loss": 0.0678,
|
| 1816 |
+
"step": 257
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 1.664516129032258,
|
| 1820 |
+
"grad_norm": 0.1688033836669086,
|
| 1821 |
+
"learning_rate": 9.849690222526698e-05,
|
| 1822 |
+
"loss": 0.0958,
|
| 1823 |
+
"step": 258
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 1.6709677419354838,
|
| 1827 |
+
"grad_norm": 0.1338126992775965,
|
| 1828 |
+
"learning_rate": 9.77454594695308e-05,
|
| 1829 |
+
"loss": 0.0782,
|
| 1830 |
+
"step": 259
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 1.6774193548387095,
|
| 1834 |
+
"grad_norm": 0.1332820451604909,
|
| 1835 |
+
"learning_rate": 9.699414406503654e-05,
|
| 1836 |
+
"loss": 0.0856,
|
| 1837 |
+
"step": 260
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 1.6838709677419355,
|
| 1841 |
+
"grad_norm": 0.1449401720605745,
|
| 1842 |
+
"learning_rate": 9.624299845100795e-05,
|
| 1843 |
+
"loss": 0.0882,
|
| 1844 |
+
"step": 261
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 1.6903225806451614,
|
| 1848 |
+
"grad_norm": 0.19395703151279187,
|
| 1849 |
+
"learning_rate": 9.549206505707777e-05,
|
| 1850 |
+
"loss": 0.1148,
|
| 1851 |
+
"step": 262
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 1.696774193548387,
|
| 1855 |
+
"grad_norm": 0.13482277559178169,
|
| 1856 |
+
"learning_rate": 9.474138630089124e-05,
|
| 1857 |
+
"loss": 0.0644,
|
| 1858 |
+
"step": 263
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 1.7032258064516128,
|
| 1862 |
+
"grad_norm": 0.18530593840863338,
|
| 1863 |
+
"learning_rate": 9.399100458570997e-05,
|
| 1864 |
+
"loss": 0.1074,
|
| 1865 |
+
"step": 264
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 1.7096774193548387,
|
| 1869 |
+
"grad_norm": 0.1897995033582595,
|
| 1870 |
+
"learning_rate": 9.324096229801674e-05,
|
| 1871 |
+
"loss": 0.0867,
|
| 1872 |
+
"step": 265
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 1.7161290322580647,
|
| 1876 |
+
"grad_norm": 0.16472780150681127,
|
| 1877 |
+
"learning_rate": 9.249130180512118e-05,
|
| 1878 |
+
"loss": 0.0896,
|
| 1879 |
+
"step": 266
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 1.7225806451612904,
|
| 1883 |
+
"grad_norm": 0.15525608679066774,
|
| 1884 |
+
"learning_rate": 9.174206545276677e-05,
|
| 1885 |
+
"loss": 0.0865,
|
| 1886 |
+
"step": 267
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 1.729032258064516,
|
| 1890 |
+
"grad_norm": 0.15426999703424252,
|
| 1891 |
+
"learning_rate": 9.099329556273866e-05,
|
| 1892 |
+
"loss": 0.0853,
|
| 1893 |
+
"step": 268
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 1.7354838709677418,
|
| 1897 |
+
"grad_norm": 0.19103115451320254,
|
| 1898 |
+
"learning_rate": 9.024503443047319e-05,
|
| 1899 |
+
"loss": 0.0993,
|
| 1900 |
+
"step": 269
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 1.7419354838709677,
|
| 1904 |
+
"grad_norm": 0.12323460303269068,
|
| 1905 |
+
"learning_rate": 8.949732432266866e-05,
|
| 1906 |
+
"loss": 0.0723,
|
| 1907 |
+
"step": 270
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 1.7483870967741937,
|
| 1911 |
+
"grad_norm": 0.1339668030381976,
|
| 1912 |
+
"learning_rate": 8.875020747489794e-05,
|
| 1913 |
+
"loss": 0.0852,
|
| 1914 |
+
"step": 271
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 1.7548387096774194,
|
| 1918 |
+
"grad_norm": 0.1385020226698345,
|
| 1919 |
+
"learning_rate": 8.800372608922271e-05,
|
| 1920 |
+
"loss": 0.0773,
|
| 1921 |
+
"step": 272
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 1.761290322580645,
|
| 1925 |
+
"grad_norm": 0.21835470061774626,
|
| 1926 |
+
"learning_rate": 8.72579223318095e-05,
|
| 1927 |
+
"loss": 0.1167,
|
| 1928 |
+
"step": 273
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 1.7677419354838708,
|
| 1932 |
+
"grad_norm": 0.1425322245524885,
|
| 1933 |
+
"learning_rate": 8.651283833054809e-05,
|
| 1934 |
+
"loss": 0.0801,
|
| 1935 |
+
"step": 274
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.7741935483870968,
|
| 1939 |
+
"grad_norm": 0.1794495686810833,
|
| 1940 |
+
"learning_rate": 8.57685161726715e-05,
|
| 1941 |
+
"loss": 0.0784,
|
| 1942 |
+
"step": 275
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 1.7806451612903227,
|
| 1946 |
+
"grad_norm": 0.15688310999446023,
|
| 1947 |
+
"learning_rate": 8.5024997902379e-05,
|
| 1948 |
+
"loss": 0.1001,
|
| 1949 |
+
"step": 276
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 1.7870967741935484,
|
| 1953 |
+
"grad_norm": 0.15417421931142297,
|
| 1954 |
+
"learning_rate": 8.428232551846101e-05,
|
| 1955 |
+
"loss": 0.0898,
|
| 1956 |
+
"step": 277
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 1.793548387096774,
|
| 1960 |
+
"grad_norm": 0.17958092903510822,
|
| 1961 |
+
"learning_rate": 8.35405409719266e-05,
|
| 1962 |
+
"loss": 0.0921,
|
| 1963 |
+
"step": 278
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 1.8,
|
| 1967 |
+
"grad_norm": 0.165061388870699,
|
| 1968 |
+
"learning_rate": 8.279968616363418e-05,
|
| 1969 |
+
"loss": 0.0935,
|
| 1970 |
+
"step": 279
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 1.8064516129032258,
|
| 1974 |
+
"grad_norm": 0.1116000947584692,
|
| 1975 |
+
"learning_rate": 8.205980294192421e-05,
|
| 1976 |
+
"loss": 0.0684,
|
| 1977 |
+
"step": 280
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 1.8129032258064517,
|
| 1981 |
+
"grad_norm": 0.12370745355237868,
|
| 1982 |
+
"learning_rate": 8.132093310025571e-05,
|
| 1983 |
+
"loss": 0.0764,
|
| 1984 |
+
"step": 281
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 1.8193548387096774,
|
| 1988 |
+
"grad_norm": 0.1724137684896053,
|
| 1989 |
+
"learning_rate": 8.058311837484535e-05,
|
| 1990 |
+
"loss": 0.0969,
|
| 1991 |
+
"step": 282
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 1.8258064516129031,
|
| 1995 |
+
"grad_norm": 0.16125193051943912,
|
| 1996 |
+
"learning_rate": 7.984640044230983e-05,
|
| 1997 |
+
"loss": 0.0868,
|
| 1998 |
+
"step": 283
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 1.832258064516129,
|
| 2002 |
+
"grad_norm": 0.13929471780512187,
|
| 2003 |
+
"learning_rate": 7.911082091731181e-05,
|
| 2004 |
+
"loss": 0.0701,
|
| 2005 |
+
"step": 284
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 1.838709677419355,
|
| 2009 |
+
"grad_norm": 0.1172859356132756,
|
| 2010 |
+
"learning_rate": 7.837642135020929e-05,
|
| 2011 |
+
"loss": 0.0705,
|
| 2012 |
+
"step": 285
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 1.8451612903225807,
|
| 2016 |
+
"grad_norm": 0.1416371336298496,
|
| 2017 |
+
"learning_rate": 7.764324322470841e-05,
|
| 2018 |
+
"loss": 0.0683,
|
| 2019 |
+
"step": 286
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 1.8516129032258064,
|
| 2023 |
+
"grad_norm": 0.19198949443360017,
|
| 2024 |
+
"learning_rate": 7.691132795552043e-05,
|
| 2025 |
+
"loss": 0.0894,
|
| 2026 |
+
"step": 287
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.8580645161290321,
|
| 2030 |
+
"grad_norm": 0.19384526534149363,
|
| 2031 |
+
"learning_rate": 7.618071688602199e-05,
|
| 2032 |
+
"loss": 0.0954,
|
| 2033 |
+
"step": 288
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 1.864516129032258,
|
| 2037 |
+
"grad_norm": 0.21508649416468206,
|
| 2038 |
+
"learning_rate": 7.54514512859201e-05,
|
| 2039 |
+
"loss": 0.0986,
|
| 2040 |
+
"step": 289
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 1.870967741935484,
|
| 2044 |
+
"grad_norm": 0.1952670486171591,
|
| 2045 |
+
"learning_rate": 7.472357234892082e-05,
|
| 2046 |
+
"loss": 0.1049,
|
| 2047 |
+
"step": 290
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 1.8774193548387097,
|
| 2051 |
+
"grad_norm": 0.13259304301639457,
|
| 2052 |
+
"learning_rate": 7.399712119040238e-05,
|
| 2053 |
+
"loss": 0.0784,
|
| 2054 |
+
"step": 291
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 1.8838709677419354,
|
| 2058 |
+
"grad_norm": 0.1367442063153438,
|
| 2059 |
+
"learning_rate": 7.327213884509272e-05,
|
| 2060 |
+
"loss": 0.0828,
|
| 2061 |
+
"step": 292
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 1.8903225806451613,
|
| 2065 |
+
"grad_norm": 0.16389743747526248,
|
| 2066 |
+
"learning_rate": 7.254866626475152e-05,
|
| 2067 |
+
"loss": 0.0927,
|
| 2068 |
+
"step": 293
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 1.896774193548387,
|
| 2072 |
+
"grad_norm": 0.22089121957380548,
|
| 2073 |
+
"learning_rate": 7.182674431585704e-05,
|
| 2074 |
+
"loss": 0.1077,
|
| 2075 |
+
"step": 294
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 1.903225806451613,
|
| 2079 |
+
"grad_norm": 0.14882797154919192,
|
| 2080 |
+
"learning_rate": 7.110641377729778e-05,
|
| 2081 |
+
"loss": 0.0879,
|
| 2082 |
+
"step": 295
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 1.9096774193548387,
|
| 2086 |
+
"grad_norm": 0.12794357419440644,
|
| 2087 |
+
"learning_rate": 7.038771533806884e-05,
|
| 2088 |
+
"loss": 0.0719,
|
| 2089 |
+
"step": 296
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 1.9161290322580644,
|
| 2093 |
+
"grad_norm": 0.1846824594193936,
|
| 2094 |
+
"learning_rate": 6.967068959497376e-05,
|
| 2095 |
+
"loss": 0.0895,
|
| 2096 |
+
"step": 297
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 1.9225806451612903,
|
| 2100 |
+
"grad_norm": 0.17646276082900025,
|
| 2101 |
+
"learning_rate": 6.895537705033108e-05,
|
| 2102 |
+
"loss": 0.0996,
|
| 2103 |
+
"step": 298
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 1.9290322580645163,
|
| 2107 |
+
"grad_norm": 0.15604696338212207,
|
| 2108 |
+
"learning_rate": 6.824181810968675e-05,
|
| 2109 |
+
"loss": 0.087,
|
| 2110 |
+
"step": 299
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 1.935483870967742,
|
| 2114 |
+
"grad_norm": 0.15478000466177547,
|
| 2115 |
+
"learning_rate": 6.753005307953167e-05,
|
| 2116 |
+
"loss": 0.0831,
|
| 2117 |
+
"step": 300
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.9419354838709677,
|
| 2121 |
+
"grad_norm": 0.19642059718505833,
|
| 2122 |
+
"learning_rate": 6.682012216502484e-05,
|
| 2123 |
+
"loss": 0.1073,
|
| 2124 |
+
"step": 301
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 1.9483870967741934,
|
| 2128 |
+
"grad_norm": 0.17255606883063224,
|
| 2129 |
+
"learning_rate": 6.611206546772237e-05,
|
| 2130 |
+
"loss": 0.1018,
|
| 2131 |
+
"step": 302
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 1.9548387096774194,
|
| 2135 |
+
"grad_norm": 0.1393602083174125,
|
| 2136 |
+
"learning_rate": 6.54059229833124e-05,
|
| 2137 |
+
"loss": 0.0859,
|
| 2138 |
+
"step": 303
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 1.9612903225806453,
|
| 2142 |
+
"grad_norm": 0.16652643955266602,
|
| 2143 |
+
"learning_rate": 6.47017345993556e-05,
|
| 2144 |
+
"loss": 0.0899,
|
| 2145 |
+
"step": 304
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 1.967741935483871,
|
| 2149 |
+
"grad_norm": 0.16878660937501272,
|
| 2150 |
+
"learning_rate": 6.39995400930324e-05,
|
| 2151 |
+
"loss": 0.0998,
|
| 2152 |
+
"step": 305
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 1.9741935483870967,
|
| 2156 |
+
"grad_norm": 0.12969061293334938,
|
| 2157 |
+
"learning_rate": 6.329937912889582e-05,
|
| 2158 |
+
"loss": 0.0706,
|
| 2159 |
+
"step": 306
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 1.9806451612903224,
|
| 2163 |
+
"grad_norm": 0.15739894009886643,
|
| 2164 |
+
"learning_rate": 6.260129125663106e-05,
|
| 2165 |
+
"loss": 0.0888,
|
| 2166 |
+
"step": 307
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 1.9870967741935484,
|
| 2170 |
+
"grad_norm": 0.1752067977727585,
|
| 2171 |
+
"learning_rate": 6.190531590882159e-05,
|
| 2172 |
+
"loss": 0.0826,
|
| 2173 |
+
"step": 308
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 1.9935483870967743,
|
| 2177 |
+
"grad_norm": 0.21002280616145647,
|
| 2178 |
+
"learning_rate": 6.121149239872151e-05,
|
| 2179 |
+
"loss": 0.123,
|
| 2180 |
+
"step": 309
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 2.0,
|
| 2184 |
+
"grad_norm": 0.13850366944373244,
|
| 2185 |
+
"learning_rate": 6.051985991803517e-05,
|
| 2186 |
+
"loss": 0.073,
|
| 2187 |
+
"step": 310
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 2.0,
|
| 2191 |
+
"eval_loss": 0.16690203547477722,
|
| 2192 |
+
"eval_runtime": 25.4682,
|
| 2193 |
+
"eval_samples_per_second": 5.144,
|
| 2194 |
+
"eval_steps_per_second": 0.668,
|
| 2195 |
+
"step": 310
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 2.0064516129032257,
|
| 2199 |
+
"grad_norm": 0.1124990090834071,
|
| 2200 |
+
"learning_rate": 5.983045753470308e-05,
|
| 2201 |
+
"loss": 0.0485,
|
| 2202 |
+
"step": 311
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 2.0129032258064514,
|
| 2206 |
+
"grad_norm": 0.08246086368829486,
|
| 2207 |
+
"learning_rate": 5.9143324190695196e-05,
|
| 2208 |
+
"loss": 0.0411,
|
| 2209 |
+
"step": 312
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 2.0193548387096776,
|
| 2213 |
+
"grad_norm": 0.11824209839228844,
|
| 2214 |
+
"learning_rate": 5.845849869981137e-05,
|
| 2215 |
+
"loss": 0.0578,
|
| 2216 |
+
"step": 313
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 2.0258064516129033,
|
| 2220 |
+
"grad_norm": 0.09183702414760068,
|
| 2221 |
+
"learning_rate": 5.777601974548866e-05,
|
| 2222 |
+
"loss": 0.0405,
|
| 2223 |
+
"step": 314
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 2.032258064516129,
|
| 2227 |
+
"grad_norm": 0.10138990147374953,
|
| 2228 |
+
"learning_rate": 5.709592587861637e-05,
|
| 2229 |
+
"loss": 0.045,
|
| 2230 |
+
"step": 315
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 2.0387096774193547,
|
| 2234 |
+
"grad_norm": 0.13174518232765414,
|
| 2235 |
+
"learning_rate": 5.6418255515358486e-05,
|
| 2236 |
+
"loss": 0.0417,
|
| 2237 |
+
"step": 316
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 2.0451612903225804,
|
| 2241 |
+
"grad_norm": 0.11134581768568716,
|
| 2242 |
+
"learning_rate": 5.574304693498346e-05,
|
| 2243 |
+
"loss": 0.0351,
|
| 2244 |
+
"step": 317
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 2.0516129032258066,
|
| 2248 |
+
"grad_norm": 0.10372734844639428,
|
| 2249 |
+
"learning_rate": 5.507033827770225e-05,
|
| 2250 |
+
"loss": 0.0458,
|
| 2251 |
+
"step": 318
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 2.0580645161290323,
|
| 2255 |
+
"grad_norm": 0.09893496819685892,
|
| 2256 |
+
"learning_rate": 5.4400167542513636e-05,
|
| 2257 |
+
"loss": 0.0421,
|
| 2258 |
+
"step": 319
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 2.064516129032258,
|
| 2262 |
+
"grad_norm": 0.1544849048513997,
|
| 2263 |
+
"learning_rate": 5.3732572585057974e-05,
|
| 2264 |
+
"loss": 0.0449,
|
| 2265 |
+
"step": 320
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 2.0709677419354837,
|
| 2269 |
+
"grad_norm": 0.12211629106983518,
|
| 2270 |
+
"learning_rate": 5.306759111547881e-05,
|
| 2271 |
+
"loss": 0.0341,
|
| 2272 |
+
"step": 321
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 2.07741935483871,
|
| 2276 |
+
"grad_norm": 0.1519039459170261,
|
| 2277 |
+
"learning_rate": 5.240526069629265e-05,
|
| 2278 |
+
"loss": 0.0477,
|
| 2279 |
+
"step": 322
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 2.0838709677419356,
|
| 2283 |
+
"grad_norm": 0.10882973801143421,
|
| 2284 |
+
"learning_rate": 5.174561874026741e-05,
|
| 2285 |
+
"loss": 0.0376,
|
| 2286 |
+
"step": 323
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 2.0903225806451613,
|
| 2290 |
+
"grad_norm": 0.09785322895617321,
|
| 2291 |
+
"learning_rate": 5.108870250830882e-05,
|
| 2292 |
+
"loss": 0.0305,
|
| 2293 |
+
"step": 324
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 2.096774193548387,
|
| 2297 |
+
"grad_norm": 0.14990301799413153,
|
| 2298 |
+
"learning_rate": 5.0434549107355944e-05,
|
| 2299 |
+
"loss": 0.0409,
|
| 2300 |
+
"step": 325
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 2.1032258064516127,
|
| 2304 |
+
"grad_norm": 0.1915279670852206,
|
| 2305 |
+
"learning_rate": 4.978319548828504e-05,
|
| 2306 |
+
"loss": 0.045,
|
| 2307 |
+
"step": 326
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 2.109677419354839,
|
| 2311 |
+
"grad_norm": 0.15445162243894608,
|
| 2312 |
+
"learning_rate": 4.9134678443822166e-05,
|
| 2313 |
+
"loss": 0.0334,
|
| 2314 |
+
"step": 327
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 2.1161290322580646,
|
| 2318 |
+
"grad_norm": 0.34562522599625123,
|
| 2319 |
+
"learning_rate": 4.8489034606465225e-05,
|
| 2320 |
+
"loss": 0.0395,
|
| 2321 |
+
"step": 328
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 2.1225806451612903,
|
| 2325 |
+
"grad_norm": 0.3183911745576579,
|
| 2326 |
+
"learning_rate": 4.784630044641435e-05,
|
| 2327 |
+
"loss": 0.0546,
|
| 2328 |
+
"step": 329
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 2.129032258064516,
|
| 2332 |
+
"grad_norm": 0.1619196710614726,
|
| 2333 |
+
"learning_rate": 4.7206512269512124e-05,
|
| 2334 |
+
"loss": 0.0446,
|
| 2335 |
+
"step": 330
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 2.135483870967742,
|
| 2339 |
+
"grad_norm": 0.24657637060341897,
|
| 2340 |
+
"learning_rate": 4.65697062151927e-05,
|
| 2341 |
+
"loss": 0.0349,
|
| 2342 |
+
"step": 331
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 2.141935483870968,
|
| 2346 |
+
"grad_norm": 0.11723613581448042,
|
| 2347 |
+
"learning_rate": 4.593591825444028e-05,
|
| 2348 |
+
"loss": 0.0337,
|
| 2349 |
+
"step": 332
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 2.1483870967741936,
|
| 2353 |
+
"grad_norm": 0.12413977056977685,
|
| 2354 |
+
"learning_rate": 4.530518418775733e-05,
|
| 2355 |
+
"loss": 0.0355,
|
| 2356 |
+
"step": 333
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 2.1548387096774193,
|
| 2360 |
+
"grad_norm": 0.1368012687546921,
|
| 2361 |
+
"learning_rate": 4.4677539643142454e-05,
|
| 2362 |
+
"loss": 0.0369,
|
| 2363 |
+
"step": 334
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 2.161290322580645,
|
| 2367 |
+
"grad_norm": 0.132092425970093,
|
| 2368 |
+
"learning_rate": 4.40530200740777e-05,
|
| 2369 |
+
"loss": 0.0346,
|
| 2370 |
+
"step": 335
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 2.167741935483871,
|
| 2374 |
+
"grad_norm": 0.17255652388608525,
|
| 2375 |
+
"learning_rate": 4.343166075752605e-05,
|
| 2376 |
+
"loss": 0.0425,
|
| 2377 |
+
"step": 336
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 2.174193548387097,
|
| 2381 |
+
"grad_norm": 0.10537625462278236,
|
| 2382 |
+
"learning_rate": 4.281349679193861e-05,
|
| 2383 |
+
"loss": 0.0363,
|
| 2384 |
+
"step": 337
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 2.1806451612903226,
|
| 2388 |
+
"grad_norm": 0.08446560264535163,
|
| 2389 |
+
"learning_rate": 4.2198563095272116e-05,
|
| 2390 |
+
"loss": 0.0303,
|
| 2391 |
+
"step": 338
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 2.1870967741935483,
|
| 2395 |
+
"grad_norm": 0.11155624490505092,
|
| 2396 |
+
"learning_rate": 4.158689440301657e-05,
|
| 2397 |
+
"loss": 0.0424,
|
| 2398 |
+
"step": 339
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 2.193548387096774,
|
| 2402 |
+
"grad_norm": 0.10614591478252769,
|
| 2403 |
+
"learning_rate": 4.097852526623307e-05,
|
| 2404 |
+
"loss": 0.0374,
|
| 2405 |
+
"step": 340
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.2,
|
| 2409 |
+
"grad_norm": 0.15380955207075353,
|
| 2410 |
+
"learning_rate": 4.0373490049602204e-05,
|
| 2411 |
+
"loss": 0.0506,
|
| 2412 |
+
"step": 341
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 2.206451612903226,
|
| 2416 |
+
"grad_norm": 0.11905414571153279,
|
| 2417 |
+
"learning_rate": 3.977182292948283e-05,
|
| 2418 |
+
"loss": 0.0517,
|
| 2419 |
+
"step": 342
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 2.2129032258064516,
|
| 2423 |
+
"grad_norm": 0.14425425935781702,
|
| 2424 |
+
"learning_rate": 3.9173557891981573e-05,
|
| 2425 |
+
"loss": 0.0467,
|
| 2426 |
+
"step": 343
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 2.2193548387096773,
|
| 2430 |
+
"grad_norm": 0.12453230757386329,
|
| 2431 |
+
"learning_rate": 3.857872873103322e-05,
|
| 2432 |
+
"loss": 0.0435,
|
| 2433 |
+
"step": 344
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 2.225806451612903,
|
| 2437 |
+
"grad_norm": 0.10038267160538758,
|
| 2438 |
+
"learning_rate": 3.7987369046491684e-05,
|
| 2439 |
+
"loss": 0.0387,
|
| 2440 |
+
"step": 345
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 2.232258064516129,
|
| 2444 |
+
"grad_norm": 0.1438025015993235,
|
| 2445 |
+
"learning_rate": 3.7399512242231995e-05,
|
| 2446 |
+
"loss": 0.0346,
|
| 2447 |
+
"step": 346
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.238709677419355,
|
| 2451 |
+
"grad_norm": 0.26072179652296745,
|
| 2452 |
+
"learning_rate": 3.6815191524263624e-05,
|
| 2453 |
+
"loss": 0.0545,
|
| 2454 |
+
"step": 347
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 2.2451612903225806,
|
| 2458 |
+
"grad_norm": 0.08770511437142381,
|
| 2459 |
+
"learning_rate": 3.623443989885462e-05,
|
| 2460 |
+
"loss": 0.0327,
|
| 2461 |
+
"step": 348
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 2.2516129032258063,
|
| 2465 |
+
"grad_norm": 0.12301181296016188,
|
| 2466 |
+
"learning_rate": 3.565729017066729e-05,
|
| 2467 |
+
"loss": 0.039,
|
| 2468 |
+
"step": 349
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 2.258064516129032,
|
| 2472 |
+
"grad_norm": 0.13814667341858822,
|
| 2473 |
+
"learning_rate": 3.508377494090521e-05,
|
| 2474 |
+
"loss": 0.0356,
|
| 2475 |
+
"step": 350
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 2.264516129032258,
|
| 2479 |
+
"grad_norm": 0.12413533495298362,
|
| 2480 |
+
"learning_rate": 3.45139266054715e-05,
|
| 2481 |
+
"loss": 0.0415,
|
| 2482 |
+
"step": 351
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 2.270967741935484,
|
| 2486 |
+
"grad_norm": 0.11031185879435731,
|
| 2487 |
+
"learning_rate": 3.394777735313919e-05,
|
| 2488 |
+
"loss": 0.0363,
|
| 2489 |
+
"step": 352
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 2.2774193548387096,
|
| 2493 |
+
"grad_norm": 0.11353520726841541,
|
| 2494 |
+
"learning_rate": 3.338535916373266e-05,
|
| 2495 |
+
"loss": 0.0376,
|
| 2496 |
+
"step": 353
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 2.2838709677419353,
|
| 2500 |
+
"grad_norm": 0.23000368159173218,
|
| 2501 |
+
"learning_rate": 3.2826703806321525e-05,
|
| 2502 |
+
"loss": 0.0444,
|
| 2503 |
+
"step": 354
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 2.2903225806451615,
|
| 2507 |
+
"grad_norm": 0.10390022981327145,
|
| 2508 |
+
"learning_rate": 3.227184283742591e-05,
|
| 2509 |
+
"loss": 0.0308,
|
| 2510 |
+
"step": 355
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 2.296774193548387,
|
| 2514 |
+
"grad_norm": 0.14881153611607173,
|
| 2515 |
+
"learning_rate": 3.17208075992339e-05,
|
| 2516 |
+
"loss": 0.0428,
|
| 2517 |
+
"step": 356
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 2.303225806451613,
|
| 2521 |
+
"grad_norm": 0.22847373170988944,
|
| 2522 |
+
"learning_rate": 3.117362921783134e-05,
|
| 2523 |
+
"loss": 0.0449,
|
| 2524 |
+
"step": 357
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 2.3096774193548386,
|
| 2528 |
+
"grad_norm": 0.1261660214676178,
|
| 2529 |
+
"learning_rate": 3.063033860144339e-05,
|
| 2530 |
+
"loss": 0.0354,
|
| 2531 |
+
"step": 358
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 2.3161290322580643,
|
| 2535 |
+
"grad_norm": 0.16508383182402034,
|
| 2536 |
+
"learning_rate": 3.0090966438688772e-05,
|
| 2537 |
+
"loss": 0.0404,
|
| 2538 |
+
"step": 359
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 2.3225806451612905,
|
| 2542 |
+
"grad_norm": 0.14003607495149023,
|
| 2543 |
+
"learning_rate": 2.9555543196846292e-05,
|
| 2544 |
+
"loss": 0.0417,
|
| 2545 |
+
"step": 360
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 2.329032258064516,
|
| 2549 |
+
"grad_norm": 0.14067522061651452,
|
| 2550 |
+
"learning_rate": 2.9024099120133673e-05,
|
| 2551 |
+
"loss": 0.0325,
|
| 2552 |
+
"step": 361
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 2.335483870967742,
|
| 2556 |
+
"grad_norm": 0.17575236098692995,
|
| 2557 |
+
"learning_rate": 2.8496664227999415e-05,
|
| 2558 |
+
"loss": 0.0533,
|
| 2559 |
+
"step": 362
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 2.3419354838709676,
|
| 2563 |
+
"grad_norm": 0.11116834491408094,
|
| 2564 |
+
"learning_rate": 2.7973268313426837e-05,
|
| 2565 |
+
"loss": 0.0347,
|
| 2566 |
+
"step": 363
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 2.3483870967741938,
|
| 2570 |
+
"grad_norm": 0.16113939577201203,
|
| 2571 |
+
"learning_rate": 2.745394094125141e-05,
|
| 2572 |
+
"loss": 0.0472,
|
| 2573 |
+
"step": 364
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 2.3548387096774195,
|
| 2577 |
+
"grad_norm": 0.10955543396615161,
|
| 2578 |
+
"learning_rate": 2.6938711446490606e-05,
|
| 2579 |
+
"loss": 0.0357,
|
| 2580 |
+
"step": 365
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 2.361290322580645,
|
| 2584 |
+
"grad_norm": 0.17969798691111394,
|
| 2585 |
+
"learning_rate": 2.6427608932686843e-05,
|
| 2586 |
+
"loss": 0.0584,
|
| 2587 |
+
"step": 366
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 2.367741935483871,
|
| 2591 |
+
"grad_norm": 0.08542932983952008,
|
| 2592 |
+
"learning_rate": 2.5920662270263653e-05,
|
| 2593 |
+
"loss": 0.032,
|
| 2594 |
+
"step": 367
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 2.3741935483870966,
|
| 2598 |
+
"grad_norm": 0.14296192462496796,
|
| 2599 |
+
"learning_rate": 2.5417900094894744e-05,
|
| 2600 |
+
"loss": 0.0463,
|
| 2601 |
+
"step": 368
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 2.3806451612903228,
|
| 2605 |
+
"grad_norm": 0.11857064948287055,
|
| 2606 |
+
"learning_rate": 2.4919350805886577e-05,
|
| 2607 |
+
"loss": 0.0328,
|
| 2608 |
+
"step": 369
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 2.3870967741935485,
|
| 2612 |
+
"grad_norm": 0.1614783419773078,
|
| 2613 |
+
"learning_rate": 2.4425042564574184e-05,
|
| 2614 |
+
"loss": 0.0398,
|
| 2615 |
+
"step": 370
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 2.393548387096774,
|
| 2619 |
+
"grad_norm": 0.12389155943230613,
|
| 2620 |
+
"learning_rate": 2.3935003292730296e-05,
|
| 2621 |
+
"loss": 0.0389,
|
| 2622 |
+
"step": 371
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 2.4,
|
| 2626 |
+
"grad_norm": 0.11380344182842446,
|
| 2627 |
+
"learning_rate": 2.344926067098836e-05,
|
| 2628 |
+
"loss": 0.0412,
|
| 2629 |
+
"step": 372
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 2.4064516129032256,
|
| 2633 |
+
"grad_norm": 0.09336252258103142,
|
| 2634 |
+
"learning_rate": 2.2967842137278706e-05,
|
| 2635 |
+
"loss": 0.0311,
|
| 2636 |
+
"step": 373
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 2.412903225806452,
|
| 2640 |
+
"grad_norm": 0.1312469591783424,
|
| 2641 |
+
"learning_rate": 2.2490774885278908e-05,
|
| 2642 |
+
"loss": 0.0363,
|
| 2643 |
+
"step": 374
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 2.4193548387096775,
|
| 2647 |
+
"grad_norm": 0.1928829483525128,
|
| 2648 |
+
"learning_rate": 2.201808586287757e-05,
|
| 2649 |
+
"loss": 0.0447,
|
| 2650 |
+
"step": 375
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 2.425806451612903,
|
| 2654 |
+
"grad_norm": 0.1407497782508695,
|
| 2655 |
+
"learning_rate": 2.15498017706521e-05,
|
| 2656 |
+
"loss": 0.0425,
|
| 2657 |
+
"step": 376
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 2.432258064516129,
|
| 2661 |
+
"grad_norm": 0.12625194439970722,
|
| 2662 |
+
"learning_rate": 2.1085949060360654e-05,
|
| 2663 |
+
"loss": 0.0394,
|
| 2664 |
+
"step": 377
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 2.4387096774193546,
|
| 2668 |
+
"grad_norm": 0.1650711686265322,
|
| 2669 |
+
"learning_rate": 2.0626553933447734e-05,
|
| 2670 |
+
"loss": 0.0356,
|
| 2671 |
+
"step": 378
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 2.445161290322581,
|
| 2675 |
+
"grad_norm": 0.203794329158885,
|
| 2676 |
+
"learning_rate": 2.01716423395644e-05,
|
| 2677 |
+
"loss": 0.046,
|
| 2678 |
+
"step": 379
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 2.4516129032258065,
|
| 2682 |
+
"grad_norm": 0.0983865374506401,
|
| 2683 |
+
"learning_rate": 1.9721239975102313e-05,
|
| 2684 |
+
"loss": 0.0383,
|
| 2685 |
+
"step": 380
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 2.458064516129032,
|
| 2689 |
+
"grad_norm": 0.11475237817508949,
|
| 2690 |
+
"learning_rate": 1.9275372281742242e-05,
|
| 2691 |
+
"loss": 0.0341,
|
| 2692 |
+
"step": 381
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 2.464516129032258,
|
| 2696 |
+
"grad_norm": 0.11696076741124051,
|
| 2697 |
+
"learning_rate": 1.8834064445016953e-05,
|
| 2698 |
+
"loss": 0.0351,
|
| 2699 |
+
"step": 382
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 2.4709677419354836,
|
| 2703 |
+
"grad_norm": 0.11440255447742229,
|
| 2704 |
+
"learning_rate": 1.839734139288868e-05,
|
| 2705 |
+
"loss": 0.0344,
|
| 2706 |
+
"step": 383
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 2.47741935483871,
|
| 2710 |
+
"grad_norm": 0.12722166866147971,
|
| 2711 |
+
"learning_rate": 1.7965227794340877e-05,
|
| 2712 |
+
"loss": 0.0327,
|
| 2713 |
+
"step": 384
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 2.4838709677419355,
|
| 2717 |
+
"grad_norm": 0.11977054711252778,
|
| 2718 |
+
"learning_rate": 1.753774805798486e-05,
|
| 2719 |
+
"loss": 0.0374,
|
| 2720 |
+
"step": 385
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 2.490322580645161,
|
| 2724 |
+
"grad_norm": 0.19041811914083892,
|
| 2725 |
+
"learning_rate": 1.7114926330680957e-05,
|
| 2726 |
+
"loss": 0.0448,
|
| 2727 |
+
"step": 386
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 2.496774193548387,
|
| 2731 |
+
"grad_norm": 0.12908385372034334,
|
| 2732 |
+
"learning_rate": 1.6696786496174578e-05,
|
| 2733 |
+
"loss": 0.0421,
|
| 2734 |
+
"step": 387
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 2.5032258064516126,
|
| 2738 |
+
"grad_norm": 0.11345958947234494,
|
| 2739 |
+
"learning_rate": 1.6283352173747145e-05,
|
| 2740 |
+
"loss": 0.0337,
|
| 2741 |
+
"step": 388
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 2.509677419354839,
|
| 2745 |
+
"grad_norm": 0.13882462080577593,
|
| 2746 |
+
"learning_rate": 1.587464671688187e-05,
|
| 2747 |
+
"loss": 0.0399,
|
| 2748 |
+
"step": 389
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 2.5161290322580645,
|
| 2752 |
+
"grad_norm": 0.15241110627485502,
|
| 2753 |
+
"learning_rate": 1.5470693211944643e-05,
|
| 2754 |
+
"loss": 0.0331,
|
| 2755 |
+
"step": 390
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 2.52258064516129,
|
| 2759 |
+
"grad_norm": 0.10975692065298487,
|
| 2760 |
+
"learning_rate": 1.5071514476879878e-05,
|
| 2761 |
+
"loss": 0.0379,
|
| 2762 |
+
"step": 391
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 2.5290322580645164,
|
| 2766 |
+
"grad_norm": 0.1379096296948562,
|
| 2767 |
+
"learning_rate": 1.4677133059921632e-05,
|
| 2768 |
+
"loss": 0.0362,
|
| 2769 |
+
"step": 392
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 2.535483870967742,
|
| 2773 |
+
"grad_norm": 0.10190107251103887,
|
| 2774 |
+
"learning_rate": 1.4287571238320053e-05,
|
| 2775 |
+
"loss": 0.0371,
|
| 2776 |
+
"step": 393
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 2.541935483870968,
|
| 2780 |
+
"grad_norm": 0.0875699618751271,
|
| 2781 |
+
"learning_rate": 1.3902851017082864e-05,
|
| 2782 |
+
"loss": 0.0335,
|
| 2783 |
+
"step": 394
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 2.5483870967741935,
|
| 2787 |
+
"grad_norm": 0.13655405710076232,
|
| 2788 |
+
"learning_rate": 1.3522994127732414e-05,
|
| 2789 |
+
"loss": 0.0388,
|
| 2790 |
+
"step": 395
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 2.554838709677419,
|
| 2794 |
+
"grad_norm": 0.11073272357568731,
|
| 2795 |
+
"learning_rate": 1.3148022027078222e-05,
|
| 2796 |
+
"loss": 0.0366,
|
| 2797 |
+
"step": 396
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 2.5612903225806454,
|
| 2801 |
+
"grad_norm": 0.11393589351391524,
|
| 2802 |
+
"learning_rate": 1.2777955896004812e-05,
|
| 2803 |
+
"loss": 0.0375,
|
| 2804 |
+
"step": 397
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 2.567741935483871,
|
| 2808 |
+
"grad_norm": 0.1294793091961239,
|
| 2809 |
+
"learning_rate": 1.2412816638275404e-05,
|
| 2810 |
+
"loss": 0.0447,
|
| 2811 |
+
"step": 398
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 2.574193548387097,
|
| 2815 |
+
"grad_norm": 0.12252957295031998,
|
| 2816 |
+
"learning_rate": 1.2052624879351104e-05,
|
| 2817 |
+
"loss": 0.036,
|
| 2818 |
+
"step": 399
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 2.5806451612903225,
|
| 2822 |
+
"grad_norm": 0.10701768631052949,
|
| 2823 |
+
"learning_rate": 1.1697400965225747e-05,
|
| 2824 |
+
"loss": 0.0375,
|
| 2825 |
+
"step": 400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 2.587096774193548,
|
| 2829 |
+
"grad_norm": 0.1826126683217801,
|
| 2830 |
+
"learning_rate": 1.134716496127679e-05,
|
| 2831 |
+
"loss": 0.0398,
|
| 2832 |
+
"step": 401
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 2.5935483870967744,
|
| 2836 |
+
"grad_norm": 0.12179126729794809,
|
| 2837 |
+
"learning_rate": 1.1001936651131717e-05,
|
| 2838 |
+
"loss": 0.0396,
|
| 2839 |
+
"step": 402
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 2.6,
|
| 2843 |
+
"grad_norm": 0.11972903116479075,
|
| 2844 |
+
"learning_rate": 1.0661735535550666e-05,
|
| 2845 |
+
"loss": 0.0336,
|
| 2846 |
+
"step": 403
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 2.606451612903226,
|
| 2850 |
+
"grad_norm": 0.09797391900877583,
|
| 2851 |
+
"learning_rate": 1.0326580831324817e-05,
|
| 2852 |
+
"loss": 0.0291,
|
| 2853 |
+
"step": 404
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 2.6129032258064515,
|
| 2857 |
+
"grad_norm": 0.08807259670414254,
|
| 2858 |
+
"learning_rate": 9.996491470190917e-06,
|
| 2859 |
+
"loss": 0.0318,
|
| 2860 |
+
"step": 405
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"epoch": 2.6193548387096772,
|
| 2864 |
+
"grad_norm": 0.12986706205285126,
|
| 2865 |
+
"learning_rate": 9.671486097761917e-06,
|
| 2866 |
+
"loss": 0.0408,
|
| 2867 |
+
"step": 406
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 2.6258064516129034,
|
| 2871 |
+
"grad_norm": 0.1331746828106328,
|
| 2872 |
+
"learning_rate": 9.351583072473713e-06,
|
| 2873 |
+
"loss": 0.0427,
|
| 2874 |
+
"step": 407
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 2.632258064516129,
|
| 2878 |
+
"grad_norm": 0.14912986532560119,
|
| 2879 |
+
"learning_rate": 9.036800464548157e-06,
|
| 2880 |
+
"loss": 0.0464,
|
| 2881 |
+
"step": 408
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 2.638709677419355,
|
| 2885 |
+
"grad_norm": 0.1440292521357826,
|
| 2886 |
+
"learning_rate": 8.727156054972374e-06,
|
| 2887 |
+
"loss": 0.0448,
|
| 2888 |
+
"step": 409
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 2.6451612903225805,
|
| 2892 |
+
"grad_norm": 0.16571651133316453,
|
| 2893 |
+
"learning_rate": 8.422667334494249e-06,
|
| 2894 |
+
"loss": 0.0449,
|
| 2895 |
+
"step": 410
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 2.6516129032258062,
|
| 2899 |
+
"grad_norm": 0.15006924502881824,
|
| 2900 |
+
"learning_rate": 8.123351502634625e-06,
|
| 2901 |
+
"loss": 0.0371,
|
| 2902 |
+
"step": 411
|
| 2903 |
+
},
|
| 2904 |
+
{
|
| 2905 |
+
"epoch": 2.6580645161290324,
|
| 2906 |
+
"grad_norm": 0.3766926409227814,
|
| 2907 |
+
"learning_rate": 7.82922546671555e-06,
|
| 2908 |
+
"loss": 0.0522,
|
| 2909 |
+
"step": 412
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 2.664516129032258,
|
| 2913 |
+
"grad_norm": 0.1693245533307095,
|
| 2914 |
+
"learning_rate": 7.54030584090537e-06,
|
| 2915 |
+
"loss": 0.0523,
|
| 2916 |
+
"step": 413
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 2.670967741935484,
|
| 2920 |
+
"grad_norm": 0.16933380856734512,
|
| 2921 |
+
"learning_rate": 7.256608945280319e-06,
|
| 2922 |
+
"loss": 0.0448,
|
| 2923 |
+
"step": 414
|
| 2924 |
+
},
|
| 2925 |
+
{
|
| 2926 |
+
"epoch": 2.6774193548387095,
|
| 2927 |
+
"grad_norm": 0.12372656470631721,
|
| 2928 |
+
"learning_rate": 6.97815080490245e-06,
|
| 2929 |
+
"loss": 0.041,
|
| 2930 |
+
"step": 415
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 2.6838709677419352,
|
| 2934 |
+
"grad_norm": 0.1139014768011305,
|
| 2935 |
+
"learning_rate": 6.704947148914609e-06,
|
| 2936 |
+
"loss": 0.034,
|
| 2937 |
+
"step": 416
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 2.6903225806451614,
|
| 2941 |
+
"grad_norm": 0.14509781428626975,
|
| 2942 |
+
"learning_rate": 6.437013409651849e-06,
|
| 2943 |
+
"loss": 0.0289,
|
| 2944 |
+
"step": 417
|
| 2945 |
+
},
|
| 2946 |
+
{
|
| 2947 |
+
"epoch": 2.696774193548387,
|
| 2948 |
+
"grad_norm": 0.1403482292705624,
|
| 2949 |
+
"learning_rate": 6.174364721769743e-06,
|
| 2950 |
+
"loss": 0.0493,
|
| 2951 |
+
"step": 418
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 2.703225806451613,
|
| 2955 |
+
"grad_norm": 0.12152856591733673,
|
| 2956 |
+
"learning_rate": 5.917015921389568e-06,
|
| 2957 |
+
"loss": 0.0343,
|
| 2958 |
+
"step": 419
|
| 2959 |
+
},
|
| 2960 |
+
{
|
| 2961 |
+
"epoch": 2.709677419354839,
|
| 2962 |
+
"grad_norm": 0.10494471511565585,
|
| 2963 |
+
"learning_rate": 5.664981545260073e-06,
|
| 2964 |
+
"loss": 0.03,
|
| 2965 |
+
"step": 420
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 2.7161290322580647,
|
| 2969 |
+
"grad_norm": 0.2176951472197835,
|
| 2970 |
+
"learning_rate": 5.418275829936537e-06,
|
| 2971 |
+
"loss": 0.0486,
|
| 2972 |
+
"step": 421
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 2.7225806451612904,
|
| 2976 |
+
"grad_norm": 0.09677627092271344,
|
| 2977 |
+
"learning_rate": 5.176912710976467e-06,
|
| 2978 |
+
"loss": 0.0318,
|
| 2979 |
+
"step": 422
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 2.729032258064516,
|
| 2983 |
+
"grad_norm": 0.14416553291024004,
|
| 2984 |
+
"learning_rate": 4.940905822152453e-06,
|
| 2985 |
+
"loss": 0.0356,
|
| 2986 |
+
"step": 423
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 2.735483870967742,
|
| 2990 |
+
"grad_norm": 0.10141360651760384,
|
| 2991 |
+
"learning_rate": 4.710268494682146e-06,
|
| 2992 |
+
"loss": 0.0346,
|
| 2993 |
+
"step": 424
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 2.741935483870968,
|
| 2997 |
+
"grad_norm": 0.13463566674635727,
|
| 2998 |
+
"learning_rate": 4.485013756475076e-06,
|
| 2999 |
+
"loss": 0.0443,
|
| 3000 |
+
"step": 425
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 2.7483870967741937,
|
| 3004 |
+
"grad_norm": 0.26842778961550184,
|
| 3005 |
+
"learning_rate": 4.2651543313968145e-06,
|
| 3006 |
+
"loss": 0.0445,
|
| 3007 |
+
"step": 426
|
| 3008 |
+
},
|
| 3009 |
+
{
|
| 3010 |
+
"epoch": 2.7548387096774194,
|
| 3011 |
+
"grad_norm": 0.16014595507517537,
|
| 3012 |
+
"learning_rate": 4.050702638550275e-06,
|
| 3013 |
+
"loss": 0.0469,
|
| 3014 |
+
"step": 427
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 2.761290322580645,
|
| 3018 |
+
"grad_norm": 0.2690022963645026,
|
| 3019 |
+
"learning_rate": 3.841670791574137e-06,
|
| 3020 |
+
"loss": 0.0423,
|
| 3021 |
+
"step": 428
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 2.767741935483871,
|
| 3025 |
+
"grad_norm": 0.2323526833237967,
|
| 3026 |
+
"learning_rate": 3.638070597958665e-06,
|
| 3027 |
+
"loss": 0.035,
|
| 3028 |
+
"step": 429
|
| 3029 |
+
},
|
| 3030 |
+
{
|
| 3031 |
+
"epoch": 2.774193548387097,
|
| 3032 |
+
"grad_norm": 0.11643458974488918,
|
| 3033 |
+
"learning_rate": 3.4399135583787043e-06,
|
| 3034 |
+
"loss": 0.0401,
|
| 3035 |
+
"step": 430
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 2.7806451612903227,
|
| 3039 |
+
"grad_norm": 0.12278976372135998,
|
| 3040 |
+
"learning_rate": 3.2472108660439706e-06,
|
| 3041 |
+
"loss": 0.04,
|
| 3042 |
+
"step": 431
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 2.7870967741935484,
|
| 3046 |
+
"grad_norm": 0.11369002878299092,
|
| 3047 |
+
"learning_rate": 3.059973406066963e-06,
|
| 3048 |
+
"loss": 0.0335,
|
| 3049 |
+
"step": 432
|
| 3050 |
+
},
|
| 3051 |
+
{
|
| 3052 |
+
"epoch": 2.793548387096774,
|
| 3053 |
+
"grad_norm": 0.1426764505452139,
|
| 3054 |
+
"learning_rate": 2.878211754847926e-06,
|
| 3055 |
+
"loss": 0.0431,
|
| 3056 |
+
"step": 433
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 2.8,
|
| 3060 |
+
"grad_norm": 0.17614309627774977,
|
| 3061 |
+
"learning_rate": 2.7019361794775156e-06,
|
| 3062 |
+
"loss": 0.0336,
|
| 3063 |
+
"step": 434
|
| 3064 |
+
},
|
| 3065 |
+
{
|
| 3066 |
+
"epoch": 2.806451612903226,
|
| 3067 |
+
"grad_norm": 0.13119646329419238,
|
| 3068 |
+
"learning_rate": 2.5311566371568507e-06,
|
| 3069 |
+
"loss": 0.0388,
|
| 3070 |
+
"step": 435
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"epoch": 2.8129032258064517,
|
| 3074 |
+
"grad_norm": 0.2274358694692015,
|
| 3075 |
+
"learning_rate": 2.365882774634998e-06,
|
| 3076 |
+
"loss": 0.0483,
|
| 3077 |
+
"step": 436
|
| 3078 |
+
},
|
| 3079 |
+
{
|
| 3080 |
+
"epoch": 2.8193548387096774,
|
| 3081 |
+
"grad_norm": 0.16809646480030987,
|
| 3082 |
+
"learning_rate": 2.206123927664161e-06,
|
| 3083 |
+
"loss": 0.0428,
|
| 3084 |
+
"step": 437
|
| 3085 |
+
},
|
| 3086 |
+
{
|
| 3087 |
+
"epoch": 2.825806451612903,
|
| 3088 |
+
"grad_norm": 0.08861174696327136,
|
| 3089 |
+
"learning_rate": 2.0518891204722168e-06,
|
| 3090 |
+
"loss": 0.0297,
|
| 3091 |
+
"step": 438
|
| 3092 |
+
},
|
| 3093 |
+
{
|
| 3094 |
+
"epoch": 2.832258064516129,
|
| 3095 |
+
"grad_norm": 0.1516310249747229,
|
| 3096 |
+
"learning_rate": 1.903187065253076e-06,
|
| 3097 |
+
"loss": 0.0486,
|
| 3098 |
+
"step": 439
|
| 3099 |
+
},
|
| 3100 |
+
{
|
| 3101 |
+
"epoch": 2.838709677419355,
|
| 3102 |
+
"grad_norm": 0.1154092981393085,
|
| 3103 |
+
"learning_rate": 1.7600261616745106e-06,
|
| 3104 |
+
"loss": 0.044,
|
| 3105 |
+
"step": 440
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 2.8451612903225807,
|
| 3109 |
+
"grad_norm": 0.14626098217303432,
|
| 3110 |
+
"learning_rate": 1.6224144964036681e-06,
|
| 3111 |
+
"loss": 0.045,
|
| 3112 |
+
"step": 441
|
| 3113 |
+
},
|
| 3114 |
+
{
|
| 3115 |
+
"epoch": 2.8516129032258064,
|
| 3116 |
+
"grad_norm": 0.09272906884038755,
|
| 3117 |
+
"learning_rate": 1.4903598426503241e-06,
|
| 3118 |
+
"loss": 0.0305,
|
| 3119 |
+
"step": 442
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 2.858064516129032,
|
| 3123 |
+
"grad_norm": 0.1688562089326058,
|
| 3124 |
+
"learning_rate": 1.3638696597277679e-06,
|
| 3125 |
+
"loss": 0.0402,
|
| 3126 |
+
"step": 443
|
| 3127 |
+
},
|
| 3128 |
+
{
|
| 3129 |
+
"epoch": 2.864516129032258,
|
| 3130 |
+
"grad_norm": 0.11628209549950047,
|
| 3131 |
+
"learning_rate": 1.2429510926314836e-06,
|
| 3132 |
+
"loss": 0.0309,
|
| 3133 |
+
"step": 444
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 2.870967741935484,
|
| 3137 |
+
"grad_norm": 0.14018105443781761,
|
| 3138 |
+
"learning_rate": 1.1276109716355287e-06,
|
| 3139 |
+
"loss": 0.0466,
|
| 3140 |
+
"step": 445
|
| 3141 |
+
},
|
| 3142 |
+
{
|
| 3143 |
+
"epoch": 2.8774193548387097,
|
| 3144 |
+
"grad_norm": 0.12354234788520546,
|
| 3145 |
+
"learning_rate": 1.0178558119067315e-06,
|
| 3146 |
+
"loss": 0.0315,
|
| 3147 |
+
"step": 446
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 2.8838709677419354,
|
| 3151 |
+
"grad_norm": 0.15472531322652747,
|
| 3152 |
+
"learning_rate": 9.136918131366412e-07,
|
| 3153 |
+
"loss": 0.0436,
|
| 3154 |
+
"step": 447
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 2.8903225806451616,
|
| 3158 |
+
"grad_norm": 0.16727260037004013,
|
| 3159 |
+
"learning_rate": 8.151248591913518e-07,
|
| 3160 |
+
"loss": 0.0413,
|
| 3161 |
+
"step": 448
|
| 3162 |
+
},
|
| 3163 |
+
{
|
| 3164 |
+
"epoch": 2.896774193548387,
|
| 3165 |
+
"grad_norm": 0.15358557638143366,
|
| 3166 |
+
"learning_rate": 7.221605177791691e-07,
|
| 3167 |
+
"loss": 0.0477,
|
| 3168 |
+
"step": 449
|
| 3169 |
+
},
|
| 3170 |
+
{
|
| 3171 |
+
"epoch": 2.903225806451613,
|
| 3172 |
+
"grad_norm": 0.11899732193116695,
|
| 3173 |
+
"learning_rate": 6.348040401360833e-07,
|
| 3174 |
+
"loss": 0.0337,
|
| 3175 |
+
"step": 450
|
| 3176 |
+
},
|
| 3177 |
+
{
|
| 3178 |
+
"epoch": 2.9096774193548387,
|
| 3179 |
+
"grad_norm": 0.19561410636845064,
|
| 3180 |
+
"learning_rate": 5.530603607290851e-07,
|
| 3181 |
+
"loss": 0.0556,
|
| 3182 |
+
"step": 451
|
| 3183 |
+
},
|
| 3184 |
+
{
|
| 3185 |
+
"epoch": 2.9161290322580644,
|
| 3186 |
+
"grad_norm": 0.14981745610583072,
|
| 3187 |
+
"learning_rate": 4.76934096977566e-07,
|
| 3188 |
+
"loss": 0.0372,
|
| 3189 |
+
"step": 452
|
| 3190 |
+
},
|
| 3191 |
+
{
|
| 3192 |
+
"epoch": 2.9225806451612906,
|
| 3193 |
+
"grad_norm": 0.1039480982798509,
|
| 3194 |
+
"learning_rate": 4.0642954899238197e-07,
|
| 3195 |
+
"loss": 0.0343,
|
| 3196 |
+
"step": 453
|
| 3197 |
+
},
|
| 3198 |
+
{
|
| 3199 |
+
"epoch": 2.9290322580645163,
|
| 3200 |
+
"grad_norm": 0.11333580055218463,
|
| 3201 |
+
"learning_rate": 3.415506993330153e-07,
|
| 3202 |
+
"loss": 0.0388,
|
| 3203 |
+
"step": 454
|
| 3204 |
+
},
|
| 3205 |
+
{
|
| 3206 |
+
"epoch": 2.935483870967742,
|
| 3207 |
+
"grad_norm": 0.10481858385197364,
|
| 3208 |
+
"learning_rate": 2.8230121278257637e-07,
|
| 3209 |
+
"loss": 0.0313,
|
| 3210 |
+
"step": 455
|
| 3211 |
+
},
|
| 3212 |
+
{
|
| 3213 |
+
"epoch": 2.9419354838709677,
|
| 3214 |
+
"grad_norm": 0.15064675470229316,
|
| 3215 |
+
"learning_rate": 2.2868443614082469e-07,
|
| 3216 |
+
"loss": 0.0449,
|
| 3217 |
+
"step": 456
|
| 3218 |
+
},
|
| 3219 |
+
{
|
| 3220 |
+
"epoch": 2.9483870967741934,
|
| 3221 |
+
"grad_norm": 0.13918413557689208,
|
| 3222 |
+
"learning_rate": 1.8070339803509807e-07,
|
| 3223 |
+
"loss": 0.0414,
|
| 3224 |
+
"step": 457
|
| 3225 |
+
},
|
| 3226 |
+
{
|
| 3227 |
+
"epoch": 2.9548387096774196,
|
| 3228 |
+
"grad_norm": 0.15721864030858848,
|
| 3229 |
+
"learning_rate": 1.3836080874926049e-07,
|
| 3230 |
+
"loss": 0.0342,
|
| 3231 |
+
"step": 458
|
| 3232 |
+
},
|
| 3233 |
+
{
|
| 3234 |
+
"epoch": 2.9612903225806453,
|
| 3235 |
+
"grad_norm": 0.15411456561887094,
|
| 3236 |
+
"learning_rate": 1.0165906007056914e-07,
|
| 3237 |
+
"loss": 0.0418,
|
| 3238 |
+
"step": 459
|
| 3239 |
+
},
|
| 3240 |
+
{
|
| 3241 |
+
"epoch": 2.967741935483871,
|
| 3242 |
+
"grad_norm": 0.1513944041768792,
|
| 3243 |
+
"learning_rate": 7.060022515460451e-08,
|
| 3244 |
+
"loss": 0.0364,
|
| 3245 |
+
"step": 460
|
| 3246 |
+
},
|
| 3247 |
+
{
|
| 3248 |
+
"epoch": 2.9741935483870967,
|
| 3249 |
+
"grad_norm": 0.22512284396579382,
|
| 3250 |
+
"learning_rate": 4.518605840815315e-08,
|
| 3251 |
+
"loss": 0.0447,
|
| 3252 |
+
"step": 461
|
| 3253 |
+
},
|
| 3254 |
+
{
|
| 3255 |
+
"epoch": 2.9806451612903224,
|
| 3256 |
+
"grad_norm": 0.1529295717768871,
|
| 3257 |
+
"learning_rate": 2.5417995390086824e-08,
|
| 3258 |
+
"loss": 0.0449,
|
| 3259 |
+
"step": 462
|
| 3260 |
+
},
|
| 3261 |
+
{
|
| 3262 |
+
"epoch": 2.9870967741935486,
|
| 3263 |
+
"grad_norm": 0.13942712092352366,
|
| 3264 |
+
"learning_rate": 1.129715273033849e-08,
|
| 3265 |
+
"loss": 0.0367,
|
| 3266 |
+
"step": 463
|
| 3267 |
+
},
|
| 3268 |
+
{
|
| 3269 |
+
"epoch": 2.9935483870967743,
|
| 3270 |
+
"grad_norm": 0.13569899513050301,
|
| 3271 |
+
"learning_rate": 2.824328066730608e-09,
|
| 3272 |
+
"loss": 0.0398,
|
| 3273 |
+
"step": 464
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 3.0,
|
| 3277 |
+
"grad_norm": 0.11004047385534296,
|
| 3278 |
+
"learning_rate": 0.0,
|
| 3279 |
+
"loss": 0.0305,
|
| 3280 |
+
"step": 465
|
| 3281 |
+
},
|
| 3282 |
+
{
|
| 3283 |
+
"epoch": 3.0,
|
| 3284 |
+
"eval_loss": 0.2107405662536621,
|
| 3285 |
+
"eval_runtime": 25.5157,
|
| 3286 |
+
"eval_samples_per_second": 5.134,
|
| 3287 |
+
"eval_steps_per_second": 0.666,
|
| 3288 |
+
"step": 465
|
| 3289 |
+
},
|
| 3290 |
+
{
|
| 3291 |
+
"epoch": 3.0,
|
| 3292 |
+
"step": 465,
|
| 3293 |
+
"total_flos": 261374226563072.0,
|
| 3294 |
+
"train_loss": 0.09038178783751304,
|
| 3295 |
+
"train_runtime": 3542.1658,
|
| 3296 |
+
"train_samples_per_second": 2.097,
|
| 3297 |
+
"train_steps_per_second": 0.131
|
| 3298 |
+
}
|
| 3299 |
+
],
|
| 3300 |
+
"logging_steps": 1,
|
| 3301 |
+
"max_steps": 465,
|
| 3302 |
+
"num_input_tokens_seen": 0,
|
| 3303 |
+
"num_train_epochs": 3,
|
| 3304 |
+
"save_steps": 500,
|
| 3305 |
+
"stateful_callbacks": {
|
| 3306 |
+
"TrainerControl": {
|
| 3307 |
+
"args": {
|
| 3308 |
+
"should_epoch_stop": false,
|
| 3309 |
+
"should_evaluate": false,
|
| 3310 |
+
"should_log": false,
|
| 3311 |
+
"should_save": true,
|
| 3312 |
+
"should_training_stop": true
|
| 3313 |
+
},
|
| 3314 |
+
"attributes": {}
|
| 3315 |
+
}
|
| 3316 |
+
},
|
| 3317 |
+
"total_flos": 261374226563072.0,
|
| 3318 |
+
"train_batch_size": 2,
|
| 3319 |
+
"trial_name": null,
|
| 3320 |
+
"trial_params": null
|
| 3321 |
+
}
|