Qin Liu
commited on
Commit
•
08ad122
1
Parent(s):
20e3667
Model save
Browse files- README.md +61 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +3255 -0
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: meta-llama/Meta-Llama-3-8B-Instruct
|
| 3 |
+
library_name: peft
|
| 4 |
+
license: llama3
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- sft
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: llama3-sudo-10epochs-mask
|
| 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 |
+
# llama3-sudo-10epochs-mask
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on an unknown dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 0.0002
|
| 39 |
+
- train_batch_size: 8
|
| 40 |
+
- eval_batch_size: 8
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 8
|
| 44 |
+
- total_train_batch_size: 64
|
| 45 |
+
- total_eval_batch_size: 64
|
| 46 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
+
- lr_scheduler_type: cosine
|
| 48 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
+
- num_epochs: 10
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
### Framework versions
|
| 56 |
+
|
| 57 |
+
- PEFT 0.12.0
|
| 58 |
+
- Transformers 4.44.0
|
| 59 |
+
- Pytorch 2.1.2
|
| 60 |
+
- Datasets 2.20.0
|
| 61 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"total_flos": 5211452815179776.0,
|
| 4 |
+
"train_loss": 1.1147898718779785,
|
| 5 |
+
"train_runtime": 4502.6314,
|
| 6 |
+
"train_samples": 14642,
|
| 7 |
+
"train_samples_per_second": 32.519,
|
| 8 |
+
"train_steps_per_second": 0.509
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"total_flos": 5211452815179776.0,
|
| 4 |
+
"train_loss": 1.1147898718779785,
|
| 5 |
+
"train_runtime": 4502.6314,
|
| 6 |
+
"train_samples": 14642,
|
| 7 |
+
"train_samples_per_second": 32.519,
|
| 8 |
+
"train_steps_per_second": 0.509
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3255 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2290,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.004366812227074236,
|
| 13 |
+
"grad_norm": 1.8752956704333947,
|
| 14 |
+
"learning_rate": 8.733624454148472e-07,
|
| 15 |
+
"loss": 3.7085,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.021834061135371178,
|
| 20 |
+
"grad_norm": 1.8028899921940393,
|
| 21 |
+
"learning_rate": 4.3668122270742355e-06,
|
| 22 |
+
"loss": 3.6291,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.043668122270742356,
|
| 27 |
+
"grad_norm": 1.75783914939411,
|
| 28 |
+
"learning_rate": 8.733624454148471e-06,
|
| 29 |
+
"loss": 3.547,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.06550218340611354,
|
| 34 |
+
"grad_norm": 2.1823195400139026,
|
| 35 |
+
"learning_rate": 1.3100436681222708e-05,
|
| 36 |
+
"loss": 3.6303,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.08733624454148471,
|
| 41 |
+
"grad_norm": 2.7703299276693256,
|
| 42 |
+
"learning_rate": 1.7467248908296942e-05,
|
| 43 |
+
"loss": 3.5805,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.1091703056768559,
|
| 48 |
+
"grad_norm": 3.0501154969163244,
|
| 49 |
+
"learning_rate": 2.183406113537118e-05,
|
| 50 |
+
"loss": 3.277,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.13100436681222707,
|
| 55 |
+
"grad_norm": 2.0575132099752795,
|
| 56 |
+
"learning_rate": 2.6200873362445416e-05,
|
| 57 |
+
"loss": 2.8289,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.15283842794759825,
|
| 62 |
+
"grad_norm": 1.2985876224099822,
|
| 63 |
+
"learning_rate": 3.056768558951965e-05,
|
| 64 |
+
"loss": 2.6017,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.17467248908296942,
|
| 69 |
+
"grad_norm": 0.7330786544465717,
|
| 70 |
+
"learning_rate": 3.4934497816593884e-05,
|
| 71 |
+
"loss": 2.4112,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.1965065502183406,
|
| 76 |
+
"grad_norm": 0.7933262369374595,
|
| 77 |
+
"learning_rate": 3.930131004366812e-05,
|
| 78 |
+
"loss": 2.2821,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.2183406113537118,
|
| 83 |
+
"grad_norm": 0.7790118551391272,
|
| 84 |
+
"learning_rate": 4.366812227074236e-05,
|
| 85 |
+
"loss": 2.0793,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.24017467248908297,
|
| 90 |
+
"grad_norm": 0.7082020251604142,
|
| 91 |
+
"learning_rate": 4.8034934497816594e-05,
|
| 92 |
+
"loss": 2.0505,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.26200873362445415,
|
| 97 |
+
"grad_norm": 0.5696342576311191,
|
| 98 |
+
"learning_rate": 5.240174672489083e-05,
|
| 99 |
+
"loss": 1.8173,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.2838427947598253,
|
| 104 |
+
"grad_norm": 0.5023621326635465,
|
| 105 |
+
"learning_rate": 5.6768558951965065e-05,
|
| 106 |
+
"loss": 1.8912,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.3056768558951965,
|
| 111 |
+
"grad_norm": 0.4787415611645182,
|
| 112 |
+
"learning_rate": 6.11353711790393e-05,
|
| 113 |
+
"loss": 1.7972,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.32751091703056767,
|
| 118 |
+
"grad_norm": 0.48422783401495123,
|
| 119 |
+
"learning_rate": 6.550218340611354e-05,
|
| 120 |
+
"loss": 1.7183,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.34934497816593885,
|
| 125 |
+
"grad_norm": 0.3910949318663936,
|
| 126 |
+
"learning_rate": 6.986899563318777e-05,
|
| 127 |
+
"loss": 1.6366,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.37117903930131,
|
| 132 |
+
"grad_norm": 0.3902520624627953,
|
| 133 |
+
"learning_rate": 7.423580786026201e-05,
|
| 134 |
+
"loss": 1.7019,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.3930131004366812,
|
| 139 |
+
"grad_norm": 0.3853385163155022,
|
| 140 |
+
"learning_rate": 7.860262008733625e-05,
|
| 141 |
+
"loss": 1.7606,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.4148471615720524,
|
| 146 |
+
"grad_norm": 0.3911758987224519,
|
| 147 |
+
"learning_rate": 8.296943231441049e-05,
|
| 148 |
+
"loss": 1.6395,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.4366812227074236,
|
| 153 |
+
"grad_norm": 0.42385759736776996,
|
| 154 |
+
"learning_rate": 8.733624454148472e-05,
|
| 155 |
+
"loss": 1.6328,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.4585152838427948,
|
| 160 |
+
"grad_norm": 0.3975926298253002,
|
| 161 |
+
"learning_rate": 9.170305676855896e-05,
|
| 162 |
+
"loss": 1.6775,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.48034934497816595,
|
| 167 |
+
"grad_norm": 0.4355288111027398,
|
| 168 |
+
"learning_rate": 9.606986899563319e-05,
|
| 169 |
+
"loss": 1.6008,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.5021834061135371,
|
| 174 |
+
"grad_norm": 0.44774118787630435,
|
| 175 |
+
"learning_rate": 0.00010043668122270742,
|
| 176 |
+
"loss": 1.6046,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.5240174672489083,
|
| 181 |
+
"grad_norm": 0.45365568335085893,
|
| 182 |
+
"learning_rate": 0.00010480349344978167,
|
| 183 |
+
"loss": 1.6348,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.5458515283842795,
|
| 188 |
+
"grad_norm": 0.3985940875605887,
|
| 189 |
+
"learning_rate": 0.00010917030567685591,
|
| 190 |
+
"loss": 1.616,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.5676855895196506,
|
| 195 |
+
"grad_norm": 0.4103041259632963,
|
| 196 |
+
"learning_rate": 0.00011353711790393013,
|
| 197 |
+
"loss": 1.6063,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.5895196506550219,
|
| 202 |
+
"grad_norm": 0.45594857125060284,
|
| 203 |
+
"learning_rate": 0.00011790393013100438,
|
| 204 |
+
"loss": 1.5782,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.611353711790393,
|
| 209 |
+
"grad_norm": 0.406753522288533,
|
| 210 |
+
"learning_rate": 0.0001222707423580786,
|
| 211 |
+
"loss": 1.5361,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.6331877729257642,
|
| 216 |
+
"grad_norm": 0.45489448779672886,
|
| 217 |
+
"learning_rate": 0.00012663755458515284,
|
| 218 |
+
"loss": 1.6416,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.6550218340611353,
|
| 223 |
+
"grad_norm": 0.4232268425851412,
|
| 224 |
+
"learning_rate": 0.00013100436681222707,
|
| 225 |
+
"loss": 1.5449,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.6768558951965066,
|
| 230 |
+
"grad_norm": 0.4008200720858846,
|
| 231 |
+
"learning_rate": 0.00013537117903930133,
|
| 232 |
+
"loss": 1.6322,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.6986899563318777,
|
| 237 |
+
"grad_norm": 0.41073046435729793,
|
| 238 |
+
"learning_rate": 0.00013973799126637554,
|
| 239 |
+
"loss": 1.6144,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.7205240174672489,
|
| 244 |
+
"grad_norm": 0.43150859357010535,
|
| 245 |
+
"learning_rate": 0.0001441048034934498,
|
| 246 |
+
"loss": 1.6816,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.74235807860262,
|
| 251 |
+
"grad_norm": 0.4209643739546475,
|
| 252 |
+
"learning_rate": 0.00014847161572052403,
|
| 253 |
+
"loss": 1.6049,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.7641921397379913,
|
| 258 |
+
"grad_norm": 0.4646259107508816,
|
| 259 |
+
"learning_rate": 0.00015283842794759826,
|
| 260 |
+
"loss": 1.6193,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.7860262008733624,
|
| 265 |
+
"grad_norm": 0.42132133209440126,
|
| 266 |
+
"learning_rate": 0.0001572052401746725,
|
| 267 |
+
"loss": 1.5542,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.8078602620087336,
|
| 272 |
+
"grad_norm": 0.4068655673248684,
|
| 273 |
+
"learning_rate": 0.00016157205240174672,
|
| 274 |
+
"loss": 1.5172,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.8296943231441049,
|
| 279 |
+
"grad_norm": 0.45022442420363395,
|
| 280 |
+
"learning_rate": 0.00016593886462882098,
|
| 281 |
+
"loss": 1.65,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.851528384279476,
|
| 286 |
+
"grad_norm": 0.4218769408186785,
|
| 287 |
+
"learning_rate": 0.00017030567685589521,
|
| 288 |
+
"loss": 1.7073,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.8733624454148472,
|
| 293 |
+
"grad_norm": 0.44363856749896563,
|
| 294 |
+
"learning_rate": 0.00017467248908296945,
|
| 295 |
+
"loss": 1.6647,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.8951965065502183,
|
| 300 |
+
"grad_norm": 0.39452894148369905,
|
| 301 |
+
"learning_rate": 0.00017903930131004368,
|
| 302 |
+
"loss": 1.4932,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.9170305676855895,
|
| 307 |
+
"grad_norm": 0.407234590774645,
|
| 308 |
+
"learning_rate": 0.0001834061135371179,
|
| 309 |
+
"loss": 1.5987,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.9388646288209607,
|
| 314 |
+
"grad_norm": 0.4299787387863718,
|
| 315 |
+
"learning_rate": 0.00018777292576419214,
|
| 316 |
+
"loss": 1.606,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.9606986899563319,
|
| 321 |
+
"grad_norm": 0.4459993359246055,
|
| 322 |
+
"learning_rate": 0.00019213973799126638,
|
| 323 |
+
"loss": 1.6248,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.982532751091703,
|
| 328 |
+
"grad_norm": 0.42477289910814814,
|
| 329 |
+
"learning_rate": 0.0001965065502183406,
|
| 330 |
+
"loss": 1.5145,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 1.0043668122270741,
|
| 335 |
+
"grad_norm": 0.4524366274873438,
|
| 336 |
+
"learning_rate": 0.00019999988382473225,
|
| 337 |
+
"loss": 1.6031,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.0262008733624455,
|
| 342 |
+
"grad_norm": 0.4084729303647835,
|
| 343 |
+
"learning_rate": 0.00019999581771870396,
|
| 344 |
+
"loss": 1.5467,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.0480349344978166,
|
| 349 |
+
"grad_norm": 0.5172909354442642,
|
| 350 |
+
"learning_rate": 0.0001999859431192192,
|
| 351 |
+
"loss": 1.4636,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 1.0698689956331877,
|
| 356 |
+
"grad_norm": 0.4686179138565006,
|
| 357 |
+
"learning_rate": 0.00019997026059986742,
|
| 358 |
+
"loss": 1.5244,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.091703056768559,
|
| 363 |
+
"grad_norm": 0.47437945549478044,
|
| 364 |
+
"learning_rate": 0.00019994877107160482,
|
| 365 |
+
"loss": 1.4414,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.1135371179039302,
|
| 370 |
+
"grad_norm": 0.45613574001819357,
|
| 371 |
+
"learning_rate": 0.00019992147578270142,
|
| 372 |
+
"loss": 1.4545,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 1.1353711790393013,
|
| 377 |
+
"grad_norm": 0.4440788329210085,
|
| 378 |
+
"learning_rate": 0.00019988837631866864,
|
| 379 |
+
"loss": 1.4727,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.1572052401746724,
|
| 384 |
+
"grad_norm": 0.43804218551099794,
|
| 385 |
+
"learning_rate": 0.00019984947460216707,
|
| 386 |
+
"loss": 1.5721,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.1790393013100438,
|
| 391 |
+
"grad_norm": 0.4712591099454086,
|
| 392 |
+
"learning_rate": 0.0001998047728928949,
|
| 393 |
+
"loss": 1.45,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 1.2008733624454149,
|
| 398 |
+
"grad_norm": 0.46682322093255346,
|
| 399 |
+
"learning_rate": 0.00019975427378745659,
|
| 400 |
+
"loss": 1.5364,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.222707423580786,
|
| 405 |
+
"grad_norm": 0.44776219185560584,
|
| 406 |
+
"learning_rate": 0.00019969798021921201,
|
| 407 |
+
"loss": 1.4799,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 1.244541484716157,
|
| 412 |
+
"grad_norm": 0.43694406427786026,
|
| 413 |
+
"learning_rate": 0.0001996358954581062,
|
| 414 |
+
"loss": 1.3916,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 1.2663755458515285,
|
| 419 |
+
"grad_norm": 0.4369949578213358,
|
| 420 |
+
"learning_rate": 0.00019956802311047925,
|
| 421 |
+
"loss": 1.5629,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 1.2882096069868996,
|
| 426 |
+
"grad_norm": 0.4507447731242532,
|
| 427 |
+
"learning_rate": 0.00019949436711885686,
|
| 428 |
+
"loss": 1.5553,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.3100436681222707,
|
| 433 |
+
"grad_norm": 0.4238763083224714,
|
| 434 |
+
"learning_rate": 0.00019941493176172154,
|
| 435 |
+
"loss": 1.555,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 1.3318777292576418,
|
| 440 |
+
"grad_norm": 0.44296845533811896,
|
| 441 |
+
"learning_rate": 0.0001993297216532637,
|
| 442 |
+
"loss": 1.5952,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.3537117903930131,
|
| 447 |
+
"grad_norm": 0.44395910006889616,
|
| 448 |
+
"learning_rate": 0.00019923874174311394,
|
| 449 |
+
"loss": 1.4769,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 1.3755458515283843,
|
| 454 |
+
"grad_norm": 0.44391981727033414,
|
| 455 |
+
"learning_rate": 0.00019914199731605546,
|
| 456 |
+
"loss": 1.5458,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.3973799126637554,
|
| 461 |
+
"grad_norm": 0.46314929089631696,
|
| 462 |
+
"learning_rate": 0.00019903949399171692,
|
| 463 |
+
"loss": 1.5994,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.4192139737991267,
|
| 468 |
+
"grad_norm": 0.4902612617938975,
|
| 469 |
+
"learning_rate": 0.0001989312377242463,
|
| 470 |
+
"loss": 1.5253,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.4410480349344978,
|
| 475 |
+
"grad_norm": 0.4264196408790441,
|
| 476 |
+
"learning_rate": 0.0001988172348019648,
|
| 477 |
+
"loss": 1.5378,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 1.462882096069869,
|
| 482 |
+
"grad_norm": 0.4125915654790707,
|
| 483 |
+
"learning_rate": 0.00019869749184700156,
|
| 484 |
+
"loss": 1.4231,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.48471615720524,
|
| 489 |
+
"grad_norm": 0.4121352169920894,
|
| 490 |
+
"learning_rate": 0.00019857201581490933,
|
| 491 |
+
"loss": 1.4937,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 1.5065502183406112,
|
| 496 |
+
"grad_norm": 0.46111501316021164,
|
| 497 |
+
"learning_rate": 0.00019844081399425997,
|
| 498 |
+
"loss": 1.6366,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 1.5283842794759825,
|
| 503 |
+
"grad_norm": 0.41020501591518393,
|
| 504 |
+
"learning_rate": 0.0001983038940062214,
|
| 505 |
+
"loss": 1.5345,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.5502183406113537,
|
| 510 |
+
"grad_norm": 0.42368595569558,
|
| 511 |
+
"learning_rate": 0.00019816126380411476,
|
| 512 |
+
"loss": 1.5478,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 1.572052401746725,
|
| 517 |
+
"grad_norm": 0.41418005567795424,
|
| 518 |
+
"learning_rate": 0.0001980129316729526,
|
| 519 |
+
"loss": 1.5202,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 1.5938864628820961,
|
| 524 |
+
"grad_norm": 0.4293992491929074,
|
| 525 |
+
"learning_rate": 0.0001978589062289573,
|
| 526 |
+
"loss": 1.4605,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 1.6157205240174672,
|
| 531 |
+
"grad_norm": 0.4361713850313702,
|
| 532 |
+
"learning_rate": 0.00019769919641906097,
|
| 533 |
+
"loss": 1.5154,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 1.6375545851528384,
|
| 538 |
+
"grad_norm": 0.3915518977683478,
|
| 539 |
+
"learning_rate": 0.0001975338115203854,
|
| 540 |
+
"loss": 1.3845,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 1.6593886462882095,
|
| 545 |
+
"grad_norm": 0.40952561447487074,
|
| 546 |
+
"learning_rate": 0.0001973627611397034,
|
| 547 |
+
"loss": 1.5663,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 1.6812227074235808,
|
| 552 |
+
"grad_norm": 0.41401676412381255,
|
| 553 |
+
"learning_rate": 0.00019718605521288073,
|
| 554 |
+
"loss": 1.5892,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 1.703056768558952,
|
| 559 |
+
"grad_norm": 0.40948687994041144,
|
| 560 |
+
"learning_rate": 0.00019700370400429885,
|
| 561 |
+
"loss": 1.5853,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 1.7248908296943233,
|
| 566 |
+
"grad_norm": 0.39912451760472517,
|
| 567 |
+
"learning_rate": 0.00019681571810625873,
|
| 568 |
+
"loss": 1.5086,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 1.7467248908296944,
|
| 573 |
+
"grad_norm": 0.4274906140041681,
|
| 574 |
+
"learning_rate": 0.00019662210843836574,
|
| 575 |
+
"loss": 1.5361,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 1.7685589519650655,
|
| 580 |
+
"grad_norm": 0.45496243462203195,
|
| 581 |
+
"learning_rate": 0.00019642288624689501,
|
| 582 |
+
"loss": 1.5281,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 1.7903930131004366,
|
| 587 |
+
"grad_norm": 0.40754588577831136,
|
| 588 |
+
"learning_rate": 0.00019621806310413857,
|
| 589 |
+
"loss": 1.4146,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 1.8122270742358078,
|
| 594 |
+
"grad_norm": 0.46020764826270205,
|
| 595 |
+
"learning_rate": 0.00019600765090773282,
|
| 596 |
+
"loss": 1.509,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 1.8340611353711789,
|
| 601 |
+
"grad_norm": 0.41433203928546175,
|
| 602 |
+
"learning_rate": 0.0001957916618799676,
|
| 603 |
+
"loss": 1.4521,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 1.8558951965065502,
|
| 608 |
+
"grad_norm": 0.4076208140451916,
|
| 609 |
+
"learning_rate": 0.00019557010856707617,
|
| 610 |
+
"loss": 1.5177,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 1.8777292576419216,
|
| 615 |
+
"grad_norm": 0.3861676767334766,
|
| 616 |
+
"learning_rate": 0.00019534300383850642,
|
| 617 |
+
"loss": 1.5334,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 1.8995633187772927,
|
| 622 |
+
"grad_norm": 0.38833060515579254,
|
| 623 |
+
"learning_rate": 0.00019511036088617342,
|
| 624 |
+
"loss": 1.5405,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.9213973799126638,
|
| 629 |
+
"grad_norm": 0.42695215319286783,
|
| 630 |
+
"learning_rate": 0.000194872193223693,
|
| 631 |
+
"loss": 1.5458,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 1.943231441048035,
|
| 636 |
+
"grad_norm": 0.4089286925168508,
|
| 637 |
+
"learning_rate": 0.0001946285146855968,
|
| 638 |
+
"loss": 1.5466,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 1.965065502183406,
|
| 643 |
+
"grad_norm": 0.42155330336215135,
|
| 644 |
+
"learning_rate": 0.00019437933942652885,
|
| 645 |
+
"loss": 1.566,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 1.9868995633187772,
|
| 650 |
+
"grad_norm": 0.3751384149186164,
|
| 651 |
+
"learning_rate": 0.000194124681920423,
|
| 652 |
+
"loss": 1.4511,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 2.0087336244541483,
|
| 657 |
+
"grad_norm": 0.4161005311145859,
|
| 658 |
+
"learning_rate": 0.00019386455695966253,
|
| 659 |
+
"loss": 1.4751,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 2.03056768558952,
|
| 664 |
+
"grad_norm": 0.4391122093349958,
|
| 665 |
+
"learning_rate": 0.0001935989796542207,
|
| 666 |
+
"loss": 1.4673,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 2.052401746724891,
|
| 671 |
+
"grad_norm": 0.4822422647436377,
|
| 672 |
+
"learning_rate": 0.00019332796543078314,
|
| 673 |
+
"loss": 1.4212,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 2.074235807860262,
|
| 678 |
+
"grad_norm": 0.47440157836581254,
|
| 679 |
+
"learning_rate": 0.00019305153003185165,
|
| 680 |
+
"loss": 1.4117,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 2.096069868995633,
|
| 685 |
+
"grad_norm": 0.5171235330183517,
|
| 686 |
+
"learning_rate": 0.00019276968951482986,
|
| 687 |
+
"loss": 1.377,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 2.1179039301310043,
|
| 692 |
+
"grad_norm": 0.5213453701595562,
|
| 693 |
+
"learning_rate": 0.00019248246025109045,
|
| 694 |
+
"loss": 1.3892,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 2.1397379912663754,
|
| 699 |
+
"grad_norm": 0.5031391500788147,
|
| 700 |
+
"learning_rate": 0.0001921898589250242,
|
| 701 |
+
"loss": 1.3859,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 2.1615720524017465,
|
| 706 |
+
"grad_norm": 0.48354639648956227,
|
| 707 |
+
"learning_rate": 0.00019189190253307082,
|
| 708 |
+
"loss": 1.3916,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 2.183406113537118,
|
| 713 |
+
"grad_norm": 0.5173445212952421,
|
| 714 |
+
"learning_rate": 0.00019158860838273172,
|
| 715 |
+
"loss": 1.3977,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 2.2052401746724892,
|
| 720 |
+
"grad_norm": 0.513094289634623,
|
| 721 |
+
"learning_rate": 0.00019127999409156453,
|
| 722 |
+
"loss": 1.3707,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 2.2270742358078603,
|
| 727 |
+
"grad_norm": 0.5024827997518101,
|
| 728 |
+
"learning_rate": 0.00019096607758615998,
|
| 729 |
+
"loss": 1.3482,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 2.2489082969432315,
|
| 734 |
+
"grad_norm": 0.5798563945761437,
|
| 735 |
+
"learning_rate": 0.0001906468771011003,
|
| 736 |
+
"loss": 1.4178,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 2.2707423580786026,
|
| 741 |
+
"grad_norm": 0.5150755731275006,
|
| 742 |
+
"learning_rate": 0.00019032241117790028,
|
| 743 |
+
"loss": 1.4191,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 2.2925764192139737,
|
| 748 |
+
"grad_norm": 0.5437625450565541,
|
| 749 |
+
"learning_rate": 0.00018999269866393006,
|
| 750 |
+
"loss": 1.3817,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 2.314410480349345,
|
| 755 |
+
"grad_norm": 0.4924027196915137,
|
| 756 |
+
"learning_rate": 0.00018965775871132044,
|
| 757 |
+
"loss": 1.3745,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 2.3362445414847164,
|
| 762 |
+
"grad_norm": 0.5287496980456949,
|
| 763 |
+
"learning_rate": 0.00018931761077585035,
|
| 764 |
+
"loss": 1.3749,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 2.3580786026200875,
|
| 769 |
+
"grad_norm": 0.5123509887446156,
|
| 770 |
+
"learning_rate": 0.00018897227461581672,
|
| 771 |
+
"loss": 1.4476,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 2.3799126637554586,
|
| 776 |
+
"grad_norm": 0.5054089069691902,
|
| 777 |
+
"learning_rate": 0.00018862177029088675,
|
| 778 |
+
"loss": 1.4103,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 2.4017467248908297,
|
| 783 |
+
"grad_norm": 0.5339659464397467,
|
| 784 |
+
"learning_rate": 0.00018826611816093273,
|
| 785 |
+
"loss": 1.421,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 2.423580786026201,
|
| 790 |
+
"grad_norm": 0.5367901761635062,
|
| 791 |
+
"learning_rate": 0.00018790533888484937,
|
| 792 |
+
"loss": 1.4725,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 2.445414847161572,
|
| 797 |
+
"grad_norm": 0.5453912898301467,
|
| 798 |
+
"learning_rate": 0.00018753945341935376,
|
| 799 |
+
"loss": 1.4671,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 2.467248908296943,
|
| 804 |
+
"grad_norm": 0.5114254543022997,
|
| 805 |
+
"learning_rate": 0.0001871684830177681,
|
| 806 |
+
"loss": 1.5483,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 2.489082969432314,
|
| 811 |
+
"grad_norm": 0.5183339258600979,
|
| 812 |
+
"learning_rate": 0.00018679244922878516,
|
| 813 |
+
"loss": 1.4277,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 2.5109170305676853,
|
| 818 |
+
"grad_norm": 0.5202383438539178,
|
| 819 |
+
"learning_rate": 0.00018641137389521645,
|
| 820 |
+
"loss": 1.4767,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 2.532751091703057,
|
| 825 |
+
"grad_norm": 0.5293048970285786,
|
| 826 |
+
"learning_rate": 0.0001860252791527236,
|
| 827 |
+
"loss": 1.4691,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 2.554585152838428,
|
| 832 |
+
"grad_norm": 0.5307661790603934,
|
| 833 |
+
"learning_rate": 0.0001856341874285324,
|
| 834 |
+
"loss": 1.484,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 2.576419213973799,
|
| 839 |
+
"grad_norm": 0.5298444162459437,
|
| 840 |
+
"learning_rate": 0.0001852381214401302,
|
| 841 |
+
"loss": 1.3704,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 2.5982532751091703,
|
| 846 |
+
"grad_norm": 0.5045622495753448,
|
| 847 |
+
"learning_rate": 0.00018483710419394615,
|
| 848 |
+
"loss": 1.4273,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.6200873362445414,
|
| 853 |
+
"grad_norm": 0.516020284616684,
|
| 854 |
+
"learning_rate": 0.00018443115898401504,
|
| 855 |
+
"loss": 1.5253,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 2.641921397379913,
|
| 860 |
+
"grad_norm": 0.5212259321852013,
|
| 861 |
+
"learning_rate": 0.000184020309390624,
|
| 862 |
+
"loss": 1.4966,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 2.6637554585152836,
|
| 867 |
+
"grad_norm": 0.571739861550278,
|
| 868 |
+
"learning_rate": 0.00018360457927894287,
|
| 869 |
+
"loss": 1.489,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 2.685589519650655,
|
| 874 |
+
"grad_norm": 0.5165260257002361,
|
| 875 |
+
"learning_rate": 0.00018318399279763797,
|
| 876 |
+
"loss": 1.419,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 2.7074235807860263,
|
| 881 |
+
"grad_norm": 0.5022734014528262,
|
| 882 |
+
"learning_rate": 0.00018275857437746932,
|
| 883 |
+
"loss": 1.5218,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 2.7292576419213974,
|
| 888 |
+
"grad_norm": 0.5065667081884927,
|
| 889 |
+
"learning_rate": 0.00018232834872987147,
|
| 890 |
+
"loss": 1.3765,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 2.7510917030567685,
|
| 895 |
+
"grad_norm": 0.5208990486632071,
|
| 896 |
+
"learning_rate": 0.00018189334084551826,
|
| 897 |
+
"loss": 1.4514,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 2.7729257641921397,
|
| 902 |
+
"grad_norm": 0.4881264484974513,
|
| 903 |
+
"learning_rate": 0.00018145357599287095,
|
| 904 |
+
"loss": 1.4477,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 2.7947598253275108,
|
| 909 |
+
"grad_norm": 0.513600593429205,
|
| 910 |
+
"learning_rate": 0.00018100907971671054,
|
| 911 |
+
"loss": 1.4449,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 2.816593886462882,
|
| 916 |
+
"grad_norm": 0.6046158817500052,
|
| 917 |
+
"learning_rate": 0.00018055987783665404,
|
| 918 |
+
"loss": 1.3161,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 2.8384279475982535,
|
| 923 |
+
"grad_norm": 0.5208894858087225,
|
| 924 |
+
"learning_rate": 0.00018010599644565457,
|
| 925 |
+
"loss": 1.4693,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 2.8602620087336246,
|
| 930 |
+
"grad_norm": 0.5746066357275653,
|
| 931 |
+
"learning_rate": 0.0001796474619084856,
|
| 932 |
+
"loss": 1.4347,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.8820960698689957,
|
| 937 |
+
"grad_norm": 0.5569260278390025,
|
| 938 |
+
"learning_rate": 0.00017918430086020975,
|
| 939 |
+
"loss": 1.4628,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 2.903930131004367,
|
| 944 |
+
"grad_norm": 0.5015533731915597,
|
| 945 |
+
"learning_rate": 0.0001787165402046313,
|
| 946 |
+
"loss": 1.4082,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 2.925764192139738,
|
| 951 |
+
"grad_norm": 0.4930090755602876,
|
| 952 |
+
"learning_rate": 0.0001782442071127338,
|
| 953 |
+
"loss": 1.4412,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 2.947598253275109,
|
| 958 |
+
"grad_norm": 0.5351113450288878,
|
| 959 |
+
"learning_rate": 0.0001777673290211014,
|
| 960 |
+
"loss": 1.3765,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 2.96943231441048,
|
| 965 |
+
"grad_norm": 0.5333783396073188,
|
| 966 |
+
"learning_rate": 0.00017728593363032532,
|
| 967 |
+
"loss": 1.4074,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 2.9912663755458517,
|
| 972 |
+
"grad_norm": 0.5148622195538735,
|
| 973 |
+
"learning_rate": 0.0001768000489033949,
|
| 974 |
+
"loss": 1.355,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 3.013100436681223,
|
| 979 |
+
"grad_norm": 0.5136549354887358,
|
| 980 |
+
"learning_rate": 0.00017630970306407311,
|
| 981 |
+
"loss": 1.33,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 3.034934497816594,
|
| 986 |
+
"grad_norm": 0.587330448131839,
|
| 987 |
+
"learning_rate": 0.00017581492459525712,
|
| 988 |
+
"loss": 1.267,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 3.056768558951965,
|
| 993 |
+
"grad_norm": 0.5781502863776671,
|
| 994 |
+
"learning_rate": 0.00017531574223732396,
|
| 995 |
+
"loss": 1.3391,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 3.078602620087336,
|
| 1000 |
+
"grad_norm": 0.5801593817600947,
|
| 1001 |
+
"learning_rate": 0.0001748121849864609,
|
| 1002 |
+
"loss": 1.3398,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 3.1004366812227073,
|
| 1007 |
+
"grad_norm": 0.6358481598657785,
|
| 1008 |
+
"learning_rate": 0.00017430428209298126,
|
| 1009 |
+
"loss": 1.3191,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 3.1222707423580784,
|
| 1014 |
+
"grad_norm": 0.635414065898168,
|
| 1015 |
+
"learning_rate": 0.00017379206305962526,
|
| 1016 |
+
"loss": 1.3233,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 3.14410480349345,
|
| 1021 |
+
"grad_norm": 0.6721891418870005,
|
| 1022 |
+
"learning_rate": 0.0001732755576398463,
|
| 1023 |
+
"loss": 1.2795,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 3.165938864628821,
|
| 1028 |
+
"grad_norm": 0.6262467411308055,
|
| 1029 |
+
"learning_rate": 0.00017275479583608261,
|
| 1030 |
+
"loss": 1.3117,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 3.1877729257641922,
|
| 1035 |
+
"grad_norm": 0.7147271334112754,
|
| 1036 |
+
"learning_rate": 0.00017222980789801477,
|
| 1037 |
+
"loss": 1.3604,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 3.2096069868995634,
|
| 1042 |
+
"grad_norm": 0.677887647709634,
|
| 1043 |
+
"learning_rate": 0.00017170062432080805,
|
| 1044 |
+
"loss": 1.3356,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 3.2314410480349345,
|
| 1049 |
+
"grad_norm": 0.6529188195262589,
|
| 1050 |
+
"learning_rate": 0.00017116727584334159,
|
| 1051 |
+
"loss": 1.3092,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 3.2532751091703056,
|
| 1056 |
+
"grad_norm": 0.6545432757758792,
|
| 1057 |
+
"learning_rate": 0.00017062979344642244,
|
| 1058 |
+
"loss": 1.3272,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 3.2751091703056767,
|
| 1063 |
+
"grad_norm": 0.6417623946150004,
|
| 1064 |
+
"learning_rate": 0.00017008820835098627,
|
| 1065 |
+
"loss": 1.3712,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 3.2969432314410483,
|
| 1070 |
+
"grad_norm": 0.6419119938295037,
|
| 1071 |
+
"learning_rate": 0.00016954255201628358,
|
| 1072 |
+
"loss": 1.372,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 3.3187772925764194,
|
| 1077 |
+
"grad_norm": 0.6643469655488602,
|
| 1078 |
+
"learning_rate": 0.00016899285613805246,
|
| 1079 |
+
"loss": 1.3883,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 3.3406113537117905,
|
| 1084 |
+
"grad_norm": 0.6617592247751748,
|
| 1085 |
+
"learning_rate": 0.00016843915264667746,
|
| 1086 |
+
"loss": 1.3131,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 3.3624454148471616,
|
| 1091 |
+
"grad_norm": 0.6721660291620549,
|
| 1092 |
+
"learning_rate": 0.00016788147370533482,
|
| 1093 |
+
"loss": 1.3677,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 3.3842794759825328,
|
| 1098 |
+
"grad_norm": 0.6696065641348963,
|
| 1099 |
+
"learning_rate": 0.00016731985170812414,
|
| 1100 |
+
"loss": 1.3612,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 3.406113537117904,
|
| 1105 |
+
"grad_norm": 0.6442936479861974,
|
| 1106 |
+
"learning_rate": 0.00016675431927818678,
|
| 1107 |
+
"loss": 1.3288,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 3.427947598253275,
|
| 1112 |
+
"grad_norm": 0.6665292623524364,
|
| 1113 |
+
"learning_rate": 0.00016618490926581086,
|
| 1114 |
+
"loss": 1.3302,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 3.449781659388646,
|
| 1119 |
+
"grad_norm": 0.6832138001277586,
|
| 1120 |
+
"learning_rate": 0.00016561165474652292,
|
| 1121 |
+
"loss": 1.296,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 3.4716157205240172,
|
| 1126 |
+
"grad_norm": 0.6676829472258946,
|
| 1127 |
+
"learning_rate": 0.0001650345890191669,
|
| 1128 |
+
"loss": 1.258,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 3.493449781659389,
|
| 1133 |
+
"grad_norm": 0.6462889175077688,
|
| 1134 |
+
"learning_rate": 0.00016445374560396974,
|
| 1135 |
+
"loss": 1.3108,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 3.51528384279476,
|
| 1140 |
+
"grad_norm": 0.6538188867916314,
|
| 1141 |
+
"learning_rate": 0.00016386915824059427,
|
| 1142 |
+
"loss": 1.2225,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 3.537117903930131,
|
| 1147 |
+
"grad_norm": 0.6573367032320154,
|
| 1148 |
+
"learning_rate": 0.0001632808608861794,
|
| 1149 |
+
"loss": 1.2692,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 3.558951965065502,
|
| 1154 |
+
"grad_norm": 0.6707468426806011,
|
| 1155 |
+
"learning_rate": 0.0001626888877133677,
|
| 1156 |
+
"loss": 1.2621,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 3.5807860262008733,
|
| 1161 |
+
"grad_norm": 0.6607806582929415,
|
| 1162 |
+
"learning_rate": 0.00016209327310832028,
|
| 1163 |
+
"loss": 1.3217,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 3.6026200873362444,
|
| 1168 |
+
"grad_norm": 0.6695542325826566,
|
| 1169 |
+
"learning_rate": 0.00016149405166871947,
|
| 1170 |
+
"loss": 1.2445,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 3.6244541484716155,
|
| 1175 |
+
"grad_norm": 0.6854945885270477,
|
| 1176 |
+
"learning_rate": 0.00016089125820175913,
|
| 1177 |
+
"loss": 1.2334,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 3.646288209606987,
|
| 1182 |
+
"grad_norm": 0.6998882406491346,
|
| 1183 |
+
"learning_rate": 0.00016028492772212277,
|
| 1184 |
+
"loss": 1.3228,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 3.668122270742358,
|
| 1189 |
+
"grad_norm": 0.6481594617699246,
|
| 1190 |
+
"learning_rate": 0.00015967509544994959,
|
| 1191 |
+
"loss": 1.3119,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 3.6899563318777293,
|
| 1196 |
+
"grad_norm": 0.6775238829866298,
|
| 1197 |
+
"learning_rate": 0.00015906179680878876,
|
| 1198 |
+
"loss": 1.2587,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 3.7117903930131004,
|
| 1203 |
+
"grad_norm": 0.667250254807951,
|
| 1204 |
+
"learning_rate": 0.00015844506742354164,
|
| 1205 |
+
"loss": 1.335,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 3.7336244541484715,
|
| 1210 |
+
"grad_norm": 0.6850864766225458,
|
| 1211 |
+
"learning_rate": 0.00015782494311839248,
|
| 1212 |
+
"loss": 1.3585,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 3.7554585152838427,
|
| 1217 |
+
"grad_norm": 0.7100136757699026,
|
| 1218 |
+
"learning_rate": 0.00015720145991472746,
|
| 1219 |
+
"loss": 1.3494,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 3.777292576419214,
|
| 1224 |
+
"grad_norm": 0.6455753266089419,
|
| 1225 |
+
"learning_rate": 0.00015657465402904239,
|
| 1226 |
+
"loss": 1.32,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 3.7991266375545854,
|
| 1231 |
+
"grad_norm": 0.6774008615272976,
|
| 1232 |
+
"learning_rate": 0.00015594456187083887,
|
| 1233 |
+
"loss": 1.3053,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 3.8209606986899565,
|
| 1238 |
+
"grad_norm": 0.6412117270718072,
|
| 1239 |
+
"learning_rate": 0.0001553112200405094,
|
| 1240 |
+
"loss": 1.3468,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 3.8427947598253276,
|
| 1245 |
+
"grad_norm": 0.6299987093163749,
|
| 1246 |
+
"learning_rate": 0.00015467466532721136,
|
| 1247 |
+
"loss": 1.2464,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 3.8646288209606987,
|
| 1252 |
+
"grad_norm": 0.6648608762625429,
|
| 1253 |
+
"learning_rate": 0.00015403493470673006,
|
| 1254 |
+
"loss": 1.4054,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 3.88646288209607,
|
| 1259 |
+
"grad_norm": 0.6973437539106749,
|
| 1260 |
+
"learning_rate": 0.00015339206533933087,
|
| 1261 |
+
"loss": 1.3005,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 3.908296943231441,
|
| 1266 |
+
"grad_norm": 0.6609891540439728,
|
| 1267 |
+
"learning_rate": 0.00015274609456760073,
|
| 1268 |
+
"loss": 1.3751,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 3.930131004366812,
|
| 1273 |
+
"grad_norm": 0.6243806427317503,
|
| 1274 |
+
"learning_rate": 0.0001520970599142789,
|
| 1275 |
+
"loss": 1.309,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 3.9519650655021836,
|
| 1280 |
+
"grad_norm": 0.6701604629698161,
|
| 1281 |
+
"learning_rate": 0.00015144499908007757,
|
| 1282 |
+
"loss": 1.3302,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 3.9737991266375547,
|
| 1287 |
+
"grad_norm": 0.6417002641455068,
|
| 1288 |
+
"learning_rate": 0.00015078994994149167,
|
| 1289 |
+
"loss": 1.3244,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 3.995633187772926,
|
| 1294 |
+
"grad_norm": 0.6227627176738441,
|
| 1295 |
+
"learning_rate": 0.00015013195054859894,
|
| 1296 |
+
"loss": 1.3739,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 4.0174672489082965,
|
| 1301 |
+
"grad_norm": 0.69565512329475,
|
| 1302 |
+
"learning_rate": 0.00014947103912284958,
|
| 1303 |
+
"loss": 1.1587,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 4.039301310043668,
|
| 1308 |
+
"grad_norm": 0.9168936710349829,
|
| 1309 |
+
"learning_rate": 0.0001488072540548461,
|
| 1310 |
+
"loss": 1.183,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 4.06113537117904,
|
| 1315 |
+
"grad_norm": 0.768969897857725,
|
| 1316 |
+
"learning_rate": 0.00014814063390211334,
|
| 1317 |
+
"loss": 1.1114,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 4.08296943231441,
|
| 1322 |
+
"grad_norm": 0.8549092972669946,
|
| 1323 |
+
"learning_rate": 0.00014747121738685874,
|
| 1324 |
+
"loss": 1.2111,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 4.104803493449782,
|
| 1329 |
+
"grad_norm": 0.8316566189643829,
|
| 1330 |
+
"learning_rate": 0.00014679904339372302,
|
| 1331 |
+
"loss": 1.1581,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 4.126637554585153,
|
| 1336 |
+
"grad_norm": 0.8226115993114171,
|
| 1337 |
+
"learning_rate": 0.00014612415096752155,
|
| 1338 |
+
"loss": 1.1881,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 4.148471615720524,
|
| 1343 |
+
"grad_norm": 0.8505226801184157,
|
| 1344 |
+
"learning_rate": 0.0001454465793109763,
|
| 1345 |
+
"loss": 1.135,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 4.170305676855895,
|
| 1350 |
+
"grad_norm": 0.8021211984624651,
|
| 1351 |
+
"learning_rate": 0.00014476636778243878,
|
| 1352 |
+
"loss": 1.1768,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 4.192139737991266,
|
| 1357 |
+
"grad_norm": 0.8578093150750806,
|
| 1358 |
+
"learning_rate": 0.00014408355589360348,
|
| 1359 |
+
"loss": 1.0631,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 4.213973799126638,
|
| 1364 |
+
"grad_norm": 0.8812497659618362,
|
| 1365 |
+
"learning_rate": 0.00014339818330721314,
|
| 1366 |
+
"loss": 1.1288,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 4.235807860262009,
|
| 1371 |
+
"grad_norm": 0.7816446218878502,
|
| 1372 |
+
"learning_rate": 0.0001427102898347546,
|
| 1373 |
+
"loss": 1.1777,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 4.25764192139738,
|
| 1378 |
+
"grad_norm": 0.8163412581216741,
|
| 1379 |
+
"learning_rate": 0.0001420199154341464,
|
| 1380 |
+
"loss": 1.1469,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 4.279475982532751,
|
| 1385 |
+
"grad_norm": 0.8410801009958802,
|
| 1386 |
+
"learning_rate": 0.0001413271002074176,
|
| 1387 |
+
"loss": 1.1547,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 4.301310043668122,
|
| 1392 |
+
"grad_norm": 0.8957226442397316,
|
| 1393 |
+
"learning_rate": 0.00014063188439837832,
|
| 1394 |
+
"loss": 1.1054,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 4.323144104803493,
|
| 1399 |
+
"grad_norm": 0.8533966074014762,
|
| 1400 |
+
"learning_rate": 0.0001399343083902824,
|
| 1401 |
+
"loss": 1.1468,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 4.344978165938865,
|
| 1406 |
+
"grad_norm": 0.7969709395883895,
|
| 1407 |
+
"learning_rate": 0.00013923441270348124,
|
| 1408 |
+
"loss": 1.1661,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 4.366812227074236,
|
| 1413 |
+
"grad_norm": 0.830675985424985,
|
| 1414 |
+
"learning_rate": 0.00013853223799307031,
|
| 1415 |
+
"loss": 1.1714,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 4.388646288209607,
|
| 1420 |
+
"grad_norm": 0.879665119495671,
|
| 1421 |
+
"learning_rate": 0.00013782782504652763,
|
| 1422 |
+
"loss": 1.2237,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 4.4104803493449785,
|
| 1427 |
+
"grad_norm": 0.8513132295064585,
|
| 1428 |
+
"learning_rate": 0.0001371212147813443,
|
| 1429 |
+
"loss": 1.2524,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 4.432314410480349,
|
| 1434 |
+
"grad_norm": 0.8345543594033096,
|
| 1435 |
+
"learning_rate": 0.00013641244824264803,
|
| 1436 |
+
"loss": 1.2055,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 4.454148471615721,
|
| 1441 |
+
"grad_norm": 0.8449282094486232,
|
| 1442 |
+
"learning_rate": 0.00013570156660081868,
|
| 1443 |
+
"loss": 1.1459,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 4.475982532751091,
|
| 1448 |
+
"grad_norm": 0.8491089050635324,
|
| 1449 |
+
"learning_rate": 0.00013498861114909685,
|
| 1450 |
+
"loss": 1.165,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 4.497816593886463,
|
| 1455 |
+
"grad_norm": 0.8675954453498238,
|
| 1456 |
+
"learning_rate": 0.00013427362330118543,
|
| 1457 |
+
"loss": 1.1048,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 4.5196506550218345,
|
| 1462 |
+
"grad_norm": 0.9120386243780424,
|
| 1463 |
+
"learning_rate": 0.0001335566445888437,
|
| 1464 |
+
"loss": 1.2427,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 4.541484716157205,
|
| 1469 |
+
"grad_norm": 0.8105081633081175,
|
| 1470 |
+
"learning_rate": 0.00013283771665947505,
|
| 1471 |
+
"loss": 1.278,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 4.563318777292577,
|
| 1476 |
+
"grad_norm": 0.8869239311496004,
|
| 1477 |
+
"learning_rate": 0.00013211688127370784,
|
| 1478 |
+
"loss": 1.1099,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 4.585152838427947,
|
| 1483 |
+
"grad_norm": 0.8873085989909458,
|
| 1484 |
+
"learning_rate": 0.00013139418030296937,
|
| 1485 |
+
"loss": 1.1783,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 4.606986899563319,
|
| 1490 |
+
"grad_norm": 0.80023844006387,
|
| 1491 |
+
"learning_rate": 0.00013066965572705401,
|
| 1492 |
+
"loss": 1.1504,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 4.62882096069869,
|
| 1497 |
+
"grad_norm": 0.8438162486126547,
|
| 1498 |
+
"learning_rate": 0.00012994334963168443,
|
| 1499 |
+
"loss": 1.2292,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 4.650655021834061,
|
| 1504 |
+
"grad_norm": 0.8687319952846376,
|
| 1505 |
+
"learning_rate": 0.00012921530420606714,
|
| 1506 |
+
"loss": 1.2132,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 4.672489082969433,
|
| 1511 |
+
"grad_norm": 0.8481724475183398,
|
| 1512 |
+
"learning_rate": 0.00012848556174044183,
|
| 1513 |
+
"loss": 1.2114,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 4.6943231441048034,
|
| 1518 |
+
"grad_norm": 0.8170589588250686,
|
| 1519 |
+
"learning_rate": 0.00012775416462362457,
|
| 1520 |
+
"loss": 1.2152,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 4.716157205240175,
|
| 1525 |
+
"grad_norm": 0.8800579649975868,
|
| 1526 |
+
"learning_rate": 0.00012702115534054593,
|
| 1527 |
+
"loss": 1.1693,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 4.737991266375546,
|
| 1532 |
+
"grad_norm": 0.8601610801550544,
|
| 1533 |
+
"learning_rate": 0.0001262865764697829,
|
| 1534 |
+
"loss": 1.1846,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 4.759825327510917,
|
| 1539 |
+
"grad_norm": 0.8386680065719362,
|
| 1540 |
+
"learning_rate": 0.00012555047068108568,
|
| 1541 |
+
"loss": 1.249,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 4.781659388646288,
|
| 1546 |
+
"grad_norm": 0.8517305343726155,
|
| 1547 |
+
"learning_rate": 0.00012481288073289912,
|
| 1548 |
+
"loss": 1.1364,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 4.8034934497816595,
|
| 1553 |
+
"grad_norm": 0.8088860139786535,
|
| 1554 |
+
"learning_rate": 0.00012407384946987898,
|
| 1555 |
+
"loss": 1.1527,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 4.825327510917031,
|
| 1560 |
+
"grad_norm": 0.8583326581924249,
|
| 1561 |
+
"learning_rate": 0.00012333341982040323,
|
| 1562 |
+
"loss": 1.1515,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 4.847161572052402,
|
| 1567 |
+
"grad_norm": 0.9360072468671379,
|
| 1568 |
+
"learning_rate": 0.00012259163479407832,
|
| 1569 |
+
"loss": 1.0865,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 4.868995633187773,
|
| 1574 |
+
"grad_norm": 0.8650317997007926,
|
| 1575 |
+
"learning_rate": 0.00012184853747924112,
|
| 1576 |
+
"loss": 1.131,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 4.890829694323144,
|
| 1581 |
+
"grad_norm": 0.8102946053666945,
|
| 1582 |
+
"learning_rate": 0.00012110417104045575,
|
| 1583 |
+
"loss": 1.111,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 4.9126637554585155,
|
| 1588 |
+
"grad_norm": 0.9358255576727259,
|
| 1589 |
+
"learning_rate": 0.00012035857871600649,
|
| 1590 |
+
"loss": 1.2429,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 4.934497816593886,
|
| 1595 |
+
"grad_norm": 0.8877109383416729,
|
| 1596 |
+
"learning_rate": 0.00011961180381538599,
|
| 1597 |
+
"loss": 1.1798,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 4.956331877729258,
|
| 1602 |
+
"grad_norm": 0.8523579756065384,
|
| 1603 |
+
"learning_rate": 0.0001188638897167797,
|
| 1604 |
+
"loss": 1.1524,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 4.978165938864628,
|
| 1609 |
+
"grad_norm": 0.8309844177132272,
|
| 1610 |
+
"learning_rate": 0.00011811487986454612,
|
| 1611 |
+
"loss": 1.2469,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 5.0,
|
| 1616 |
+
"grad_norm": 0.8482507314409842,
|
| 1617 |
+
"learning_rate": 0.00011736481776669306,
|
| 1618 |
+
"loss": 1.1823,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 5.021834061135372,
|
| 1623 |
+
"grad_norm": 1.0322243914151594,
|
| 1624 |
+
"learning_rate": 0.00011661374699235057,
|
| 1625 |
+
"loss": 1.0325,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 5.043668122270742,
|
| 1630 |
+
"grad_norm": 1.0324089865002601,
|
| 1631 |
+
"learning_rate": 0.00011586171116924014,
|
| 1632 |
+
"loss": 1.0234,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 5.065502183406114,
|
| 1637 |
+
"grad_norm": 0.9762094077290032,
|
| 1638 |
+
"learning_rate": 0.00011510875398114027,
|
| 1639 |
+
"loss": 1.0794,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 5.0873362445414845,
|
| 1644 |
+
"grad_norm": 1.1952296969892835,
|
| 1645 |
+
"learning_rate": 0.00011435491916534919,
|
| 1646 |
+
"loss": 1.0145,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 5.109170305676856,
|
| 1651 |
+
"grad_norm": 1.0895050768888697,
|
| 1652 |
+
"learning_rate": 0.0001136002505101442,
|
| 1653 |
+
"loss": 1.0151,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 5.131004366812227,
|
| 1658 |
+
"grad_norm": 1.0235853080588493,
|
| 1659 |
+
"learning_rate": 0.00011284479185223812,
|
| 1660 |
+
"loss": 1.0388,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 5.152838427947598,
|
| 1665 |
+
"grad_norm": 1.0490311233025102,
|
| 1666 |
+
"learning_rate": 0.00011208858707423299,
|
| 1667 |
+
"loss": 1.0072,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 5.17467248908297,
|
| 1672 |
+
"grad_norm": 1.1856197953264118,
|
| 1673 |
+
"learning_rate": 0.00011133168010207091,
|
| 1674 |
+
"loss": 1.0504,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 5.1965065502183405,
|
| 1679 |
+
"grad_norm": 1.0065811557379292,
|
| 1680 |
+
"learning_rate": 0.00011057411490248266,
|
| 1681 |
+
"loss": 0.9977,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 5.218340611353712,
|
| 1686 |
+
"grad_norm": 1.1043632872539175,
|
| 1687 |
+
"learning_rate": 0.00010981593548043374,
|
| 1688 |
+
"loss": 0.9932,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 5.240174672489083,
|
| 1693 |
+
"grad_norm": 1.0753914347833422,
|
| 1694 |
+
"learning_rate": 0.00010905718587656811,
|
| 1695 |
+
"loss": 1.092,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 5.262008733624454,
|
| 1700 |
+
"grad_norm": 1.0266820326577377,
|
| 1701 |
+
"learning_rate": 0.0001082979101646502,
|
| 1702 |
+
"loss": 1.0655,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 5.283842794759825,
|
| 1707 |
+
"grad_norm": 0.9941194317158725,
|
| 1708 |
+
"learning_rate": 0.00010753815244900458,
|
| 1709 |
+
"loss": 0.9828,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 5.3056768558951966,
|
| 1714 |
+
"grad_norm": 1.084324048580049,
|
| 1715 |
+
"learning_rate": 0.00010677795686195422,
|
| 1716 |
+
"loss": 1.0229,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 5.327510917030567,
|
| 1721 |
+
"grad_norm": 1.051439528201926,
|
| 1722 |
+
"learning_rate": 0.00010601736756125685,
|
| 1723 |
+
"loss": 1.0168,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 5.349344978165939,
|
| 1728 |
+
"grad_norm": 1.1580814102197374,
|
| 1729 |
+
"learning_rate": 0.00010525642872753996,
|
| 1730 |
+
"loss": 0.935,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 5.37117903930131,
|
| 1735 |
+
"grad_norm": 1.0710336894680983,
|
| 1736 |
+
"learning_rate": 0.00010449518456173456,
|
| 1737 |
+
"loss": 1.067,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 5.393013100436681,
|
| 1742 |
+
"grad_norm": 1.0463850478020345,
|
| 1743 |
+
"learning_rate": 0.00010373367928250749,
|
| 1744 |
+
"loss": 1.0489,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 5.414847161572053,
|
| 1749 |
+
"grad_norm": 1.1057562523337745,
|
| 1750 |
+
"learning_rate": 0.00010297195712369311,
|
| 1751 |
+
"loss": 0.954,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 5.436681222707423,
|
| 1756 |
+
"grad_norm": 1.0647508212261871,
|
| 1757 |
+
"learning_rate": 0.0001022100623317237,
|
| 1758 |
+
"loss": 0.9094,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 5.458515283842795,
|
| 1763 |
+
"grad_norm": 1.0854376270937827,
|
| 1764 |
+
"learning_rate": 0.00010144803916305925,
|
| 1765 |
+
"loss": 0.9996,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 5.4803493449781655,
|
| 1770 |
+
"grad_norm": 1.0768231071114585,
|
| 1771 |
+
"learning_rate": 0.00010068593188161697,
|
| 1772 |
+
"loss": 1.0098,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 5.502183406113537,
|
| 1777 |
+
"grad_norm": 1.0629470347878223,
|
| 1778 |
+
"learning_rate": 9.992378475619981e-05,
|
| 1779 |
+
"loss": 1.0252,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 5.524017467248909,
|
| 1784 |
+
"grad_norm": 1.051250411684801,
|
| 1785 |
+
"learning_rate": 9.916164205792527e-05,
|
| 1786 |
+
"loss": 0.9879,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 5.545851528384279,
|
| 1791 |
+
"grad_norm": 1.0100574406024927,
|
| 1792 |
+
"learning_rate": 9.839954805765364e-05,
|
| 1793 |
+
"loss": 1.0638,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 5.567685589519651,
|
| 1798 |
+
"grad_norm": 1.0416962539798005,
|
| 1799 |
+
"learning_rate": 9.763754702341646e-05,
|
| 1800 |
+
"loss": 0.9556,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 5.5895196506550215,
|
| 1805 |
+
"grad_norm": 1.041656249717949,
|
| 1806 |
+
"learning_rate": 9.687568321784509e-05,
|
| 1807 |
+
"loss": 1.0295,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 5.611353711790393,
|
| 1812 |
+
"grad_norm": 1.057095267929098,
|
| 1813 |
+
"learning_rate": 9.611400089559975e-05,
|
| 1814 |
+
"loss": 1.0233,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 5.633187772925764,
|
| 1819 |
+
"grad_norm": 1.0153521926215252,
|
| 1820 |
+
"learning_rate": 9.535254430079864e-05,
|
| 1821 |
+
"loss": 0.9867,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 5.655021834061135,
|
| 1826 |
+
"grad_norm": 1.1345238181227135,
|
| 1827 |
+
"learning_rate": 9.459135766444815e-05,
|
| 1828 |
+
"loss": 1.0027,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 5.676855895196507,
|
| 1833 |
+
"grad_norm": 1.1187108189925221,
|
| 1834 |
+
"learning_rate": 9.383048520187344e-05,
|
| 1835 |
+
"loss": 0.9987,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 5.698689956331878,
|
| 1840 |
+
"grad_norm": 1.0636989170846824,
|
| 1841 |
+
"learning_rate": 9.306997111015014e-05,
|
| 1842 |
+
"loss": 1.0486,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 5.720524017467249,
|
| 1847 |
+
"grad_norm": 1.0706396029944643,
|
| 1848 |
+
"learning_rate": 9.23098595655371e-05,
|
| 1849 |
+
"loss": 0.9931,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 5.74235807860262,
|
| 1854 |
+
"grad_norm": 1.0475360658888055,
|
| 1855 |
+
"learning_rate": 9.155019472091022e-05,
|
| 1856 |
+
"loss": 0.9749,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 5.764192139737991,
|
| 1861 |
+
"grad_norm": 1.0214139834738702,
|
| 1862 |
+
"learning_rate": 9.079102070319786e-05,
|
| 1863 |
+
"loss": 1.0693,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 5.786026200873362,
|
| 1868 |
+
"grad_norm": 1.0530973134949948,
|
| 1869 |
+
"learning_rate": 9.003238161081743e-05,
|
| 1870 |
+
"loss": 1.0228,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 5.807860262008734,
|
| 1875 |
+
"grad_norm": 1.103677474707846,
|
| 1876 |
+
"learning_rate": 8.9274321511114e-05,
|
| 1877 |
+
"loss": 0.9761,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 5.829694323144105,
|
| 1882 |
+
"grad_norm": 1.0644633976825475,
|
| 1883 |
+
"learning_rate": 8.851688443780043e-05,
|
| 1884 |
+
"loss": 1.0239,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 5.851528384279476,
|
| 1889 |
+
"grad_norm": 1.0555551246349646,
|
| 1890 |
+
"learning_rate": 8.776011438839977e-05,
|
| 1891 |
+
"loss": 1.0473,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 5.873362445414847,
|
| 1896 |
+
"grad_norm": 1.122056836899885,
|
| 1897 |
+
"learning_rate": 8.70040553216892e-05,
|
| 1898 |
+
"loss": 0.9723,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 5.895196506550218,
|
| 1903 |
+
"grad_norm": 0.9963873601564418,
|
| 1904 |
+
"learning_rate": 8.624875115514697e-05,
|
| 1905 |
+
"loss": 1.0268,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 5.91703056768559,
|
| 1910 |
+
"grad_norm": 1.0336940380478288,
|
| 1911 |
+
"learning_rate": 8.549424576240102e-05,
|
| 1912 |
+
"loss": 0.9574,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 5.93886462882096,
|
| 1917 |
+
"grad_norm": 1.0760501661341102,
|
| 1918 |
+
"learning_rate": 8.474058297068071e-05,
|
| 1919 |
+
"loss": 1.0979,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 5.960698689956332,
|
| 1924 |
+
"grad_norm": 1.0890942343064978,
|
| 1925 |
+
"learning_rate": 8.398780655827096e-05,
|
| 1926 |
+
"loss": 0.9427,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 5.9825327510917035,
|
| 1931 |
+
"grad_norm": 1.1415589599070428,
|
| 1932 |
+
"learning_rate": 8.323596025196911e-05,
|
| 1933 |
+
"loss": 1.0041,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 6.004366812227074,
|
| 1938 |
+
"grad_norm": 1.1509480354124522,
|
| 1939 |
+
"learning_rate": 8.248508772454529e-05,
|
| 1940 |
+
"loss": 0.9545,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 6.026200873362446,
|
| 1945 |
+
"grad_norm": 1.4252192004046074,
|
| 1946 |
+
"learning_rate": 8.173523259220521e-05,
|
| 1947 |
+
"loss": 0.8584,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 6.048034934497816,
|
| 1952 |
+
"grad_norm": 1.37449264500959,
|
| 1953 |
+
"learning_rate": 8.098643841205685e-05,
|
| 1954 |
+
"loss": 0.8417,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 6.069868995633188,
|
| 1959 |
+
"grad_norm": 1.1455224268481572,
|
| 1960 |
+
"learning_rate": 8.023874867958027e-05,
|
| 1961 |
+
"loss": 0.8365,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 6.091703056768559,
|
| 1966 |
+
"grad_norm": 1.4517829308762,
|
| 1967 |
+
"learning_rate": 7.949220682610109e-05,
|
| 1968 |
+
"loss": 0.8772,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 6.11353711790393,
|
| 1973 |
+
"grad_norm": 1.2198167511326543,
|
| 1974 |
+
"learning_rate": 7.874685621626767e-05,
|
| 1975 |
+
"loss": 0.7638,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 6.135371179039302,
|
| 1980 |
+
"grad_norm": 1.2285129578196883,
|
| 1981 |
+
"learning_rate": 7.80027401455321e-05,
|
| 1982 |
+
"loss": 0.8632,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 6.157205240174672,
|
| 1987 |
+
"grad_norm": 1.3128304307256697,
|
| 1988 |
+
"learning_rate": 7.725990183763541e-05,
|
| 1989 |
+
"loss": 0.7864,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 6.179039301310044,
|
| 1994 |
+
"grad_norm": 1.2223078968334138,
|
| 1995 |
+
"learning_rate": 7.651838444209678e-05,
|
| 1996 |
+
"loss": 0.8107,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 6.200873362445415,
|
| 2001 |
+
"grad_norm": 1.2116494464520935,
|
| 2002 |
+
"learning_rate": 7.577823103170695e-05,
|
| 2003 |
+
"loss": 0.7665,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 6.222707423580786,
|
| 2008 |
+
"grad_norm": 1.2710441404657735,
|
| 2009 |
+
"learning_rate": 7.503948460002651e-05,
|
| 2010 |
+
"loss": 0.8755,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 6.244541484716157,
|
| 2015 |
+
"grad_norm": 1.3090624013120402,
|
| 2016 |
+
"learning_rate": 7.430218805888831e-05,
|
| 2017 |
+
"loss": 0.8635,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 6.2663755458515285,
|
| 2022 |
+
"grad_norm": 1.3421074934086965,
|
| 2023 |
+
"learning_rate": 7.356638423590485e-05,
|
| 2024 |
+
"loss": 0.8408,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 6.2882096069869,
|
| 2029 |
+
"grad_norm": 1.2190845509165837,
|
| 2030 |
+
"learning_rate": 7.283211587198056e-05,
|
| 2031 |
+
"loss": 0.901,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 6.310043668122271,
|
| 2036 |
+
"grad_norm": 1.2848215636192764,
|
| 2037 |
+
"learning_rate": 7.209942561882914e-05,
|
| 2038 |
+
"loss": 0.8183,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 6.331877729257642,
|
| 2043 |
+
"grad_norm": 1.2782891875098124,
|
| 2044 |
+
"learning_rate": 7.136835603649599e-05,
|
| 2045 |
+
"loss": 0.8144,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 6.353711790393013,
|
| 2050 |
+
"grad_norm": 1.3643650683015465,
|
| 2051 |
+
"learning_rate": 7.0638949590886e-05,
|
| 2052 |
+
"loss": 0.815,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 6.3755458515283845,
|
| 2057 |
+
"grad_norm": 1.3747527932453862,
|
| 2058 |
+
"learning_rate": 6.991124865129683e-05,
|
| 2059 |
+
"loss": 0.8058,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 6.397379912663755,
|
| 2064 |
+
"grad_norm": 1.324272429862148,
|
| 2065 |
+
"learning_rate": 6.918529548795781e-05,
|
| 2066 |
+
"loss": 0.8359,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 6.419213973799127,
|
| 2071 |
+
"grad_norm": 1.329733488330745,
|
| 2072 |
+
"learning_rate": 6.846113226957456e-05,
|
| 2073 |
+
"loss": 0.8081,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 6.441048034934497,
|
| 2078 |
+
"grad_norm": 1.3973165024504683,
|
| 2079 |
+
"learning_rate": 6.773880106087945e-05,
|
| 2080 |
+
"loss": 0.9255,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 6.462882096069869,
|
| 2085 |
+
"grad_norm": 1.2579802684349493,
|
| 2086 |
+
"learning_rate": 6.701834382018832e-05,
|
| 2087 |
+
"loss": 0.8932,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 6.4847161572052405,
|
| 2092 |
+
"grad_norm": 1.3102021281865468,
|
| 2093 |
+
"learning_rate": 6.629980239696315e-05,
|
| 2094 |
+
"loss": 0.8651,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 6.506550218340611,
|
| 2099 |
+
"grad_norm": 1.3096833467828455,
|
| 2100 |
+
"learning_rate": 6.558321852938099e-05,
|
| 2101 |
+
"loss": 0.8145,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 6.528384279475983,
|
| 2106 |
+
"grad_norm": 1.3466234774293075,
|
| 2107 |
+
"learning_rate": 6.486863384190987e-05,
|
| 2108 |
+
"loss": 0.8885,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 6.550218340611353,
|
| 2113 |
+
"grad_norm": 1.296177045867574,
|
| 2114 |
+
"learning_rate": 6.415608984289052e-05,
|
| 2115 |
+
"loss": 0.8546,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 6.572052401746725,
|
| 2120 |
+
"grad_norm": 1.2633240958805778,
|
| 2121 |
+
"learning_rate": 6.344562792212554e-05,
|
| 2122 |
+
"loss": 0.8685,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 6.593886462882097,
|
| 2127 |
+
"grad_norm": 1.3534959976317207,
|
| 2128 |
+
"learning_rate": 6.273728934847516e-05,
|
| 2129 |
+
"loss": 0.7986,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 6.615720524017467,
|
| 2134 |
+
"grad_norm": 1.252619828877404,
|
| 2135 |
+
"learning_rate": 6.203111526745985e-05,
|
| 2136 |
+
"loss": 0.8332,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 6.637554585152839,
|
| 2141 |
+
"grad_norm": 1.3249258619095206,
|
| 2142 |
+
"learning_rate": 6.132714669887044e-05,
|
| 2143 |
+
"loss": 0.8308,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 6.6593886462882095,
|
| 2148 |
+
"grad_norm": 1.2054299066608396,
|
| 2149 |
+
"learning_rate": 6.0625424534385425e-05,
|
| 2150 |
+
"loss": 0.8697,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 6.681222707423581,
|
| 2155 |
+
"grad_norm": 1.1955192517041973,
|
| 2156 |
+
"learning_rate": 5.99259895351955e-05,
|
| 2157 |
+
"loss": 0.8591,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 6.703056768558952,
|
| 2162 |
+
"grad_norm": 1.2570222378899258,
|
| 2163 |
+
"learning_rate": 5.9228882329636094e-05,
|
| 2164 |
+
"loss": 0.7953,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 6.724890829694323,
|
| 2169 |
+
"grad_norm": 1.3956159571082007,
|
| 2170 |
+
"learning_rate": 5.8534143410827104e-05,
|
| 2171 |
+
"loss": 0.8367,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 6.746724890829694,
|
| 2176 |
+
"grad_norm": 1.3649251909287845,
|
| 2177 |
+
"learning_rate": 5.7841813134320975e-05,
|
| 2178 |
+
"loss": 0.8553,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 6.7685589519650655,
|
| 2183 |
+
"grad_norm": 1.3726673140188062,
|
| 2184 |
+
"learning_rate": 5.715193171575842e-05,
|
| 2185 |
+
"loss": 0.8649,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 6.790393013100436,
|
| 2190 |
+
"grad_norm": 1.2184325480564675,
|
| 2191 |
+
"learning_rate": 5.64645392285325e-05,
|
| 2192 |
+
"loss": 0.8222,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 6.812227074235808,
|
| 2197 |
+
"grad_norm": 1.3383861408747,
|
| 2198 |
+
"learning_rate": 5.577967560146077e-05,
|
| 2199 |
+
"loss": 0.851,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 6.834061135371179,
|
| 2204 |
+
"grad_norm": 1.3171908370457348,
|
| 2205 |
+
"learning_rate": 5.5097380616466057e-05,
|
| 2206 |
+
"loss": 0.8662,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 6.85589519650655,
|
| 2211 |
+
"grad_norm": 1.3054511507141975,
|
| 2212 |
+
"learning_rate": 5.4417693906265365e-05,
|
| 2213 |
+
"loss": 0.8979,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 6.877729257641922,
|
| 2218 |
+
"grad_norm": 1.2387745464762083,
|
| 2219 |
+
"learning_rate": 5.374065495206805e-05,
|
| 2220 |
+
"loss": 0.8119,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 6.899563318777292,
|
| 2225 |
+
"grad_norm": 1.3948067425415336,
|
| 2226 |
+
"learning_rate": 5.306630308128229e-05,
|
| 2227 |
+
"loss": 0.8409,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 6.921397379912664,
|
| 2232 |
+
"grad_norm": 1.3502073303444249,
|
| 2233 |
+
"learning_rate": 5.239467746523048e-05,
|
| 2234 |
+
"loss": 0.8391,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 6.9432314410480345,
|
| 2239 |
+
"grad_norm": 1.3334535387387385,
|
| 2240 |
+
"learning_rate": 5.172581711687438e-05,
|
| 2241 |
+
"loss": 0.8577,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 6.965065502183406,
|
| 2246 |
+
"grad_norm": 1.3381286547460995,
|
| 2247 |
+
"learning_rate": 5.105976088854842e-05,
|
| 2248 |
+
"loss": 0.8925,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 6.986899563318778,
|
| 2253 |
+
"grad_norm": 1.2107404270853181,
|
| 2254 |
+
"learning_rate": 5.0396547469703106e-05,
|
| 2255 |
+
"loss": 0.8894,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 7.008733624454148,
|
| 2260 |
+
"grad_norm": 1.2856577830397042,
|
| 2261 |
+
"learning_rate": 4.973621538465768e-05,
|
| 2262 |
+
"loss": 0.8269,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 7.03056768558952,
|
| 2267 |
+
"grad_norm": 1.5720811121732976,
|
| 2268 |
+
"learning_rate": 4.907880299036234e-05,
|
| 2269 |
+
"loss": 0.6532,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 7.0524017467248905,
|
| 2274 |
+
"grad_norm": 1.513790910492325,
|
| 2275 |
+
"learning_rate": 4.8424348474170014e-05,
|
| 2276 |
+
"loss": 0.6398,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 7.074235807860262,
|
| 2281 |
+
"grad_norm": 1.3731923874867311,
|
| 2282 |
+
"learning_rate": 4.7772889851618405e-05,
|
| 2283 |
+
"loss": 0.7323,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 7.096069868995633,
|
| 2288 |
+
"grad_norm": 1.3538833010462115,
|
| 2289 |
+
"learning_rate": 4.712446496422165e-05,
|
| 2290 |
+
"loss": 0.6906,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 7.117903930131004,
|
| 2295 |
+
"grad_norm": 1.4454754507310241,
|
| 2296 |
+
"learning_rate": 4.647911147727209e-05,
|
| 2297 |
+
"loss": 0.7328,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 7.139737991266376,
|
| 2302 |
+
"grad_norm": 1.449171514767732,
|
| 2303 |
+
"learning_rate": 4.583686687765264e-05,
|
| 2304 |
+
"loss": 0.6782,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 7.1615720524017465,
|
| 2309 |
+
"grad_norm": 1.6050731544814476,
|
| 2310 |
+
"learning_rate": 4.5197768471659104e-05,
|
| 2311 |
+
"loss": 0.7385,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 7.183406113537118,
|
| 2316 |
+
"grad_norm": 1.3388820740639182,
|
| 2317 |
+
"learning_rate": 4.4561853382833206e-05,
|
| 2318 |
+
"loss": 0.6937,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 7.205240174672489,
|
| 2323 |
+
"grad_norm": 1.3724590934014314,
|
| 2324 |
+
"learning_rate": 4.3929158549806096e-05,
|
| 2325 |
+
"loss": 0.6899,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 7.22707423580786,
|
| 2330 |
+
"grad_norm": 1.5165378554181088,
|
| 2331 |
+
"learning_rate": 4.32997207241528e-05,
|
| 2332 |
+
"loss": 0.7044,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 7.248908296943231,
|
| 2337 |
+
"grad_norm": 1.6117091341741596,
|
| 2338 |
+
"learning_rate": 4.267357646825746e-05,
|
| 2339 |
+
"loss": 0.7093,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 7.270742358078603,
|
| 2344 |
+
"grad_norm": 1.4490430208656846,
|
| 2345 |
+
"learning_rate": 4.205076215318925e-05,
|
| 2346 |
+
"loss": 0.6967,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 7.292576419213974,
|
| 2351 |
+
"grad_norm": 1.469380632694007,
|
| 2352 |
+
"learning_rate": 4.143131395658996e-05,
|
| 2353 |
+
"loss": 0.7164,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 7.314410480349345,
|
| 2358 |
+
"grad_norm": 1.519813749480573,
|
| 2359 |
+
"learning_rate": 4.081526786057254e-05,
|
| 2360 |
+
"loss": 0.6724,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 7.336244541484716,
|
| 2365 |
+
"grad_norm": 1.4588612796193572,
|
| 2366 |
+
"learning_rate": 4.020265964963066e-05,
|
| 2367 |
+
"loss": 0.731,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 7.358078602620087,
|
| 2372 |
+
"grad_norm": 1.429975103899558,
|
| 2373 |
+
"learning_rate": 3.9593524908560464e-05,
|
| 2374 |
+
"loss": 0.7327,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 7.379912663755459,
|
| 2379 |
+
"grad_norm": 1.6113371519439155,
|
| 2380 |
+
"learning_rate": 3.898789902039338e-05,
|
| 2381 |
+
"loss": 0.709,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 7.401746724890829,
|
| 2386 |
+
"grad_norm": 1.5463250020435173,
|
| 2387 |
+
"learning_rate": 3.8385817164340723e-05,
|
| 2388 |
+
"loss": 0.7246,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 7.423580786026201,
|
| 2393 |
+
"grad_norm": 1.463616658449462,
|
| 2394 |
+
"learning_rate": 3.778731431375041e-05,
|
| 2395 |
+
"loss": 0.7013,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 7.445414847161572,
|
| 2400 |
+
"grad_norm": 1.5566283347360372,
|
| 2401 |
+
"learning_rate": 3.719242523407539e-05,
|
| 2402 |
+
"loss": 0.7344,
|
| 2403 |
+
"step": 1705
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 7.467248908296943,
|
| 2407 |
+
"grad_norm": 1.4398876621434327,
|
| 2408 |
+
"learning_rate": 3.6601184480854066e-05,
|
| 2409 |
+
"loss": 0.7323,
|
| 2410 |
+
"step": 1710
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 7.489082969432315,
|
| 2414 |
+
"grad_norm": 1.5252195085267406,
|
| 2415 |
+
"learning_rate": 3.601362639770328e-05,
|
| 2416 |
+
"loss": 0.7091,
|
| 2417 |
+
"step": 1715
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 7.510917030567685,
|
| 2421 |
+
"grad_norm": 1.6881923973765582,
|
| 2422 |
+
"learning_rate": 3.542978511432325e-05,
|
| 2423 |
+
"loss": 0.7585,
|
| 2424 |
+
"step": 1720
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 7.532751091703057,
|
| 2428 |
+
"grad_norm": 1.450411669170985,
|
| 2429 |
+
"learning_rate": 3.484969454451511e-05,
|
| 2430 |
+
"loss": 0.7258,
|
| 2431 |
+
"step": 1725
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 7.554585152838428,
|
| 2435 |
+
"grad_norm": 1.523075629927926,
|
| 2436 |
+
"learning_rate": 3.4273388384210855e-05,
|
| 2437 |
+
"loss": 0.6716,
|
| 2438 |
+
"step": 1730
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 7.576419213973799,
|
| 2442 |
+
"grad_norm": 1.4565661875305633,
|
| 2443 |
+
"learning_rate": 3.3700900109516184e-05,
|
| 2444 |
+
"loss": 0.6586,
|
| 2445 |
+
"step": 1735
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 7.598253275109171,
|
| 2449 |
+
"grad_norm": 1.6164061159781367,
|
| 2450 |
+
"learning_rate": 3.3132262974765906e-05,
|
| 2451 |
+
"loss": 0.7123,
|
| 2452 |
+
"step": 1740
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 7.620087336244541,
|
| 2456 |
+
"grad_norm": 1.5157276652461606,
|
| 2457 |
+
"learning_rate": 3.256751001059214e-05,
|
| 2458 |
+
"loss": 0.723,
|
| 2459 |
+
"step": 1745
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 7.641921397379913,
|
| 2463 |
+
"grad_norm": 1.5600828466807661,
|
| 2464 |
+
"learning_rate": 3.200667402200586e-05,
|
| 2465 |
+
"loss": 0.7477,
|
| 2466 |
+
"step": 1750
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 7.663755458515284,
|
| 2470 |
+
"grad_norm": 1.5728766811598356,
|
| 2471 |
+
"learning_rate": 3.144978758649133e-05,
|
| 2472 |
+
"loss": 0.7001,
|
| 2473 |
+
"step": 1755
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 7.685589519650655,
|
| 2477 |
+
"grad_norm": 1.458458820326098,
|
| 2478 |
+
"learning_rate": 3.0896883052113525e-05,
|
| 2479 |
+
"loss": 0.7066,
|
| 2480 |
+
"step": 1760
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 7.707423580786026,
|
| 2484 |
+
"grad_norm": 1.5980960571332508,
|
| 2485 |
+
"learning_rate": 3.034799253563939e-05,
|
| 2486 |
+
"loss": 0.6878,
|
| 2487 |
+
"step": 1765
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 7.729257641921397,
|
| 2491 |
+
"grad_norm": 1.544990160417705,
|
| 2492 |
+
"learning_rate": 2.9803147920672146e-05,
|
| 2493 |
+
"loss": 0.6894,
|
| 2494 |
+
"step": 1770
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 7.751091703056769,
|
| 2498 |
+
"grad_norm": 1.6353637199811462,
|
| 2499 |
+
"learning_rate": 2.9262380855799164e-05,
|
| 2500 |
+
"loss": 0.7297,
|
| 2501 |
+
"step": 1775
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 7.77292576419214,
|
| 2505 |
+
"grad_norm": 1.5830327061313785,
|
| 2506 |
+
"learning_rate": 2.872572275275379e-05,
|
| 2507 |
+
"loss": 0.6983,
|
| 2508 |
+
"step": 1780
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 7.794759825327511,
|
| 2512 |
+
"grad_norm": 1.4630236788620592,
|
| 2513 |
+
"learning_rate": 2.8193204784590597e-05,
|
| 2514 |
+
"loss": 0.7176,
|
| 2515 |
+
"step": 1785
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 7.816593886462882,
|
| 2519 |
+
"grad_norm": 1.3928429612080049,
|
| 2520 |
+
"learning_rate": 2.766485788387455e-05,
|
| 2521 |
+
"loss": 0.7269,
|
| 2522 |
+
"step": 1790
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 7.8384279475982535,
|
| 2526 |
+
"grad_norm": 1.5318894677152983,
|
| 2527 |
+
"learning_rate": 2.7140712740884376e-05,
|
| 2528 |
+
"loss": 0.7094,
|
| 2529 |
+
"step": 1795
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 7.860262008733624,
|
| 2533 |
+
"grad_norm": 1.5006081477924398,
|
| 2534 |
+
"learning_rate": 2.6620799801829765e-05,
|
| 2535 |
+
"loss": 0.7356,
|
| 2536 |
+
"step": 1800
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 7.882096069868996,
|
| 2540 |
+
"grad_norm": 1.4913103972222401,
|
| 2541 |
+
"learning_rate": 2.610514926708285e-05,
|
| 2542 |
+
"loss": 0.7563,
|
| 2543 |
+
"step": 1805
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 7.903930131004367,
|
| 2547 |
+
"grad_norm": 1.5640971425352013,
|
| 2548 |
+
"learning_rate": 2.5593791089423858e-05,
|
| 2549 |
+
"loss": 0.6974,
|
| 2550 |
+
"step": 1810
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 7.925764192139738,
|
| 2554 |
+
"grad_norm": 1.5288946416263425,
|
| 2555 |
+
"learning_rate": 2.5086754972301384e-05,
|
| 2556 |
+
"loss": 0.7597,
|
| 2557 |
+
"step": 1815
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 7.9475982532751095,
|
| 2561 |
+
"grad_norm": 1.4850580236939783,
|
| 2562 |
+
"learning_rate": 2.4584070368106928e-05,
|
| 2563 |
+
"loss": 0.731,
|
| 2564 |
+
"step": 1820
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 7.96943231441048,
|
| 2568 |
+
"grad_norm": 1.5394321618591782,
|
| 2569 |
+
"learning_rate": 2.4085766476463967e-05,
|
| 2570 |
+
"loss": 0.712,
|
| 2571 |
+
"step": 1825
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 7.991266375545852,
|
| 2575 |
+
"grad_norm": 1.6849009773020913,
|
| 2576 |
+
"learning_rate": 2.3591872242532066e-05,
|
| 2577 |
+
"loss": 0.7327,
|
| 2578 |
+
"step": 1830
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 8.013100436681222,
|
| 2582 |
+
"grad_norm": 1.4739518847631212,
|
| 2583 |
+
"learning_rate": 2.310241635532531e-05,
|
| 2584 |
+
"loss": 0.6777,
|
| 2585 |
+
"step": 1835
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 8.034934497816593,
|
| 2589 |
+
"grad_norm": 1.4710344726002955,
|
| 2590 |
+
"learning_rate": 2.2617427246045973e-05,
|
| 2591 |
+
"loss": 0.5886,
|
| 2592 |
+
"step": 1840
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 8.056768558951966,
|
| 2596 |
+
"grad_norm": 1.8620287227429924,
|
| 2597 |
+
"learning_rate": 2.2136933086432955e-05,
|
| 2598 |
+
"loss": 0.6258,
|
| 2599 |
+
"step": 1845
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 8.078602620087336,
|
| 2603 |
+
"grad_norm": 1.4930799825826104,
|
| 2604 |
+
"learning_rate": 2.1660961787125388e-05,
|
| 2605 |
+
"loss": 0.6041,
|
| 2606 |
+
"step": 1850
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 8.100436681222707,
|
| 2610 |
+
"grad_norm": 1.469008850941141,
|
| 2611 |
+
"learning_rate": 2.1189540996041313e-05,
|
| 2612 |
+
"loss": 0.647,
|
| 2613 |
+
"step": 1855
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 8.12227074235808,
|
| 2617 |
+
"grad_norm": 1.6154095310822976,
|
| 2618 |
+
"learning_rate": 2.0722698096771832e-05,
|
| 2619 |
+
"loss": 0.5866,
|
| 2620 |
+
"step": 1860
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 8.14410480349345,
|
| 2624 |
+
"grad_norm": 1.7217187185906804,
|
| 2625 |
+
"learning_rate": 2.026046020699035e-05,
|
| 2626 |
+
"loss": 0.6718,
|
| 2627 |
+
"step": 1865
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 8.16593886462882,
|
| 2631 |
+
"grad_norm": 1.5816769270478201,
|
| 2632 |
+
"learning_rate": 1.980285417687735e-05,
|
| 2633 |
+
"loss": 0.6303,
|
| 2634 |
+
"step": 1870
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 8.187772925764191,
|
| 2638 |
+
"grad_norm": 1.5330680888020691,
|
| 2639 |
+
"learning_rate": 1.9349906587560862e-05,
|
| 2640 |
+
"loss": 0.6166,
|
| 2641 |
+
"step": 1875
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 8.209606986899564,
|
| 2645 |
+
"grad_norm": 1.4811035124378678,
|
| 2646 |
+
"learning_rate": 1.8901643749572374e-05,
|
| 2647 |
+
"loss": 0.6245,
|
| 2648 |
+
"step": 1880
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 8.231441048034934,
|
| 2652 |
+
"grad_norm": 1.6432943675024339,
|
| 2653 |
+
"learning_rate": 1.8458091701318504e-05,
|
| 2654 |
+
"loss": 0.6261,
|
| 2655 |
+
"step": 1885
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 8.253275109170305,
|
| 2659 |
+
"grad_norm": 1.6763677385421005,
|
| 2660 |
+
"learning_rate": 1.801927620756847e-05,
|
| 2661 |
+
"loss": 0.6468,
|
| 2662 |
+
"step": 1890
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 8.275109170305678,
|
| 2666 |
+
"grad_norm": 1.5425807262553166,
|
| 2667 |
+
"learning_rate": 1.7585222757957576e-05,
|
| 2668 |
+
"loss": 0.6059,
|
| 2669 |
+
"step": 1895
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 8.296943231441048,
|
| 2673 |
+
"grad_norm": 1.7341819757802275,
|
| 2674 |
+
"learning_rate": 1.7155956565506547e-05,
|
| 2675 |
+
"loss": 0.6728,
|
| 2676 |
+
"step": 1900
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 8.318777292576419,
|
| 2680 |
+
"grad_norm": 1.4483276727621572,
|
| 2681 |
+
"learning_rate": 1.6731502565156875e-05,
|
| 2682 |
+
"loss": 0.6033,
|
| 2683 |
+
"step": 1905
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 8.34061135371179,
|
| 2687 |
+
"grad_norm": 1.5596252995782947,
|
| 2688 |
+
"learning_rate": 1.6311885412322602e-05,
|
| 2689 |
+
"loss": 0.63,
|
| 2690 |
+
"step": 1910
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 8.362445414847162,
|
| 2694 |
+
"grad_norm": 1.747814876036389,
|
| 2695 |
+
"learning_rate": 1.5897129481457996e-05,
|
| 2696 |
+
"loss": 0.5621,
|
| 2697 |
+
"step": 1915
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 8.384279475982533,
|
| 2701 |
+
"grad_norm": 1.5550499771354138,
|
| 2702 |
+
"learning_rate": 1.5487258864641717e-05,
|
| 2703 |
+
"loss": 0.6306,
|
| 2704 |
+
"step": 1920
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 8.406113537117903,
|
| 2708 |
+
"grad_norm": 1.5708977251660243,
|
| 2709 |
+
"learning_rate": 1.50822973701775e-05,
|
| 2710 |
+
"loss": 0.6281,
|
| 2711 |
+
"step": 1925
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 8.427947598253276,
|
| 2715 |
+
"grad_norm": 1.5052322890768695,
|
| 2716 |
+
"learning_rate": 1.4682268521211073e-05,
|
| 2717 |
+
"loss": 0.5805,
|
| 2718 |
+
"step": 1930
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 8.449781659388647,
|
| 2722 |
+
"grad_norm": 1.5682376306714874,
|
| 2723 |
+
"learning_rate": 1.4287195554363718e-05,
|
| 2724 |
+
"loss": 0.6103,
|
| 2725 |
+
"step": 1935
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 8.471615720524017,
|
| 2729 |
+
"grad_norm": 1.5926672273219815,
|
| 2730 |
+
"learning_rate": 1.3897101418382663e-05,
|
| 2731 |
+
"loss": 0.6086,
|
| 2732 |
+
"step": 1940
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 8.493449781659388,
|
| 2736 |
+
"grad_norm": 1.6920842457267133,
|
| 2737 |
+
"learning_rate": 1.3512008772807993e-05,
|
| 2738 |
+
"loss": 0.6075,
|
| 2739 |
+
"step": 1945
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 8.51528384279476,
|
| 2743 |
+
"grad_norm": 1.7066301478594386,
|
| 2744 |
+
"learning_rate": 1.3131939986656305e-05,
|
| 2745 |
+
"loss": 0.6037,
|
| 2746 |
+
"step": 1950
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 8.537117903930131,
|
| 2750 |
+
"grad_norm": 1.6263309583877439,
|
| 2751 |
+
"learning_rate": 1.2756917137121527e-05,
|
| 2752 |
+
"loss": 0.6137,
|
| 2753 |
+
"step": 1955
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 8.558951965065502,
|
| 2757 |
+
"grad_norm": 1.63476708848148,
|
| 2758 |
+
"learning_rate": 1.2386962008292413e-05,
|
| 2759 |
+
"loss": 0.5858,
|
| 2760 |
+
"step": 1960
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 8.580786026200874,
|
| 2764 |
+
"grad_norm": 1.558434256481925,
|
| 2765 |
+
"learning_rate": 1.2022096089887191e-05,
|
| 2766 |
+
"loss": 0.6426,
|
| 2767 |
+
"step": 1965
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 8.602620087336245,
|
| 2771 |
+
"grad_norm": 1.551619834712178,
|
| 2772 |
+
"learning_rate": 1.1662340576005216e-05,
|
| 2773 |
+
"loss": 0.6084,
|
| 2774 |
+
"step": 1970
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 8.624454148471616,
|
| 2778 |
+
"grad_norm": 1.555097964766268,
|
| 2779 |
+
"learning_rate": 1.130771636389596e-05,
|
| 2780 |
+
"loss": 0.6687,
|
| 2781 |
+
"step": 1975
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 8.646288209606986,
|
| 2785 |
+
"grad_norm": 1.6643047208497839,
|
| 2786 |
+
"learning_rate": 1.0958244052745126e-05,
|
| 2787 |
+
"loss": 0.6155,
|
| 2788 |
+
"step": 1980
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 8.668122270742359,
|
| 2792 |
+
"grad_norm": 1.611005114630636,
|
| 2793 |
+
"learning_rate": 1.0613943942478e-05,
|
| 2794 |
+
"loss": 0.6089,
|
| 2795 |
+
"step": 1985
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 8.68995633187773,
|
| 2799 |
+
"grad_norm": 1.5244955674562928,
|
| 2800 |
+
"learning_rate": 1.0274836032580415e-05,
|
| 2801 |
+
"loss": 0.6487,
|
| 2802 |
+
"step": 1990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 8.7117903930131,
|
| 2806 |
+
"grad_norm": 1.6030907080208872,
|
| 2807 |
+
"learning_rate": 9.940940020936951e-06,
|
| 2808 |
+
"loss": 0.6293,
|
| 2809 |
+
"step": 1995
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 8.733624454148472,
|
| 2813 |
+
"grad_norm": 1.5700446980317904,
|
| 2814 |
+
"learning_rate": 9.612275302686713e-06,
|
| 2815 |
+
"loss": 0.6326,
|
| 2816 |
+
"step": 2000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 8.755458515283843,
|
| 2820 |
+
"grad_norm": 1.5394752226532917,
|
| 2821 |
+
"learning_rate": 9.288860969096857e-06,
|
| 2822 |
+
"loss": 0.6107,
|
| 2823 |
+
"step": 2005
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 8.777292576419214,
|
| 2827 |
+
"grad_norm": 1.667355715027956,
|
| 2828 |
+
"learning_rate": 8.970715806453489e-06,
|
| 2829 |
+
"loss": 0.636,
|
| 2830 |
+
"step": 2010
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 8.799126637554584,
|
| 2834 |
+
"grad_norm": 1.7421757272877927,
|
| 2835 |
+
"learning_rate": 8.657858294970412e-06,
|
| 2836 |
+
"loss": 0.6358,
|
| 2837 |
+
"step": 2015
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 8.820960698689957,
|
| 2841 |
+
"grad_norm": 1.48526713960321,
|
| 2842 |
+
"learning_rate": 8.350306607715774e-06,
|
| 2843 |
+
"loss": 0.6456,
|
| 2844 |
+
"step": 2020
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 8.842794759825328,
|
| 2848 |
+
"grad_norm": 1.7150442937256956,
|
| 2849 |
+
"learning_rate": 8.048078609556386e-06,
|
| 2850 |
+
"loss": 0.6443,
|
| 2851 |
+
"step": 2025
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 8.864628820960698,
|
| 2855 |
+
"grad_norm": 1.6511037454437418,
|
| 2856 |
+
"learning_rate": 7.751191856119932e-06,
|
| 2857 |
+
"loss": 0.671,
|
| 2858 |
+
"step": 2030
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 8.886462882096069,
|
| 2862 |
+
"grad_norm": 1.6987081532253472,
|
| 2863 |
+
"learning_rate": 7.459663592775334e-06,
|
| 2864 |
+
"loss": 0.6577,
|
| 2865 |
+
"step": 2035
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 8.908296943231441,
|
| 2869 |
+
"grad_norm": 1.534952723839,
|
| 2870 |
+
"learning_rate": 7.173510753630919e-06,
|
| 2871 |
+
"loss": 0.6233,
|
| 2872 |
+
"step": 2040
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 8.930131004366812,
|
| 2876 |
+
"grad_norm": 1.641736331583932,
|
| 2877 |
+
"learning_rate": 6.892749960550815e-06,
|
| 2878 |
+
"loss": 0.6289,
|
| 2879 |
+
"step": 2045
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 8.951965065502183,
|
| 2883 |
+
"grad_norm": 1.5438647483260877,
|
| 2884 |
+
"learning_rate": 6.6173975221893615e-06,
|
| 2885 |
+
"loss": 0.5888,
|
| 2886 |
+
"step": 2050
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 8.973799126637555,
|
| 2890 |
+
"grad_norm": 1.5243321966646095,
|
| 2891 |
+
"learning_rate": 6.347469433043851e-06,
|
| 2892 |
+
"loss": 0.6707,
|
| 2893 |
+
"step": 2055
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 8.995633187772926,
|
| 2897 |
+
"grad_norm": 1.6062799594767991,
|
| 2898 |
+
"learning_rate": 6.082981372525487e-06,
|
| 2899 |
+
"loss": 0.5971,
|
| 2900 |
+
"step": 2060
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 9.017467248908297,
|
| 2904 |
+
"grad_norm": 1.462091614630793,
|
| 2905 |
+
"learning_rate": 5.823948704048443e-06,
|
| 2906 |
+
"loss": 0.5631,
|
| 2907 |
+
"step": 2065
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 9.039301310043669,
|
| 2911 |
+
"grad_norm": 1.4923902769131498,
|
| 2912 |
+
"learning_rate": 5.570386474137623e-06,
|
| 2913 |
+
"loss": 0.5617,
|
| 2914 |
+
"step": 2070
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 9.06113537117904,
|
| 2918 |
+
"grad_norm": 1.6248158642620525,
|
| 2919 |
+
"learning_rate": 5.322309411554582e-06,
|
| 2920 |
+
"loss": 0.6111,
|
| 2921 |
+
"step": 2075
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 9.08296943231441,
|
| 2925 |
+
"grad_norm": 1.6615679454812948,
|
| 2926 |
+
"learning_rate": 5.0797319264419105e-06,
|
| 2927 |
+
"loss": 0.563,
|
| 2928 |
+
"step": 2080
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 9.104803493449781,
|
| 2932 |
+
"grad_norm": 1.6126706246375568,
|
| 2933 |
+
"learning_rate": 4.84266810948627e-06,
|
| 2934 |
+
"loss": 0.5686,
|
| 2935 |
+
"step": 2085
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 9.126637554585153,
|
| 2939 |
+
"grad_norm": 1.5809643369161743,
|
| 2940 |
+
"learning_rate": 4.611131731099905e-06,
|
| 2941 |
+
"loss": 0.5533,
|
| 2942 |
+
"step": 2090
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 9.148471615720524,
|
| 2946 |
+
"grad_norm": 1.6161716990207524,
|
| 2947 |
+
"learning_rate": 4.385136240620657e-06,
|
| 2948 |
+
"loss": 0.5962,
|
| 2949 |
+
"step": 2095
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 9.170305676855895,
|
| 2953 |
+
"grad_norm": 1.5154717234389132,
|
| 2954 |
+
"learning_rate": 4.164694765530841e-06,
|
| 2955 |
+
"loss": 0.5946,
|
| 2956 |
+
"step": 2100
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 9.192139737991265,
|
| 2960 |
+
"grad_norm": 1.5768496797034472,
|
| 2961 |
+
"learning_rate": 3.94982011069468e-06,
|
| 2962 |
+
"loss": 0.5383,
|
| 2963 |
+
"step": 2105
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 9.213973799126638,
|
| 2967 |
+
"grad_norm": 1.6583634682510404,
|
| 2968 |
+
"learning_rate": 3.7405247576144054e-06,
|
| 2969 |
+
"loss": 0.6018,
|
| 2970 |
+
"step": 2110
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 9.235807860262009,
|
| 2974 |
+
"grad_norm": 1.5801691223578864,
|
| 2975 |
+
"learning_rate": 3.5368208637053702e-06,
|
| 2976 |
+
"loss": 0.5564,
|
| 2977 |
+
"step": 2115
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 9.25764192139738,
|
| 2981 |
+
"grad_norm": 1.702349073192532,
|
| 2982 |
+
"learning_rate": 3.338720261589823e-06,
|
| 2983 |
+
"loss": 0.578,
|
| 2984 |
+
"step": 2120
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 9.279475982532752,
|
| 2988 |
+
"grad_norm": 1.5184620025021562,
|
| 2989 |
+
"learning_rate": 3.146234458409525e-06,
|
| 2990 |
+
"loss": 0.5649,
|
| 2991 |
+
"step": 2125
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 9.301310043668122,
|
| 2995 |
+
"grad_norm": 1.4782870440067606,
|
| 2996 |
+
"learning_rate": 2.959374635157364e-06,
|
| 2997 |
+
"loss": 0.5708,
|
| 2998 |
+
"step": 2130
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 9.323144104803493,
|
| 3002 |
+
"grad_norm": 1.5908300327141753,
|
| 3003 |
+
"learning_rate": 2.7781516460279157e-06,
|
| 3004 |
+
"loss": 0.5719,
|
| 3005 |
+
"step": 2135
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 9.344978165938866,
|
| 3009 |
+
"grad_norm": 1.653067982991365,
|
| 3010 |
+
"learning_rate": 2.6025760177869063e-06,
|
| 3011 |
+
"loss": 0.5914,
|
| 3012 |
+
"step": 2140
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 9.366812227074236,
|
| 3016 |
+
"grad_norm": 1.5475103396862675,
|
| 3017 |
+
"learning_rate": 2.4326579491597333e-06,
|
| 3018 |
+
"loss": 0.5931,
|
| 3019 |
+
"step": 2145
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 9.388646288209607,
|
| 3023 |
+
"grad_norm": 1.6475600779910462,
|
| 3024 |
+
"learning_rate": 2.2684073102391066e-06,
|
| 3025 |
+
"loss": 0.5978,
|
| 3026 |
+
"step": 2150
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 9.410480349344978,
|
| 3030 |
+
"grad_norm": 1.5891384582888524,
|
| 3031 |
+
"learning_rate": 2.1098336419116625e-06,
|
| 3032 |
+
"loss": 0.5656,
|
| 3033 |
+
"step": 2155
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 9.43231441048035,
|
| 3037 |
+
"grad_norm": 1.5178759944544706,
|
| 3038 |
+
"learning_rate": 1.956946155303785e-06,
|
| 3039 |
+
"loss": 0.6198,
|
| 3040 |
+
"step": 2160
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 9.45414847161572,
|
| 3044 |
+
"grad_norm": 1.641246543950594,
|
| 3045 |
+
"learning_rate": 1.809753731246544e-06,
|
| 3046 |
+
"loss": 0.5829,
|
| 3047 |
+
"step": 2165
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 9.475982532751091,
|
| 3051 |
+
"grad_norm": 1.665331469610374,
|
| 3052 |
+
"learning_rate": 1.6682649197598433e-06,
|
| 3053 |
+
"loss": 0.5871,
|
| 3054 |
+
"step": 2170
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 9.497816593886462,
|
| 3058 |
+
"grad_norm": 1.823654408381571,
|
| 3059 |
+
"learning_rate": 1.5324879395557933e-06,
|
| 3060 |
+
"loss": 0.5906,
|
| 3061 |
+
"step": 2175
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 9.519650655021834,
|
| 3065 |
+
"grad_norm": 1.564912693530831,
|
| 3066 |
+
"learning_rate": 1.4024306775612283e-06,
|
| 3067 |
+
"loss": 0.6207,
|
| 3068 |
+
"step": 2180
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 9.541484716157205,
|
| 3072 |
+
"grad_norm": 1.5863861199272236,
|
| 3073 |
+
"learning_rate": 1.2781006884596825e-06,
|
| 3074 |
+
"loss": 0.6267,
|
| 3075 |
+
"step": 2185
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 9.563318777292576,
|
| 3079 |
+
"grad_norm": 1.6063472205433305,
|
| 3080 |
+
"learning_rate": 1.1595051942524637e-06,
|
| 3081 |
+
"loss": 0.5755,
|
| 3082 |
+
"step": 2190
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 9.585152838427948,
|
| 3086 |
+
"grad_norm": 1.5366202264694289,
|
| 3087 |
+
"learning_rate": 1.0466510838392229e-06,
|
| 3088 |
+
"loss": 0.5384,
|
| 3089 |
+
"step": 2195
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 9.606986899563319,
|
| 3093 |
+
"grad_norm": 1.61763818006577,
|
| 3094 |
+
"learning_rate": 9.395449126177291e-07,
|
| 3095 |
+
"loss": 0.6435,
|
| 3096 |
+
"step": 2200
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 9.62882096069869,
|
| 3100 |
+
"grad_norm": 1.5959225303743334,
|
| 3101 |
+
"learning_rate": 8.381929021031409e-07,
|
| 3102 |
+
"loss": 0.5587,
|
| 3103 |
+
"step": 2205
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 9.65065502183406,
|
| 3107 |
+
"grad_norm": 1.5317193865637022,
|
| 3108 |
+
"learning_rate": 7.426009395665734e-07,
|
| 3109 |
+
"loss": 0.6166,
|
| 3110 |
+
"step": 2210
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 9.672489082969433,
|
| 3114 |
+
"grad_norm": 1.7050859904255076,
|
| 3115 |
+
"learning_rate": 6.527745776931382e-07,
|
| 3116 |
+
"loss": 0.6078,
|
| 3117 |
+
"step": 2215
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 9.694323144104803,
|
| 3121 |
+
"grad_norm": 1.626127532872475,
|
| 3122 |
+
"learning_rate": 5.687190342594239e-07,
|
| 3123 |
+
"loss": 0.6022,
|
| 3124 |
+
"step": 2220
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 9.716157205240174,
|
| 3128 |
+
"grad_norm": 1.6752520068014993,
|
| 3129 |
+
"learning_rate": 4.904391918303608e-07,
|
| 3130 |
+
"loss": 0.6126,
|
| 3131 |
+
"step": 2225
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 9.737991266375547,
|
| 3135 |
+
"grad_norm": 1.6226696728292287,
|
| 3136 |
+
"learning_rate": 4.1793959747565836e-07,
|
| 3137 |
+
"loss": 0.5671,
|
| 3138 |
+
"step": 2230
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 9.759825327510917,
|
| 3142 |
+
"grad_norm": 1.6182868533567345,
|
| 3143 |
+
"learning_rate": 3.5122446250562825e-07,
|
| 3144 |
+
"loss": 0.5859,
|
| 3145 |
+
"step": 2235
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 9.781659388646288,
|
| 3149 |
+
"grad_norm": 1.686162675556006,
|
| 3150 |
+
"learning_rate": 2.902976622265907e-07,
|
| 3151 |
+
"loss": 0.5634,
|
| 3152 |
+
"step": 2240
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 9.803493449781659,
|
| 3156 |
+
"grad_norm": 1.5169542450593143,
|
| 3157 |
+
"learning_rate": 2.3516273571577708e-07,
|
| 3158 |
+
"loss": 0.5578,
|
| 3159 |
+
"step": 2245
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 9.825327510917031,
|
| 3163 |
+
"grad_norm": 1.6287222229266196,
|
| 3164 |
+
"learning_rate": 1.8582288561573847e-07,
|
| 3165 |
+
"loss": 0.5543,
|
| 3166 |
+
"step": 2250
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 9.847161572052402,
|
| 3170 |
+
"grad_norm": 1.4791363351235698,
|
| 3171 |
+
"learning_rate": 1.4228097794828366e-07,
|
| 3172 |
+
"loss": 0.5705,
|
| 3173 |
+
"step": 2255
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 9.868995633187772,
|
| 3177 |
+
"grad_norm": 1.5782969298269662,
|
| 3178 |
+
"learning_rate": 1.045395419480677e-07,
|
| 3179 |
+
"loss": 0.5742,
|
| 3180 |
+
"step": 2260
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 9.890829694323145,
|
| 3184 |
+
"grad_norm": 1.5934772169541918,
|
| 3185 |
+
"learning_rate": 7.260076991560949e-08,
|
| 3186 |
+
"loss": 0.6509,
|
| 3187 |
+
"step": 2265
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 9.912663755458516,
|
| 3191 |
+
"grad_norm": 1.5072353913111483,
|
| 3192 |
+
"learning_rate": 4.646651708998251e-08,
|
| 3193 |
+
"loss": 0.5509,
|
| 3194 |
+
"step": 2270
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 9.934497816593886,
|
| 3198 |
+
"grad_norm": 1.5448684833958992,
|
| 3199 |
+
"learning_rate": 2.6138301541056564e-08,
|
| 3200 |
+
"loss": 0.5666,
|
| 3201 |
+
"step": 2275
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 9.956331877729257,
|
| 3205 |
+
"grad_norm": 1.504682951530182,
|
| 3206 |
+
"learning_rate": 1.1617304081268376e-08,
|
| 3207 |
+
"loss": 0.5865,
|
| 3208 |
+
"step": 2280
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 9.97816593886463,
|
| 3212 |
+
"grad_norm": 1.5933513163764192,
|
| 3213 |
+
"learning_rate": 2.9043681970875035e-09,
|
| 3214 |
+
"loss": 0.5653,
|
| 3215 |
+
"step": 2285
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 10.0,
|
| 3219 |
+
"grad_norm": 1.5918112804762306,
|
| 3220 |
+
"learning_rate": 0.0,
|
| 3221 |
+
"loss": 0.5502,
|
| 3222 |
+
"step": 2290
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 10.0,
|
| 3226 |
+
"step": 2290,
|
| 3227 |
+
"total_flos": 5211452815179776.0,
|
| 3228 |
+
"train_loss": 1.1147898718779785,
|
| 3229 |
+
"train_runtime": 4502.6314,
|
| 3230 |
+
"train_samples_per_second": 32.519,
|
| 3231 |
+
"train_steps_per_second": 0.509
|
| 3232 |
+
}
|
| 3233 |
+
],
|
| 3234 |
+
"logging_steps": 5,
|
| 3235 |
+
"max_steps": 2290,
|
| 3236 |
+
"num_input_tokens_seen": 0,
|
| 3237 |
+
"num_train_epochs": 10,
|
| 3238 |
+
"save_steps": 100,
|
| 3239 |
+
"stateful_callbacks": {
|
| 3240 |
+
"TrainerControl": {
|
| 3241 |
+
"args": {
|
| 3242 |
+
"should_epoch_stop": false,
|
| 3243 |
+
"should_evaluate": false,
|
| 3244 |
+
"should_log": false,
|
| 3245 |
+
"should_save": true,
|
| 3246 |
+
"should_training_stop": true
|
| 3247 |
+
},
|
| 3248 |
+
"attributes": {}
|
| 3249 |
+
}
|
| 3250 |
+
},
|
| 3251 |
+
"total_flos": 5211452815179776.0,
|
| 3252 |
+
"train_batch_size": 8,
|
| 3253 |
+
"trial_name": null,
|
| 3254 |
+
"trial_params": null
|
| 3255 |
+
}
|