Model save
Browse files- README.md +66 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- runs/Aug19_01-30-44_simpo-8-0-0/events.out.tfevents.1724056267.simpo-8-0-0.29615.0 +2 -2
- train_results.json +9 -0
- trainer_state.json +3088 -0
README.md
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: mistralai/Mistral-7B-v0.1
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- sft
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
datasets:
|
| 9 |
+
- generator
|
| 10 |
+
model-index:
|
| 11 |
+
- name: zephyr-7b-sft-full
|
| 12 |
+
results: []
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 16 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 17 |
+
|
| 18 |
+
# zephyr-7b-sft-full
|
| 19 |
+
|
| 20 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the generator dataset.
|
| 21 |
+
It achieves the following results on the evaluation set:
|
| 22 |
+
- Loss: 0.9540
|
| 23 |
+
|
| 24 |
+
## Model description
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Intended uses & limitations
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training and evaluation data
|
| 33 |
+
|
| 34 |
+
More information needed
|
| 35 |
+
|
| 36 |
+
## Training procedure
|
| 37 |
+
|
| 38 |
+
### Training hyperparameters
|
| 39 |
+
|
| 40 |
+
The following hyperparameters were used during training:
|
| 41 |
+
- learning_rate: 2e-05
|
| 42 |
+
- train_batch_size: 8
|
| 43 |
+
- eval_batch_size: 8
|
| 44 |
+
- seed: 42
|
| 45 |
+
- distributed_type: multi-GPU
|
| 46 |
+
- num_devices: 8
|
| 47 |
+
- total_train_batch_size: 64
|
| 48 |
+
- total_eval_batch_size: 64
|
| 49 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 50 |
+
- lr_scheduler_type: cosine
|
| 51 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 52 |
+
- num_epochs: 1
|
| 53 |
+
|
| 54 |
+
### Training results
|
| 55 |
+
|
| 56 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 57 |
+
|:-------------:|:-----:|:----:|:---------------:|
|
| 58 |
+
| 0.973 | 1.0 | 2167 | 0.9540 |
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
### Framework versions
|
| 62 |
+
|
| 63 |
+
- Transformers 4.44.0
|
| 64 |
+
- Pytorch 2.1.2+cu121
|
| 65 |
+
- Datasets 2.21.0
|
| 66 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 453725713858560.0,
|
| 4 |
+
"train_loss": 1.0076359760722753,
|
| 5 |
+
"train_runtime": 13620.7718,
|
| 6 |
+
"train_samples": 207864,
|
| 7 |
+
"train_samples_per_second": 10.182,
|
| 8 |
+
"train_steps_per_second": 0.159
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.44.0"
|
| 6 |
+
}
|
runs/Aug19_01-30-44_simpo-8-0-0/events.out.tfevents.1724056267.simpo-8-0-0.29615.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:43f7d9ab526565bf90a7e066c227ba26ef7cc416ef943cbfead2f5121f938367
|
| 3 |
+
size 97594
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 453725713858560.0,
|
| 4 |
+
"train_loss": 1.0076359760722753,
|
| 5 |
+
"train_runtime": 13620.7718,
|
| 6 |
+
"train_samples": 207864,
|
| 7 |
+
"train_samples_per_second": 10.182,
|
| 8 |
+
"train_steps_per_second": 0.159
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3088 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2167,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.00046146746654360867,
|
| 13 |
+
"grad_norm": 9.51411938145824,
|
| 14 |
+
"learning_rate": 9.216589861751152e-08,
|
| 15 |
+
"loss": 1.1509,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0023073373327180432,
|
| 20 |
+
"grad_norm": 9.594129890510615,
|
| 21 |
+
"learning_rate": 4.608294930875577e-07,
|
| 22 |
+
"loss": 1.1369,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0046146746654360865,
|
| 27 |
+
"grad_norm": 5.144220521250804,
|
| 28 |
+
"learning_rate": 9.216589861751154e-07,
|
| 29 |
+
"loss": 1.0949,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.00692201199815413,
|
| 34 |
+
"grad_norm": 3.40382562328973,
|
| 35 |
+
"learning_rate": 1.382488479262673e-06,
|
| 36 |
+
"loss": 1.0191,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.009229349330872173,
|
| 41 |
+
"grad_norm": 3.215915212337555,
|
| 42 |
+
"learning_rate": 1.8433179723502307e-06,
|
| 43 |
+
"loss": 1.0296,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.011536686663590217,
|
| 48 |
+
"grad_norm": 2.650742433924389,
|
| 49 |
+
"learning_rate": 2.3041474654377884e-06,
|
| 50 |
+
"loss": 0.976,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.01384402399630826,
|
| 55 |
+
"grad_norm": 2.7902393320247687,
|
| 56 |
+
"learning_rate": 2.764976958525346e-06,
|
| 57 |
+
"loss": 1.0089,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.016151361329026302,
|
| 62 |
+
"grad_norm": 2.4654209985126148,
|
| 63 |
+
"learning_rate": 3.225806451612903e-06,
|
| 64 |
+
"loss": 0.9951,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.018458698661744346,
|
| 69 |
+
"grad_norm": 2.527650306992279,
|
| 70 |
+
"learning_rate": 3.6866359447004615e-06,
|
| 71 |
+
"loss": 0.9988,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.02076603599446239,
|
| 76 |
+
"grad_norm": 2.655474988385568,
|
| 77 |
+
"learning_rate": 4.147465437788019e-06,
|
| 78 |
+
"loss": 1.004,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.023073373327180433,
|
| 83 |
+
"grad_norm": 2.537803872777302,
|
| 84 |
+
"learning_rate": 4.608294930875577e-06,
|
| 85 |
+
"loss": 0.9863,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.025380710659898477,
|
| 90 |
+
"grad_norm": 2.6464740488754366,
|
| 91 |
+
"learning_rate": 5.0691244239631346e-06,
|
| 92 |
+
"loss": 0.9694,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.02768804799261652,
|
| 97 |
+
"grad_norm": 2.7507142522196566,
|
| 98 |
+
"learning_rate": 5.529953917050692e-06,
|
| 99 |
+
"loss": 0.9688,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.029995385325334564,
|
| 104 |
+
"grad_norm": 2.836749382191462,
|
| 105 |
+
"learning_rate": 5.9907834101382485e-06,
|
| 106 |
+
"loss": 0.9968,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.032302722658052604,
|
| 111 |
+
"grad_norm": 3.048439725741993,
|
| 112 |
+
"learning_rate": 6.451612903225806e-06,
|
| 113 |
+
"loss": 1.0023,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.03461005999077065,
|
| 118 |
+
"grad_norm": 3.072103382083384,
|
| 119 |
+
"learning_rate": 6.912442396313365e-06,
|
| 120 |
+
"loss": 0.9909,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.03691739732348869,
|
| 125 |
+
"grad_norm": 2.4206649546182386,
|
| 126 |
+
"learning_rate": 7.373271889400923e-06,
|
| 127 |
+
"loss": 1.0277,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.03922473465620674,
|
| 132 |
+
"grad_norm": 2.6546178755277254,
|
| 133 |
+
"learning_rate": 7.83410138248848e-06,
|
| 134 |
+
"loss": 1.0123,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.04153207198892478,
|
| 139 |
+
"grad_norm": 2.3888921486796058,
|
| 140 |
+
"learning_rate": 8.294930875576038e-06,
|
| 141 |
+
"loss": 0.9688,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.043839409321642826,
|
| 146 |
+
"grad_norm": 2.32499977135365,
|
| 147 |
+
"learning_rate": 8.755760368663595e-06,
|
| 148 |
+
"loss": 0.9903,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.046146746654360866,
|
| 153 |
+
"grad_norm": 2.392443248966377,
|
| 154 |
+
"learning_rate": 9.216589861751153e-06,
|
| 155 |
+
"loss": 0.9893,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.048454083987078914,
|
| 160 |
+
"grad_norm": 2.680320857358668,
|
| 161 |
+
"learning_rate": 9.67741935483871e-06,
|
| 162 |
+
"loss": 0.9846,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.050761421319796954,
|
| 167 |
+
"grad_norm": 2.532469504703905,
|
| 168 |
+
"learning_rate": 1.0138248847926269e-05,
|
| 169 |
+
"loss": 1.0089,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.053068758652515,
|
| 174 |
+
"grad_norm": 3.1046898569172945,
|
| 175 |
+
"learning_rate": 1.0599078341013826e-05,
|
| 176 |
+
"loss": 1.0266,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.05537609598523304,
|
| 181 |
+
"grad_norm": 2.5574603903328743,
|
| 182 |
+
"learning_rate": 1.1059907834101385e-05,
|
| 183 |
+
"loss": 1.002,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.05768343331795108,
|
| 188 |
+
"grad_norm": 2.452071743693235,
|
| 189 |
+
"learning_rate": 1.152073732718894e-05,
|
| 190 |
+
"loss": 1.03,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.05999077065066913,
|
| 195 |
+
"grad_norm": 2.418236004711402,
|
| 196 |
+
"learning_rate": 1.1981566820276497e-05,
|
| 197 |
+
"loss": 1.006,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.06229810798338717,
|
| 202 |
+
"grad_norm": 2.378674843033103,
|
| 203 |
+
"learning_rate": 1.2442396313364056e-05,
|
| 204 |
+
"loss": 0.9717,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.06460544531610521,
|
| 209 |
+
"grad_norm": 2.288433336347559,
|
| 210 |
+
"learning_rate": 1.2903225806451613e-05,
|
| 211 |
+
"loss": 1.0247,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.06691278264882326,
|
| 216 |
+
"grad_norm": 2.7611308296401282,
|
| 217 |
+
"learning_rate": 1.3364055299539171e-05,
|
| 218 |
+
"loss": 1.003,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.0692201199815413,
|
| 223 |
+
"grad_norm": 2.421569626628109,
|
| 224 |
+
"learning_rate": 1.382488479262673e-05,
|
| 225 |
+
"loss": 1.0197,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.07152745731425934,
|
| 230 |
+
"grad_norm": 2.3978303307399247,
|
| 231 |
+
"learning_rate": 1.4285714285714287e-05,
|
| 232 |
+
"loss": 0.9898,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.07383479464697738,
|
| 237 |
+
"grad_norm": 2.980112762291027,
|
| 238 |
+
"learning_rate": 1.4746543778801846e-05,
|
| 239 |
+
"loss": 1.0275,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.07614213197969544,
|
| 244 |
+
"grad_norm": 2.6176775912790404,
|
| 245 |
+
"learning_rate": 1.5207373271889403e-05,
|
| 246 |
+
"loss": 1.0382,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.07844946931241348,
|
| 251 |
+
"grad_norm": 2.44164739077761,
|
| 252 |
+
"learning_rate": 1.566820276497696e-05,
|
| 253 |
+
"loss": 1.0236,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.08075680664513152,
|
| 258 |
+
"grad_norm": 2.404929271261824,
|
| 259 |
+
"learning_rate": 1.6129032258064517e-05,
|
| 260 |
+
"loss": 1.0304,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.08306414397784956,
|
| 265 |
+
"grad_norm": 2.3566906067196105,
|
| 266 |
+
"learning_rate": 1.6589861751152075e-05,
|
| 267 |
+
"loss": 1.0355,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.0853714813105676,
|
| 272 |
+
"grad_norm": 2.6391147388298246,
|
| 273 |
+
"learning_rate": 1.705069124423963e-05,
|
| 274 |
+
"loss": 1.0417,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.08767881864328565,
|
| 279 |
+
"grad_norm": 2.684091771591401,
|
| 280 |
+
"learning_rate": 1.751152073732719e-05,
|
| 281 |
+
"loss": 1.0434,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.08998615597600369,
|
| 286 |
+
"grad_norm": 2.482023036660176,
|
| 287 |
+
"learning_rate": 1.7972350230414748e-05,
|
| 288 |
+
"loss": 1.0638,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.09229349330872173,
|
| 293 |
+
"grad_norm": 2.4167958259613944,
|
| 294 |
+
"learning_rate": 1.8433179723502307e-05,
|
| 295 |
+
"loss": 1.0422,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.09460083064143977,
|
| 300 |
+
"grad_norm": 2.5336845010658586,
|
| 301 |
+
"learning_rate": 1.8894009216589862e-05,
|
| 302 |
+
"loss": 1.0711,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.09690816797415783,
|
| 307 |
+
"grad_norm": 2.51621301326488,
|
| 308 |
+
"learning_rate": 1.935483870967742e-05,
|
| 309 |
+
"loss": 1.0881,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.09921550530687587,
|
| 314 |
+
"grad_norm": 2.4344384988965175,
|
| 315 |
+
"learning_rate": 1.981566820276498e-05,
|
| 316 |
+
"loss": 1.0735,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.10152284263959391,
|
| 321 |
+
"grad_norm": 2.5052346642024728,
|
| 322 |
+
"learning_rate": 1.9999883200175286e-05,
|
| 323 |
+
"loss": 1.0593,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.10383017997231195,
|
| 328 |
+
"grad_norm": 2.3275673568454986,
|
| 329 |
+
"learning_rate": 1.9999169433349454e-05,
|
| 330 |
+
"loss": 1.0766,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.10613751730503,
|
| 335 |
+
"grad_norm": 2.444969349213072,
|
| 336 |
+
"learning_rate": 1.9997806834748455e-05,
|
| 337 |
+
"loss": 1.0805,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.10844485463774804,
|
| 342 |
+
"grad_norm": 2.5019688689621455,
|
| 343 |
+
"learning_rate": 1.9995795492789368e-05,
|
| 344 |
+
"loss": 1.0795,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.11075219197046608,
|
| 349 |
+
"grad_norm": 2.41754553967893,
|
| 350 |
+
"learning_rate": 1.9993135537985285e-05,
|
| 351 |
+
"loss": 1.0419,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.11305952930318412,
|
| 356 |
+
"grad_norm": 2.4632725320939297,
|
| 357 |
+
"learning_rate": 1.9989827142936864e-05,
|
| 358 |
+
"loss": 1.1022,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.11536686663590216,
|
| 363 |
+
"grad_norm": 2.1283515103999004,
|
| 364 |
+
"learning_rate": 1.9985870522321118e-05,
|
| 365 |
+
"loss": 1.0727,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.11767420396862022,
|
| 370 |
+
"grad_norm": 2.5373992715116316,
|
| 371 |
+
"learning_rate": 1.9981265932877486e-05,
|
| 372 |
+
"loss": 1.0595,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.11998154130133826,
|
| 377 |
+
"grad_norm": 2.382000156883209,
|
| 378 |
+
"learning_rate": 1.9976013673391185e-05,
|
| 379 |
+
"loss": 1.0585,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.1222888786340563,
|
| 384 |
+
"grad_norm": 2.4022206882693857,
|
| 385 |
+
"learning_rate": 1.9970114084673796e-05,
|
| 386 |
+
"loss": 1.089,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.12459621596677434,
|
| 391 |
+
"grad_norm": 2.4577010071803707,
|
| 392 |
+
"learning_rate": 1.996356754954119e-05,
|
| 393 |
+
"loss": 1.0971,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.12690355329949238,
|
| 398 |
+
"grad_norm": 2.973750019978815,
|
| 399 |
+
"learning_rate": 1.995637449278864e-05,
|
| 400 |
+
"loss": 1.083,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.12921089063221042,
|
| 405 |
+
"grad_norm": 2.771443155668778,
|
| 406 |
+
"learning_rate": 1.994853538116329e-05,
|
| 407 |
+
"loss": 1.0948,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.13151822796492849,
|
| 412 |
+
"grad_norm": 2.2945686154224902,
|
| 413 |
+
"learning_rate": 1.9940050723333867e-05,
|
| 414 |
+
"loss": 1.0684,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.13382556529764653,
|
| 419 |
+
"grad_norm": 2.2402037197255864,
|
| 420 |
+
"learning_rate": 1.9930921069857653e-05,
|
| 421 |
+
"loss": 1.0605,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.13613290263036457,
|
| 426 |
+
"grad_norm": 3.0321973969800955,
|
| 427 |
+
"learning_rate": 1.9921147013144782e-05,
|
| 428 |
+
"loss": 1.0629,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.1384402399630826,
|
| 433 |
+
"grad_norm": 2.4295781427476215,
|
| 434 |
+
"learning_rate": 1.991072918741978e-05,
|
| 435 |
+
"loss": 1.0353,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.14074757729580065,
|
| 440 |
+
"grad_norm": 2.5121504192318698,
|
| 441 |
+
"learning_rate": 1.9899668268680438e-05,
|
| 442 |
+
"loss": 1.1156,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.1430549146285187,
|
| 447 |
+
"grad_norm": 2.454863605026453,
|
| 448 |
+
"learning_rate": 1.988796497465392e-05,
|
| 449 |
+
"loss": 1.0921,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.14536225196123673,
|
| 454 |
+
"grad_norm": 2.2532325296884532,
|
| 455 |
+
"learning_rate": 1.98756200647502e-05,
|
| 456 |
+
"loss": 1.0683,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.14766958929395477,
|
| 461 |
+
"grad_norm": 2.3557373973476334,
|
| 462 |
+
"learning_rate": 1.9862634340012796e-05,
|
| 463 |
+
"loss": 1.0559,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.1499769266266728,
|
| 468 |
+
"grad_norm": 2.493215513816939,
|
| 469 |
+
"learning_rate": 1.9849008643066774e-05,
|
| 470 |
+
"loss": 1.0725,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.15228426395939088,
|
| 475 |
+
"grad_norm": 3.2539520651857594,
|
| 476 |
+
"learning_rate": 1.983474385806408e-05,
|
| 477 |
+
"loss": 1.0674,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.15459160129210892,
|
| 482 |
+
"grad_norm": 2.2486130234020165,
|
| 483 |
+
"learning_rate": 1.9819840910626174e-05,
|
| 484 |
+
"loss": 1.0705,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.15689893862482696,
|
| 489 |
+
"grad_norm": 2.237162636902425,
|
| 490 |
+
"learning_rate": 1.9804300767783958e-05,
|
| 491 |
+
"loss": 1.0772,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.159206275957545,
|
| 496 |
+
"grad_norm": 6.51182752032121,
|
| 497 |
+
"learning_rate": 1.9788124437915034e-05,
|
| 498 |
+
"loss": 1.0837,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.16151361329026304,
|
| 503 |
+
"grad_norm": 2.322519163352806,
|
| 504 |
+
"learning_rate": 1.9771312970678258e-05,
|
| 505 |
+
"loss": 1.0405,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.16382095062298108,
|
| 510 |
+
"grad_norm": 2.2704370082078773,
|
| 511 |
+
"learning_rate": 1.9753867456945653e-05,
|
| 512 |
+
"loss": 1.0632,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.16612828795569912,
|
| 517 |
+
"grad_norm": 2.1581678351113602,
|
| 518 |
+
"learning_rate": 1.9735789028731603e-05,
|
| 519 |
+
"loss": 1.0818,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.16843562528841716,
|
| 524 |
+
"grad_norm": 2.0342475926569583,
|
| 525 |
+
"learning_rate": 1.971707885911941e-05,
|
| 526 |
+
"loss": 1.0679,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.1707429626211352,
|
| 531 |
+
"grad_norm": 2.2080262528668957,
|
| 532 |
+
"learning_rate": 1.9697738162185163e-05,
|
| 533 |
+
"loss": 1.0813,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.17305029995385326,
|
| 538 |
+
"grad_norm": 2.627310254608405,
|
| 539 |
+
"learning_rate": 1.9677768192918973e-05,
|
| 540 |
+
"loss": 1.0682,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.1753576372865713,
|
| 545 |
+
"grad_norm": 31.427071093107436,
|
| 546 |
+
"learning_rate": 1.9657170247143526e-05,
|
| 547 |
+
"loss": 1.06,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.17766497461928935,
|
| 552 |
+
"grad_norm": 2.4322378895965016,
|
| 553 |
+
"learning_rate": 1.9635945661430006e-05,
|
| 554 |
+
"loss": 1.0648,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.17997231195200739,
|
| 559 |
+
"grad_norm": 2.446435285466414,
|
| 560 |
+
"learning_rate": 1.9614095813011366e-05,
|
| 561 |
+
"loss": 1.0795,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.18227964928472543,
|
| 566 |
+
"grad_norm": 2.6478307008342745,
|
| 567 |
+
"learning_rate": 1.9591622119692953e-05,
|
| 568 |
+
"loss": 1.0655,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.18458698661744347,
|
| 573 |
+
"grad_norm": 3.5462973637007433,
|
| 574 |
+
"learning_rate": 1.956852603976052e-05,
|
| 575 |
+
"loss": 1.074,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.1868943239501615,
|
| 580 |
+
"grad_norm": 2.1952483345123834,
|
| 581 |
+
"learning_rate": 1.9544809071885603e-05,
|
| 582 |
+
"loss": 1.066,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.18920166128287955,
|
| 587 |
+
"grad_norm": 2.2118543537817397,
|
| 588 |
+
"learning_rate": 1.9520472755028256e-05,
|
| 589 |
+
"loss": 1.0567,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.1915089986155976,
|
| 594 |
+
"grad_norm": 2.1220225112366085,
|
| 595 |
+
"learning_rate": 1.9495518668337204e-05,
|
| 596 |
+
"loss": 1.0485,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.19381633594831565,
|
| 601 |
+
"grad_norm": 2.0821093984759433,
|
| 602 |
+
"learning_rate": 1.946994843104737e-05,
|
| 603 |
+
"loss": 1.0374,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.1961236732810337,
|
| 608 |
+
"grad_norm": 2.0190113966629433,
|
| 609 |
+
"learning_rate": 1.944376370237481e-05,
|
| 610 |
+
"loss": 1.064,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.19843101061375173,
|
| 615 |
+
"grad_norm": 2.056684419876963,
|
| 616 |
+
"learning_rate": 1.9416966181409047e-05,
|
| 617 |
+
"loss": 1.0524,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.20073834794646978,
|
| 622 |
+
"grad_norm": 2.366243541981115,
|
| 623 |
+
"learning_rate": 1.9389557607002808e-05,
|
| 624 |
+
"loss": 1.0587,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.20304568527918782,
|
| 629 |
+
"grad_norm": 2.3366923572205356,
|
| 630 |
+
"learning_rate": 1.9361539757659212e-05,
|
| 631 |
+
"loss": 1.104,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.20535302261190586,
|
| 636 |
+
"grad_norm": 2.135135120153321,
|
| 637 |
+
"learning_rate": 1.933291445141635e-05,
|
| 638 |
+
"loss": 1.0838,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.2076603599446239,
|
| 643 |
+
"grad_norm": 2.1686749928093367,
|
| 644 |
+
"learning_rate": 1.930368354572932e-05,
|
| 645 |
+
"loss": 1.0889,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.20996769727734194,
|
| 650 |
+
"grad_norm": 2.18295591084296,
|
| 651 |
+
"learning_rate": 1.9273848937349712e-05,
|
| 652 |
+
"loss": 1.0517,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.21227503461006,
|
| 657 |
+
"grad_norm": 2.015301818923552,
|
| 658 |
+
"learning_rate": 1.92434125622025e-05,
|
| 659 |
+
"loss": 1.0766,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.21458237194277804,
|
| 664 |
+
"grad_norm": 2.0741949328544984,
|
| 665 |
+
"learning_rate": 1.9212376395260447e-05,
|
| 666 |
+
"loss": 1.0896,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.21688970927549608,
|
| 671 |
+
"grad_norm": 2.15451982178122,
|
| 672 |
+
"learning_rate": 1.9180742450415962e-05,
|
| 673 |
+
"loss": 1.0763,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.21919704660821412,
|
| 678 |
+
"grad_norm": 2.0906260465476967,
|
| 679 |
+
"learning_rate": 1.9148512780350384e-05,
|
| 680 |
+
"loss": 1.086,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.22150438394093216,
|
| 685 |
+
"grad_norm": 2.031559261836197,
|
| 686 |
+
"learning_rate": 1.9115689476400817e-05,
|
| 687 |
+
"loss": 1.059,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.2238117212736502,
|
| 692 |
+
"grad_norm": 2.07650174311531,
|
| 693 |
+
"learning_rate": 1.9082274668424423e-05,
|
| 694 |
+
"loss": 1.0679,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.22611905860636825,
|
| 699 |
+
"grad_norm": 2.126208239890011,
|
| 700 |
+
"learning_rate": 1.9048270524660197e-05,
|
| 701 |
+
"loss": 1.0809,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.22842639593908629,
|
| 706 |
+
"grad_norm": 1.9929349716624978,
|
| 707 |
+
"learning_rate": 1.9013679251588304e-05,
|
| 708 |
+
"loss": 1.085,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.23073373327180433,
|
| 713 |
+
"grad_norm": 3.001431077273745,
|
| 714 |
+
"learning_rate": 1.8978503093786882e-05,
|
| 715 |
+
"loss": 1.0558,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.2330410706045224,
|
| 720 |
+
"grad_norm": 1.9284000626000521,
|
| 721 |
+
"learning_rate": 1.89427443337864e-05,
|
| 722 |
+
"loss": 1.0685,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.23534840793724043,
|
| 727 |
+
"grad_norm": 2.156814659249471,
|
| 728 |
+
"learning_rate": 1.890640529192155e-05,
|
| 729 |
+
"loss": 1.0857,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.23765574526995847,
|
| 734 |
+
"grad_norm": 2.2063349330204174,
|
| 735 |
+
"learning_rate": 1.8869488326180682e-05,
|
| 736 |
+
"loss": 1.092,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.23996308260267651,
|
| 741 |
+
"grad_norm": 1.8963715836357997,
|
| 742 |
+
"learning_rate": 1.8831995832052802e-05,
|
| 743 |
+
"loss": 1.0694,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.24227041993539455,
|
| 748 |
+
"grad_norm": 2.0285632136378613,
|
| 749 |
+
"learning_rate": 1.8793930242372117e-05,
|
| 750 |
+
"loss": 1.0795,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.2445777572681126,
|
| 755 |
+
"grad_norm": 2.099474069037447,
|
| 756 |
+
"learning_rate": 1.8755294027160203e-05,
|
| 757 |
+
"loss": 1.0893,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.24688509460083063,
|
| 762 |
+
"grad_norm": 2.0358502445768165,
|
| 763 |
+
"learning_rate": 1.8716089693465696e-05,
|
| 764 |
+
"loss": 1.086,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.24919243193354867,
|
| 769 |
+
"grad_norm": 2.1218454361521633,
|
| 770 |
+
"learning_rate": 1.8676319785201617e-05,
|
| 771 |
+
"loss": 1.0842,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.2514997692662667,
|
| 776 |
+
"grad_norm": 2.0341225955626583,
|
| 777 |
+
"learning_rate": 1.8635986882980325e-05,
|
| 778 |
+
"loss": 1.0625,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.25380710659898476,
|
| 783 |
+
"grad_norm": 2.3910625184538747,
|
| 784 |
+
"learning_rate": 1.8595093603946053e-05,
|
| 785 |
+
"loss": 1.0727,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.2561144439317028,
|
| 790 |
+
"grad_norm": 1.98644765469211,
|
| 791 |
+
"learning_rate": 1.855364260160507e-05,
|
| 792 |
+
"loss": 1.0595,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.25842178126442084,
|
| 797 |
+
"grad_norm": 2.208738179396901,
|
| 798 |
+
"learning_rate": 1.851163656565351e-05,
|
| 799 |
+
"loss": 1.0936,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.2607291185971389,
|
| 804 |
+
"grad_norm": 2.0209364206754645,
|
| 805 |
+
"learning_rate": 1.846907822180286e-05,
|
| 806 |
+
"loss": 1.0684,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.26303645592985697,
|
| 811 |
+
"grad_norm": 1.925369099665116,
|
| 812 |
+
"learning_rate": 1.842597033160306e-05,
|
| 813 |
+
"loss": 1.0669,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.265343793262575,
|
| 818 |
+
"grad_norm": 2.169504176441067,
|
| 819 |
+
"learning_rate": 1.8382315692263324e-05,
|
| 820 |
+
"loss": 1.0914,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.26765113059529305,
|
| 825 |
+
"grad_norm": 2.0095996014073503,
|
| 826 |
+
"learning_rate": 1.8338117136470648e-05,
|
| 827 |
+
"loss": 1.0679,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.2699584679280111,
|
| 832 |
+
"grad_norm": 2.0780448467433468,
|
| 833 |
+
"learning_rate": 1.829337753220597e-05,
|
| 834 |
+
"loss": 1.0823,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.27226580526072913,
|
| 839 |
+
"grad_norm": 1.9092130149771946,
|
| 840 |
+
"learning_rate": 1.8248099782558103e-05,
|
| 841 |
+
"loss": 1.0485,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.2745731425934472,
|
| 846 |
+
"grad_norm": 2.2904699258286914,
|
| 847 |
+
"learning_rate": 1.820228682553533e-05,
|
| 848 |
+
"loss": 1.0676,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.2768804799261652,
|
| 853 |
+
"grad_norm": 2.1592942059891884,
|
| 854 |
+
"learning_rate": 1.8155941633874787e-05,
|
| 855 |
+
"loss": 1.0862,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.27918781725888325,
|
| 860 |
+
"grad_norm": 1.9056960337173154,
|
| 861 |
+
"learning_rate": 1.810906721484954e-05,
|
| 862 |
+
"loss": 1.027,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.2814951545916013,
|
| 867 |
+
"grad_norm": 2.029717811241469,
|
| 868 |
+
"learning_rate": 1.8061666610073465e-05,
|
| 869 |
+
"loss": 1.0638,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.28380249192431933,
|
| 874 |
+
"grad_norm": 2.0411421295106873,
|
| 875 |
+
"learning_rate": 1.8013742895303883e-05,
|
| 876 |
+
"loss": 1.0667,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.2861098292570374,
|
| 881 |
+
"grad_norm": 1.924799713813513,
|
| 882 |
+
"learning_rate": 1.7965299180241963e-05,
|
| 883 |
+
"loss": 1.0685,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.2884171665897554,
|
| 888 |
+
"grad_norm": 2.171875799314523,
|
| 889 |
+
"learning_rate": 1.791633860833096e-05,
|
| 890 |
+
"loss": 1.0463,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.29072450392247345,
|
| 895 |
+
"grad_norm": 2.068372996208825,
|
| 896 |
+
"learning_rate": 1.7866864356552215e-05,
|
| 897 |
+
"loss": 1.0715,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.2930318412551915,
|
| 902 |
+
"grad_norm": 1.8601211490681129,
|
| 903 |
+
"learning_rate": 1.7816879635219028e-05,
|
| 904 |
+
"loss": 1.0576,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.29533917858790953,
|
| 909 |
+
"grad_norm": 1.9725316785259686,
|
| 910 |
+
"learning_rate": 1.7766387687768338e-05,
|
| 911 |
+
"loss": 1.0648,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.2976465159206276,
|
| 916 |
+
"grad_norm": 2.1844471268704515,
|
| 917 |
+
"learning_rate": 1.7715391790550255e-05,
|
| 918 |
+
"loss": 1.0637,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.2999538532533456,
|
| 923 |
+
"grad_norm": 1.933021204525043,
|
| 924 |
+
"learning_rate": 1.766389525261547e-05,
|
| 925 |
+
"loss": 1.0803,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.30226119058606365,
|
| 930 |
+
"grad_norm": 2.1351960039602695,
|
| 931 |
+
"learning_rate": 1.7611901415500536e-05,
|
| 932 |
+
"loss": 1.0979,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.30456852791878175,
|
| 937 |
+
"grad_norm": 1.9611319269471612,
|
| 938 |
+
"learning_rate": 1.7559413653011027e-05,
|
| 939 |
+
"loss": 1.0652,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.3068758652514998,
|
| 944 |
+
"grad_norm": 2.033679368734863,
|
| 945 |
+
"learning_rate": 1.7506435371002635e-05,
|
| 946 |
+
"loss": 1.0749,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.30918320258421783,
|
| 951 |
+
"grad_norm": 2.0279720872015354,
|
| 952 |
+
"learning_rate": 1.745297000716016e-05,
|
| 953 |
+
"loss": 1.078,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.31149053991693587,
|
| 958 |
+
"grad_norm": 1.989733876253561,
|
| 959 |
+
"learning_rate": 1.7399021030774443e-05,
|
| 960 |
+
"loss": 1.0639,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.3137978772496539,
|
| 965 |
+
"grad_norm": 1.9037569950190747,
|
| 966 |
+
"learning_rate": 1.734459194251725e-05,
|
| 967 |
+
"loss": 1.0721,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.31610521458237195,
|
| 972 |
+
"grad_norm": 2.183774346551292,
|
| 973 |
+
"learning_rate": 1.7289686274214116e-05,
|
| 974 |
+
"loss": 1.0755,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.31841255191509,
|
| 979 |
+
"grad_norm": 1.992668508208317,
|
| 980 |
+
"learning_rate": 1.7234307588615177e-05,
|
| 981 |
+
"loss": 1.0761,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.32071988924780803,
|
| 986 |
+
"grad_norm": 1.9985850630928745,
|
| 987 |
+
"learning_rate": 1.717845947916398e-05,
|
| 988 |
+
"loss": 1.0575,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.3230272265805261,
|
| 993 |
+
"grad_norm": 2.1480400724448883,
|
| 994 |
+
"learning_rate": 1.712214556976431e-05,
|
| 995 |
+
"loss": 1.0404,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.3253345639132441,
|
| 1000 |
+
"grad_norm": 1.9503339704430334,
|
| 1001 |
+
"learning_rate": 1.7065369514545054e-05,
|
| 1002 |
+
"loss": 1.0579,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.32764190124596215,
|
| 1007 |
+
"grad_norm": 1.971699140050545,
|
| 1008 |
+
"learning_rate": 1.7008134997623066e-05,
|
| 1009 |
+
"loss": 1.0629,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.3299492385786802,
|
| 1014 |
+
"grad_norm": 1.97358622805482,
|
| 1015 |
+
"learning_rate": 1.695044573286413e-05,
|
| 1016 |
+
"loss": 1.039,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.33225657591139823,
|
| 1021 |
+
"grad_norm": 1.8903289514072814,
|
| 1022 |
+
"learning_rate": 1.6892305463641967e-05,
|
| 1023 |
+
"loss": 1.0996,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.3345639132441163,
|
| 1028 |
+
"grad_norm": 1.9171530806208752,
|
| 1029 |
+
"learning_rate": 1.6833717962595327e-05,
|
| 1030 |
+
"loss": 1.0587,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.3368712505768343,
|
| 1035 |
+
"grad_norm": 2.0292768253738855,
|
| 1036 |
+
"learning_rate": 1.677468703138319e-05,
|
| 1037 |
+
"loss": 1.0534,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.33917858790955235,
|
| 1042 |
+
"grad_norm": 1.9857495035997068,
|
| 1043 |
+
"learning_rate": 1.6715216500438093e-05,
|
| 1044 |
+
"loss": 1.0805,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.3414859252422704,
|
| 1049 |
+
"grad_norm": 5.464414796561983,
|
| 1050 |
+
"learning_rate": 1.6655310228717565e-05,
|
| 1051 |
+
"loss": 1.0802,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.3437932625749885,
|
| 1056 |
+
"grad_norm": 1.909425909480839,
|
| 1057 |
+
"learning_rate": 1.6594972103453727e-05,
|
| 1058 |
+
"loss": 1.0813,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.34610059990770653,
|
| 1063 |
+
"grad_norm": 1.9164783421961078,
|
| 1064 |
+
"learning_rate": 1.6534206039901057e-05,
|
| 1065 |
+
"loss": 1.0466,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.34840793724042457,
|
| 1070 |
+
"grad_norm": 4.300395520109931,
|
| 1071 |
+
"learning_rate": 1.647301598108234e-05,
|
| 1072 |
+
"loss": 1.0326,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.3507152745731426,
|
| 1077 |
+
"grad_norm": 2.2226539635666827,
|
| 1078 |
+
"learning_rate": 1.64114058975328e-05,
|
| 1079 |
+
"loss": 1.0824,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.35302261190586065,
|
| 1084 |
+
"grad_norm": 2.08738140836867,
|
| 1085 |
+
"learning_rate": 1.6349379787042478e-05,
|
| 1086 |
+
"loss": 1.0445,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.3553299492385787,
|
| 1091 |
+
"grad_norm": 1.7956395961308758,
|
| 1092 |
+
"learning_rate": 1.6286941674396788e-05,
|
| 1093 |
+
"loss": 1.0283,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.35763728657129673,
|
| 1098 |
+
"grad_norm": 1.9001406773036147,
|
| 1099 |
+
"learning_rate": 1.6224095611115385e-05,
|
| 1100 |
+
"loss": 1.0558,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.35994462390401477,
|
| 1105 |
+
"grad_norm": 1.8584506601925908,
|
| 1106 |
+
"learning_rate": 1.6160845675189254e-05,
|
| 1107 |
+
"loss": 1.0315,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.3622519612367328,
|
| 1112 |
+
"grad_norm": 1.8994200106765273,
|
| 1113 |
+
"learning_rate": 1.6097195970816094e-05,
|
| 1114 |
+
"loss": 1.0736,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.36455929856945085,
|
| 1119 |
+
"grad_norm": 2.396252821544053,
|
| 1120 |
+
"learning_rate": 1.603315062813401e-05,
|
| 1121 |
+
"loss": 1.0605,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.3668666359021689,
|
| 1126 |
+
"grad_norm": 1.798952489279231,
|
| 1127 |
+
"learning_rate": 1.596871380295351e-05,
|
| 1128 |
+
"loss": 1.0439,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.36917397323488693,
|
| 1133 |
+
"grad_norm": 1.8907451459454219,
|
| 1134 |
+
"learning_rate": 1.5903889676487832e-05,
|
| 1135 |
+
"loss": 1.047,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.37148131056760497,
|
| 1140 |
+
"grad_norm": 1.9272449578154556,
|
| 1141 |
+
"learning_rate": 1.5838682455081657e-05,
|
| 1142 |
+
"loss": 1.0557,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.373788647900323,
|
| 1147 |
+
"grad_norm": 1.8500068465129675,
|
| 1148 |
+
"learning_rate": 1.5773096369938125e-05,
|
| 1149 |
+
"loss": 1.0448,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.37609598523304105,
|
| 1154 |
+
"grad_norm": 2.024989082401722,
|
| 1155 |
+
"learning_rate": 1.570713567684432e-05,
|
| 1156 |
+
"loss": 1.0444,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.3784033225657591,
|
| 1161 |
+
"grad_norm": 1.864545609223796,
|
| 1162 |
+
"learning_rate": 1.5640804655895086e-05,
|
| 1163 |
+
"loss": 1.0316,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.38071065989847713,
|
| 1168 |
+
"grad_norm": 1.9810163039010853,
|
| 1169 |
+
"learning_rate": 1.557410761121532e-05,
|
| 1170 |
+
"loss": 1.0476,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.3830179972311952,
|
| 1175 |
+
"grad_norm": 1.9715732068507474,
|
| 1176 |
+
"learning_rate": 1.5507048870680668e-05,
|
| 1177 |
+
"loss": 1.0092,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.38532533456391327,
|
| 1182 |
+
"grad_norm": 1.9346233566232378,
|
| 1183 |
+
"learning_rate": 1.5439632785636707e-05,
|
| 1184 |
+
"loss": 1.0834,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.3876326718966313,
|
| 1189 |
+
"grad_norm": 2.2247085368619164,
|
| 1190 |
+
"learning_rate": 1.5371863730616586e-05,
|
| 1191 |
+
"loss": 1.0608,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.38994000922934935,
|
| 1196 |
+
"grad_norm": 1.855445999462738,
|
| 1197 |
+
"learning_rate": 1.5303746103057163e-05,
|
| 1198 |
+
"loss": 1.0311,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.3922473465620674,
|
| 1203 |
+
"grad_norm": 1.8433500481185805,
|
| 1204 |
+
"learning_rate": 1.5235284323013674e-05,
|
| 1205 |
+
"loss": 1.0513,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.39455468389478543,
|
| 1210 |
+
"grad_norm": 1.9238020550812749,
|
| 1211 |
+
"learning_rate": 1.5166482832872923e-05,
|
| 1212 |
+
"loss": 1.0611,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.39686202122750347,
|
| 1217 |
+
"grad_norm": 2.034539491931288,
|
| 1218 |
+
"learning_rate": 1.5097346097065008e-05,
|
| 1219 |
+
"loss": 1.0369,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.3991693585602215,
|
| 1224 |
+
"grad_norm": 1.8719773240320596,
|
| 1225 |
+
"learning_rate": 1.5027878601773633e-05,
|
| 1226 |
+
"loss": 1.031,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.40147669589293955,
|
| 1231 |
+
"grad_norm": 1.8218549093076317,
|
| 1232 |
+
"learning_rate": 1.4958084854645018e-05,
|
| 1233 |
+
"loss": 1.027,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.4037840332256576,
|
| 1238 |
+
"grad_norm": 1.908377236915306,
|
| 1239 |
+
"learning_rate": 1.4887969384495403e-05,
|
| 1240 |
+
"loss": 1.0505,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.40609137055837563,
|
| 1245 |
+
"grad_norm": 1.882874108335332,
|
| 1246 |
+
"learning_rate": 1.4817536741017153e-05,
|
| 1247 |
+
"loss": 1.0421,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.40839870789109367,
|
| 1252 |
+
"grad_norm": 2.0217065440618622,
|
| 1253 |
+
"learning_rate": 1.4746791494483584e-05,
|
| 1254 |
+
"loss": 1.0533,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.4107060452238117,
|
| 1259 |
+
"grad_norm": 1.8717323475177303,
|
| 1260 |
+
"learning_rate": 1.4675738235452352e-05,
|
| 1261 |
+
"loss": 1.0279,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.41301338255652975,
|
| 1266 |
+
"grad_norm": 1.9788825364185045,
|
| 1267 |
+
"learning_rate": 1.4604381574467616e-05,
|
| 1268 |
+
"loss": 1.042,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.4153207198892478,
|
| 1273 |
+
"grad_norm": 1.9327030504589935,
|
| 1274 |
+
"learning_rate": 1.4532726141760849e-05,
|
| 1275 |
+
"loss": 1.06,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.41762805722196583,
|
| 1280 |
+
"grad_norm": 1.8050202525178007,
|
| 1281 |
+
"learning_rate": 1.4460776586950393e-05,
|
| 1282 |
+
"loss": 1.0176,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.41993539455468387,
|
| 1287 |
+
"grad_norm": 1.7140772518888605,
|
| 1288 |
+
"learning_rate": 1.438853757873975e-05,
|
| 1289 |
+
"loss": 1.0336,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.4222427318874019,
|
| 1294 |
+
"grad_norm": 1.9381284110778458,
|
| 1295 |
+
"learning_rate": 1.4316013804614644e-05,
|
| 1296 |
+
"loss": 1.0283,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.42455006922012,
|
| 1301 |
+
"grad_norm": 1.8569863683755345,
|
| 1302 |
+
"learning_rate": 1.4243209970538846e-05,
|
| 1303 |
+
"loss": 1.0295,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.42685740655283805,
|
| 1308 |
+
"grad_norm": 1.7584125894267681,
|
| 1309 |
+
"learning_rate": 1.4170130800648814e-05,
|
| 1310 |
+
"loss": 1.0451,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.4291647438855561,
|
| 1315 |
+
"grad_norm": 1.8468563006595364,
|
| 1316 |
+
"learning_rate": 1.4096781036947159e-05,
|
| 1317 |
+
"loss": 1.0329,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.43147208121827413,
|
| 1322 |
+
"grad_norm": 1.8005410726866136,
|
| 1323 |
+
"learning_rate": 1.4023165438994933e-05,
|
| 1324 |
+
"loss": 1.0523,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.43377941855099217,
|
| 1329 |
+
"grad_norm": 1.7881203203680747,
|
| 1330 |
+
"learning_rate": 1.394928878360279e-05,
|
| 1331 |
+
"loss": 1.052,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.4360867558837102,
|
| 1336 |
+
"grad_norm": 1.9402582404112974,
|
| 1337 |
+
"learning_rate": 1.3875155864521031e-05,
|
| 1338 |
+
"loss": 1.0418,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.43839409321642825,
|
| 1343 |
+
"grad_norm": 1.926002050119894,
|
| 1344 |
+
"learning_rate": 1.3800771492128537e-05,
|
| 1345 |
+
"loss": 1.0491,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.4407014305491463,
|
| 1350 |
+
"grad_norm": 1.8807563751664647,
|
| 1351 |
+
"learning_rate": 1.3726140493120639e-05,
|
| 1352 |
+
"loss": 1.032,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.44300876788186433,
|
| 1357 |
+
"grad_norm": 1.9189485256851713,
|
| 1358 |
+
"learning_rate": 1.3651267710195909e-05,
|
| 1359 |
+
"loss": 1.0355,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.44531610521458237,
|
| 1364 |
+
"grad_norm": 1.9803461150155048,
|
| 1365 |
+
"learning_rate": 1.3576158001741932e-05,
|
| 1366 |
+
"loss": 1.0569,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.4476234425473004,
|
| 1371 |
+
"grad_norm": 2.018305042251882,
|
| 1372 |
+
"learning_rate": 1.3500816241520059e-05,
|
| 1373 |
+
"loss": 1.04,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.44993077988001845,
|
| 1378 |
+
"grad_norm": 1.7580864552202506,
|
| 1379 |
+
"learning_rate": 1.3425247318349137e-05,
|
| 1380 |
+
"loss": 1.0075,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.4522381172127365,
|
| 1385 |
+
"grad_norm": 1.8967387580024155,
|
| 1386 |
+
"learning_rate": 1.3349456135788298e-05,
|
| 1387 |
+
"loss": 1.0429,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.45454545454545453,
|
| 1392 |
+
"grad_norm": 2.2301409193225985,
|
| 1393 |
+
"learning_rate": 1.3273447611818768e-05,
|
| 1394 |
+
"loss": 1.0244,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.45685279187817257,
|
| 1399 |
+
"grad_norm": 1.8218900361937265,
|
| 1400 |
+
"learning_rate": 1.3197226678524739e-05,
|
| 1401 |
+
"loss": 1.0006,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.4591601292108906,
|
| 1406 |
+
"grad_norm": 1.8705871989801575,
|
| 1407 |
+
"learning_rate": 1.3120798281773346e-05,
|
| 1408 |
+
"loss": 1.0382,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.46146746654360865,
|
| 1413 |
+
"grad_norm": 1.8993921065361903,
|
| 1414 |
+
"learning_rate": 1.3044167380893726e-05,
|
| 1415 |
+
"loss": 1.0543,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.46377480387632675,
|
| 1420 |
+
"grad_norm": 1.762141098208751,
|
| 1421 |
+
"learning_rate": 1.2967338948355217e-05,
|
| 1422 |
+
"loss": 1.031,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.4660821412090448,
|
| 1427 |
+
"grad_norm": 1.8349620217027005,
|
| 1428 |
+
"learning_rate": 1.2890317969444716e-05,
|
| 1429 |
+
"loss": 1.0104,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.4683894785417628,
|
| 1434 |
+
"grad_norm": 1.9527169882770812,
|
| 1435 |
+
"learning_rate": 1.2813109441943166e-05,
|
| 1436 |
+
"loss": 1.0325,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.47069681587448087,
|
| 1441 |
+
"grad_norm": 1.7594153130967782,
|
| 1442 |
+
"learning_rate": 1.273571837580127e-05,
|
| 1443 |
+
"loss": 1.0476,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.4730041532071989,
|
| 1448 |
+
"grad_norm": 1.7823802580407797,
|
| 1449 |
+
"learning_rate": 1.2658149792814405e-05,
|
| 1450 |
+
"loss": 1.0397,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.47531149053991695,
|
| 1455 |
+
"grad_norm": 1.7288773807653248,
|
| 1456 |
+
"learning_rate": 1.258040872629676e-05,
|
| 1457 |
+
"loss": 1.0419,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.477618827872635,
|
| 1462 |
+
"grad_norm": 1.8969879276116197,
|
| 1463 |
+
"learning_rate": 1.2502500220754736e-05,
|
| 1464 |
+
"loss": 1.0538,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.47992616520535303,
|
| 1469 |
+
"grad_norm": 1.9748280209096565,
|
| 1470 |
+
"learning_rate": 1.242442933155961e-05,
|
| 1471 |
+
"loss": 1.0088,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.48223350253807107,
|
| 1476 |
+
"grad_norm": 1.8630834023430007,
|
| 1477 |
+
"learning_rate": 1.2346201124619502e-05,
|
| 1478 |
+
"loss": 1.0041,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.4845408398707891,
|
| 1483 |
+
"grad_norm": 1.93282001404706,
|
| 1484 |
+
"learning_rate": 1.2267820676050657e-05,
|
| 1485 |
+
"loss": 1.0117,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.48684817720350715,
|
| 1490 |
+
"grad_norm": 1.7732915883551568,
|
| 1491 |
+
"learning_rate": 1.2189293071848051e-05,
|
| 1492 |
+
"loss": 1.0395,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.4891555145362252,
|
| 1497 |
+
"grad_norm": 1.7668917225682153,
|
| 1498 |
+
"learning_rate": 1.2110623407555398e-05,
|
| 1499 |
+
"loss": 1.0055,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.49146285186894323,
|
| 1504 |
+
"grad_norm": 1.770548333794891,
|
| 1505 |
+
"learning_rate": 1.2031816787934465e-05,
|
| 1506 |
+
"loss": 1.0198,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.49377018920166127,
|
| 1511 |
+
"grad_norm": 1.804504292487286,
|
| 1512 |
+
"learning_rate": 1.1952878326633872e-05,
|
| 1513 |
+
"loss": 0.9925,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.4960775265343793,
|
| 1518 |
+
"grad_norm": 1.8384240011850799,
|
| 1519 |
+
"learning_rate": 1.187381314585725e-05,
|
| 1520 |
+
"loss": 1.0167,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.49838486386709735,
|
| 1525 |
+
"grad_norm": 1.8738489698458378,
|
| 1526 |
+
"learning_rate": 1.1794626376030866e-05,
|
| 1527 |
+
"loss": 1.0266,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.5006922011998154,
|
| 1532 |
+
"grad_norm": 1.7963716989600227,
|
| 1533 |
+
"learning_rate": 1.1715323155470745e-05,
|
| 1534 |
+
"loss": 1.0203,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.5029995385325334,
|
| 1539 |
+
"grad_norm": 1.7950924324700734,
|
| 1540 |
+
"learning_rate": 1.163590863004922e-05,
|
| 1541 |
+
"loss": 1.0014,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.5053068758652515,
|
| 1546 |
+
"grad_norm": 1.7996992785566162,
|
| 1547 |
+
"learning_rate": 1.1556387952861036e-05,
|
| 1548 |
+
"loss": 1.0147,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.5076142131979695,
|
| 1553 |
+
"grad_norm": 1.9262189643769105,
|
| 1554 |
+
"learning_rate": 1.1476766283888986e-05,
|
| 1555 |
+
"loss": 1.0176,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.5099215505306876,
|
| 1560 |
+
"grad_norm": 1.738673259571015,
|
| 1561 |
+
"learning_rate": 1.1397048789669061e-05,
|
| 1562 |
+
"loss": 1.0221,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.5122288878634056,
|
| 1567 |
+
"grad_norm": 1.7993896869653003,
|
| 1568 |
+
"learning_rate": 1.1317240642955226e-05,
|
| 1569 |
+
"loss": 1.0232,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.5145362251961236,
|
| 1574 |
+
"grad_norm": 1.8492729168966688,
|
| 1575 |
+
"learning_rate": 1.1237347022383747e-05,
|
| 1576 |
+
"loss": 1.0138,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.5168435625288417,
|
| 1581 |
+
"grad_norm": 1.792127722956897,
|
| 1582 |
+
"learning_rate": 1.1157373112137171e-05,
|
| 1583 |
+
"loss": 1.011,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.5191508998615597,
|
| 1588 |
+
"grad_norm": 1.767761412954839,
|
| 1589 |
+
"learning_rate": 1.107732410160793e-05,
|
| 1590 |
+
"loss": 0.9917,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.5214582371942778,
|
| 1595 |
+
"grad_norm": 1.78861016845621,
|
| 1596 |
+
"learning_rate": 1.0997205185061599e-05,
|
| 1597 |
+
"loss": 1.024,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.5237655745269958,
|
| 1602 |
+
"grad_norm": 1.796111964851059,
|
| 1603 |
+
"learning_rate": 1.0917021561299864e-05,
|
| 1604 |
+
"loss": 1.0094,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.5260729118597139,
|
| 1609 |
+
"grad_norm": 1.6839699201837544,
|
| 1610 |
+
"learning_rate": 1.083677843332316e-05,
|
| 1611 |
+
"loss": 1.0019,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.528380249192432,
|
| 1616 |
+
"grad_norm": 1.735381208836221,
|
| 1617 |
+
"learning_rate": 1.0756481007993063e-05,
|
| 1618 |
+
"loss": 0.9979,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.53068758652515,
|
| 1623 |
+
"grad_norm": 1.7283238773850635,
|
| 1624 |
+
"learning_rate": 1.0676134495694439e-05,
|
| 1625 |
+
"loss": 1.0127,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.5329949238578681,
|
| 1630 |
+
"grad_norm": 1.8700927823490678,
|
| 1631 |
+
"learning_rate": 1.0595744109997326e-05,
|
| 1632 |
+
"loss": 0.9897,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.5353022611905861,
|
| 1637 |
+
"grad_norm": 1.7850898978429104,
|
| 1638 |
+
"learning_rate": 1.0515315067318652e-05,
|
| 1639 |
+
"loss": 1.0155,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.5376095985233041,
|
| 1644 |
+
"grad_norm": 1.8867979217718087,
|
| 1645 |
+
"learning_rate": 1.0434852586583737e-05,
|
| 1646 |
+
"loss": 0.9996,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.5399169358560222,
|
| 1651 |
+
"grad_norm": 1.8409069763382047,
|
| 1652 |
+
"learning_rate": 1.0354361888887642e-05,
|
| 1653 |
+
"loss": 1.0038,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.5422242731887402,
|
| 1658 |
+
"grad_norm": 1.9318197730416369,
|
| 1659 |
+
"learning_rate": 1.0273848197156401e-05,
|
| 1660 |
+
"loss": 0.9893,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.5445316105214583,
|
| 1665 |
+
"grad_norm": 1.7534602196952722,
|
| 1666 |
+
"learning_rate": 1.0193316735808085e-05,
|
| 1667 |
+
"loss": 0.993,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.5468389478541763,
|
| 1672 |
+
"grad_norm": 1.7436059512387687,
|
| 1673 |
+
"learning_rate": 1.0112772730413816e-05,
|
| 1674 |
+
"loss": 1.0079,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.5491462851868943,
|
| 1679 |
+
"grad_norm": 1.8076118426423142,
|
| 1680 |
+
"learning_rate": 1.0032221407358683e-05,
|
| 1681 |
+
"loss": 1.0336,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.5514536225196124,
|
| 1686 |
+
"grad_norm": 1.9044420451434694,
|
| 1687 |
+
"learning_rate": 9.951667993502599e-06,
|
| 1688 |
+
"loss": 1.0152,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.5537609598523304,
|
| 1693 |
+
"grad_norm": 1.842329102136153,
|
| 1694 |
+
"learning_rate": 9.871117715841151e-06,
|
| 1695 |
+
"loss": 0.9783,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.5560682971850485,
|
| 1700 |
+
"grad_norm": 1.8583622986365993,
|
| 1701 |
+
"learning_rate": 9.790575801166432e-06,
|
| 1702 |
+
"loss": 1.0054,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.5583756345177665,
|
| 1707 |
+
"grad_norm": 1.7004919095912332,
|
| 1708 |
+
"learning_rate": 9.710047475727854e-06,
|
| 1709 |
+
"loss": 1.0011,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5606829718504845,
|
| 1714 |
+
"grad_norm": 1.7280460978374188,
|
| 1715 |
+
"learning_rate": 9.629537964893063e-06,
|
| 1716 |
+
"loss": 1.0299,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.5629903091832026,
|
| 1721 |
+
"grad_norm": 1.7174011953937558,
|
| 1722 |
+
"learning_rate": 9.549052492808834e-06,
|
| 1723 |
+
"loss": 0.9946,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.5652976465159206,
|
| 1728 |
+
"grad_norm": 1.8215498597720168,
|
| 1729 |
+
"learning_rate": 9.468596282062114e-06,
|
| 1730 |
+
"loss": 1.0113,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.5676049838486387,
|
| 1735 |
+
"grad_norm": 1.6911500192296895,
|
| 1736 |
+
"learning_rate": 9.38817455334112e-06,
|
| 1737 |
+
"loss": 0.9855,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.5699123211813567,
|
| 1742 |
+
"grad_norm": 1.8405046704539174,
|
| 1743 |
+
"learning_rate": 9.307792525096582e-06,
|
| 1744 |
+
"loss": 1.0113,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.5722196585140747,
|
| 1749 |
+
"grad_norm": 1.838204327540361,
|
| 1750 |
+
"learning_rate": 9.227455413203115e-06,
|
| 1751 |
+
"loss": 0.9947,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.5745269958467928,
|
| 1756 |
+
"grad_norm": 1.656688699609939,
|
| 1757 |
+
"learning_rate": 9.147168430620788e-06,
|
| 1758 |
+
"loss": 0.9892,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.5768343331795108,
|
| 1763 |
+
"grad_norm": 1.7231036061816765,
|
| 1764 |
+
"learning_rate": 9.066936787056843e-06,
|
| 1765 |
+
"loss": 0.9944,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.5791416705122289,
|
| 1770 |
+
"grad_norm": 1.645605940940624,
|
| 1771 |
+
"learning_rate": 8.986765688627652e-06,
|
| 1772 |
+
"loss": 0.9936,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.5814490078449469,
|
| 1777 |
+
"grad_norm": 1.8141527360329759,
|
| 1778 |
+
"learning_rate": 8.906660337520903e-06,
|
| 1779 |
+
"loss": 1.0096,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.583756345177665,
|
| 1784 |
+
"grad_norm": 1.856808726362016,
|
| 1785 |
+
"learning_rate": 8.82662593165804e-06,
|
| 1786 |
+
"loss": 1.0032,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.586063682510383,
|
| 1791 |
+
"grad_norm": 1.8593818092553211,
|
| 1792 |
+
"learning_rate": 8.746667664356957e-06,
|
| 1793 |
+
"loss": 1.0177,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.588371019843101,
|
| 1798 |
+
"grad_norm": 1.7318701186944272,
|
| 1799 |
+
"learning_rate": 8.666790723995043e-06,
|
| 1800 |
+
"loss": 0.9933,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.5906783571758191,
|
| 1805 |
+
"grad_norm": 1.8632249625406112,
|
| 1806 |
+
"learning_rate": 8.587000293672482e-06,
|
| 1807 |
+
"loss": 1.0278,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.5929856945085371,
|
| 1812 |
+
"grad_norm": 1.8482080793994375,
|
| 1813 |
+
"learning_rate": 8.50730155087596e-06,
|
| 1814 |
+
"loss": 0.9753,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.5952930318412551,
|
| 1819 |
+
"grad_norm": 1.6654816438940703,
|
| 1820 |
+
"learning_rate": 8.427699667142681e-06,
|
| 1821 |
+
"loss": 0.9923,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.5976003691739732,
|
| 1826 |
+
"grad_norm": 1.816789112483473,
|
| 1827 |
+
"learning_rate": 8.348199807724806e-06,
|
| 1828 |
+
"loss": 0.9951,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.5999077065066912,
|
| 1833 |
+
"grad_norm": 1.8671938825009406,
|
| 1834 |
+
"learning_rate": 8.268807131254288e-06,
|
| 1835 |
+
"loss": 1.0063,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.6022150438394093,
|
| 1840 |
+
"grad_norm": 1.736173419625791,
|
| 1841 |
+
"learning_rate": 8.189526789408123e-06,
|
| 1842 |
+
"loss": 0.9942,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.6045223811721273,
|
| 1847 |
+
"grad_norm": 1.7397594354717327,
|
| 1848 |
+
"learning_rate": 8.110363926574088e-06,
|
| 1849 |
+
"loss": 0.9899,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.6068297185048455,
|
| 1854 |
+
"grad_norm": 1.7112354026341845,
|
| 1855 |
+
"learning_rate": 8.0313236795169e-06,
|
| 1856 |
+
"loss": 0.9981,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.6091370558375635,
|
| 1861 |
+
"grad_norm": 1.7633777819452738,
|
| 1862 |
+
"learning_rate": 7.952411177044923e-06,
|
| 1863 |
+
"loss": 0.9667,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.6114443931702815,
|
| 1868 |
+
"grad_norm": 1.7477692209080626,
|
| 1869 |
+
"learning_rate": 7.873631539677364e-06,
|
| 1870 |
+
"loss": 0.9979,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.6137517305029996,
|
| 1875 |
+
"grad_norm": 1.7532055508610305,
|
| 1876 |
+
"learning_rate": 7.794989879311991e-06,
|
| 1877 |
+
"loss": 0.9869,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.6160590678357176,
|
| 1882 |
+
"grad_norm": 1.8525858143415055,
|
| 1883 |
+
"learning_rate": 7.716491298893443e-06,
|
| 1884 |
+
"loss": 0.9834,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.6183664051684357,
|
| 1889 |
+
"grad_norm": 1.749585519245075,
|
| 1890 |
+
"learning_rate": 7.638140892082118e-06,
|
| 1891 |
+
"loss": 1.0092,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.6206737425011537,
|
| 1896 |
+
"grad_norm": 1.8420135288729067,
|
| 1897 |
+
"learning_rate": 7.559943742923626e-06,
|
| 1898 |
+
"loss": 0.9797,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.6229810798338717,
|
| 1903 |
+
"grad_norm": 1.7361527256574634,
|
| 1904 |
+
"learning_rate": 7.4819049255189215e-06,
|
| 1905 |
+
"loss": 1.0084,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.6252884171665898,
|
| 1910 |
+
"grad_norm": 1.8341519418326866,
|
| 1911 |
+
"learning_rate": 7.404029503695028e-06,
|
| 1912 |
+
"loss": 0.978,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.6275957544993078,
|
| 1917 |
+
"grad_norm": 1.8293945335237427,
|
| 1918 |
+
"learning_rate": 7.326322530676471e-06,
|
| 1919 |
+
"loss": 0.9949,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.6299030918320259,
|
| 1924 |
+
"grad_norm": 1.8042095660293147,
|
| 1925 |
+
"learning_rate": 7.248789048757368e-06,
|
| 1926 |
+
"loss": 0.9708,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.6322104291647439,
|
| 1931 |
+
"grad_norm": 1.845467719423503,
|
| 1932 |
+
"learning_rate": 7.171434088974252e-06,
|
| 1933 |
+
"loss": 0.9965,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.6345177664974619,
|
| 1938 |
+
"grad_norm": 1.6543843384272663,
|
| 1939 |
+
"learning_rate": 7.094262670779611e-06,
|
| 1940 |
+
"loss": 0.9745,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.63682510383018,
|
| 1945 |
+
"grad_norm": 1.78812671106571,
|
| 1946 |
+
"learning_rate": 7.017279801716177e-06,
|
| 1947 |
+
"loss": 0.9913,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.639132441162898,
|
| 1952 |
+
"grad_norm": 1.6947334759904245,
|
| 1953 |
+
"learning_rate": 6.940490477092004e-06,
|
| 1954 |
+
"loss": 0.9852,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.6414397784956161,
|
| 1959 |
+
"grad_norm": 1.8535301270043634,
|
| 1960 |
+
"learning_rate": 6.8638996796563275e-06,
|
| 1961 |
+
"loss": 1.007,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.6437471158283341,
|
| 1966 |
+
"grad_norm": 1.7676355127694694,
|
| 1967 |
+
"learning_rate": 6.78751237927623e-06,
|
| 1968 |
+
"loss": 0.9514,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.6460544531610521,
|
| 1973 |
+
"grad_norm": 1.6769380120076558,
|
| 1974 |
+
"learning_rate": 6.711333532614168e-06,
|
| 1975 |
+
"loss": 0.9698,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.6483617904937702,
|
| 1980 |
+
"grad_norm": 1.7272039849376555,
|
| 1981 |
+
"learning_rate": 6.6353680828063306e-06,
|
| 1982 |
+
"loss": 0.948,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.6506691278264882,
|
| 1987 |
+
"grad_norm": 1.7909691104530978,
|
| 1988 |
+
"learning_rate": 6.559620959141897e-06,
|
| 1989 |
+
"loss": 0.9741,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.6529764651592063,
|
| 1994 |
+
"grad_norm": 1.7584119603336634,
|
| 1995 |
+
"learning_rate": 6.48409707674317e-06,
|
| 1996 |
+
"loss": 0.9825,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.6552838024919243,
|
| 2001 |
+
"grad_norm": 1.704146715339984,
|
| 2002 |
+
"learning_rate": 6.408801336246645e-06,
|
| 2003 |
+
"loss": 0.9473,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.6575911398246423,
|
| 2008 |
+
"grad_norm": 1.675928516675119,
|
| 2009 |
+
"learning_rate": 6.3337386234850255e-06,
|
| 2010 |
+
"loss": 0.9726,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.6598984771573604,
|
| 2015 |
+
"grad_norm": 1.8640939079623915,
|
| 2016 |
+
"learning_rate": 6.258913809170169e-06,
|
| 2017 |
+
"loss": 0.9899,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.6622058144900784,
|
| 2022 |
+
"grad_norm": 1.7077639495220778,
|
| 2023 |
+
"learning_rate": 6.18433174857705e-06,
|
| 2024 |
+
"loss": 0.9856,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.6645131518227965,
|
| 2029 |
+
"grad_norm": 1.7756594990657744,
|
| 2030 |
+
"learning_rate": 6.1099972812287e-06,
|
| 2031 |
+
"loss": 0.9766,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.6668204891555145,
|
| 2036 |
+
"grad_norm": 1.9246917842171538,
|
| 2037 |
+
"learning_rate": 6.035915230582176e-06,
|
| 2038 |
+
"loss": 0.9802,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.6691278264882325,
|
| 2043 |
+
"grad_norm": 1.7061051671690723,
|
| 2044 |
+
"learning_rate": 5.962090403715592e-06,
|
| 2045 |
+
"loss": 0.9589,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.6714351638209506,
|
| 2050 |
+
"grad_norm": 1.9209301423646885,
|
| 2051 |
+
"learning_rate": 5.8885275910161574e-06,
|
| 2052 |
+
"loss": 0.9661,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.6737425011536686,
|
| 2057 |
+
"grad_norm": 1.7522165825796936,
|
| 2058 |
+
"learning_rate": 5.815231565869377e-06,
|
| 2059 |
+
"loss": 0.9683,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.6760498384863867,
|
| 2064 |
+
"grad_norm": 1.7093168808099815,
|
| 2065 |
+
"learning_rate": 5.742207084349274e-06,
|
| 2066 |
+
"loss": 0.9787,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.6783571758191047,
|
| 2071 |
+
"grad_norm": 1.7331687059615726,
|
| 2072 |
+
"learning_rate": 5.669458884909815e-06,
|
| 2073 |
+
"loss": 0.962,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.6806645131518227,
|
| 2078 |
+
"grad_norm": 1.7125984142423774,
|
| 2079 |
+
"learning_rate": 5.596991688077409e-06,
|
| 2080 |
+
"loss": 0.9749,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.6829718504845408,
|
| 2085 |
+
"grad_norm": 1.7116106483633,
|
| 2086 |
+
"learning_rate": 5.5248101961446065e-06,
|
| 2087 |
+
"loss": 0.9646,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.6852791878172588,
|
| 2092 |
+
"grad_norm": 1.7871177917200074,
|
| 2093 |
+
"learning_rate": 5.452919092864976e-06,
|
| 2094 |
+
"loss": 0.9869,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.687586525149977,
|
| 2099 |
+
"grad_norm": 1.7422263712914812,
|
| 2100 |
+
"learning_rate": 5.381323043149192e-06,
|
| 2101 |
+
"loss": 0.9598,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.689893862482695,
|
| 2106 |
+
"grad_norm": 1.827833481936086,
|
| 2107 |
+
"learning_rate": 5.310026692762316e-06,
|
| 2108 |
+
"loss": 0.9674,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.6922011998154131,
|
| 2113 |
+
"grad_norm": 1.8013595644003924,
|
| 2114 |
+
"learning_rate": 5.239034668022353e-06,
|
| 2115 |
+
"loss": 0.9573,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.6945085371481311,
|
| 2120 |
+
"grad_norm": 1.8007495209856474,
|
| 2121 |
+
"learning_rate": 5.168351575500049e-06,
|
| 2122 |
+
"loss": 0.9719,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.6968158744808491,
|
| 2127 |
+
"grad_norm": 1.7404494760342795,
|
| 2128 |
+
"learning_rate": 5.097982001719994e-06,
|
| 2129 |
+
"loss": 0.9724,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.6991232118135672,
|
| 2134 |
+
"grad_norm": 1.8207400554800481,
|
| 2135 |
+
"learning_rate": 5.027930512862976e-06,
|
| 2136 |
+
"loss": 0.9643,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.7014305491462852,
|
| 2141 |
+
"grad_norm": 1.6503774911907483,
|
| 2142 |
+
"learning_rate": 4.958201654469731e-06,
|
| 2143 |
+
"loss": 0.9718,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.7037378864790033,
|
| 2148 |
+
"grad_norm": 1.7042499652030019,
|
| 2149 |
+
"learning_rate": 4.888799951145948e-06,
|
| 2150 |
+
"loss": 0.9786,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.7060452238117213,
|
| 2155 |
+
"grad_norm": 1.691429781718496,
|
| 2156 |
+
"learning_rate": 4.8197299062687e-06,
|
| 2157 |
+
"loss": 0.9584,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.7083525611444393,
|
| 2162 |
+
"grad_norm": 1.789471718965235,
|
| 2163 |
+
"learning_rate": 4.750996001694215e-06,
|
| 2164 |
+
"loss": 0.978,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.7106598984771574,
|
| 2169 |
+
"grad_norm": 1.762516310353455,
|
| 2170 |
+
"learning_rate": 4.6826026974670665e-06,
|
| 2171 |
+
"loss": 0.9536,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.7129672358098754,
|
| 2176 |
+
"grad_norm": 1.6808685507807348,
|
| 2177 |
+
"learning_rate": 4.614554431530754e-06,
|
| 2178 |
+
"loss": 0.9453,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.7152745731425935,
|
| 2183 |
+
"grad_norm": 1.695661878440997,
|
| 2184 |
+
"learning_rate": 4.546855619439734e-06,
|
| 2185 |
+
"loss": 0.9674,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.7175819104753115,
|
| 2190 |
+
"grad_norm": 1.780885617378124,
|
| 2191 |
+
"learning_rate": 4.479510654072909e-06,
|
| 2192 |
+
"loss": 0.9724,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.7198892478080295,
|
| 2197 |
+
"grad_norm": 1.7441807694903777,
|
| 2198 |
+
"learning_rate": 4.412523905348568e-06,
|
| 2199 |
+
"loss": 0.9422,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.7221965851407476,
|
| 2204 |
+
"grad_norm": 1.7666139001524914,
|
| 2205 |
+
"learning_rate": 4.345899719940844e-06,
|
| 2206 |
+
"loss": 0.9496,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.7245039224734656,
|
| 2211 |
+
"grad_norm": 1.6463010573052135,
|
| 2212 |
+
"learning_rate": 4.279642420997655e-06,
|
| 2213 |
+
"loss": 0.9635,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.7268112598061837,
|
| 2218 |
+
"grad_norm": 1.7790169033851828,
|
| 2219 |
+
"learning_rate": 4.213756307860175e-06,
|
| 2220 |
+
"loss": 0.9795,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.7291185971389017,
|
| 2225 |
+
"grad_norm": 1.709252193673288,
|
| 2226 |
+
"learning_rate": 4.148245655783869e-06,
|
| 2227 |
+
"loss": 0.9542,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.7314259344716197,
|
| 2232 |
+
"grad_norm": 1.7064646780964507,
|
| 2233 |
+
"learning_rate": 4.083114715661069e-06,
|
| 2234 |
+
"loss": 0.9494,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.7337332718043378,
|
| 2239 |
+
"grad_norm": 1.7203471522785316,
|
| 2240 |
+
"learning_rate": 4.018367713745137e-06,
|
| 2241 |
+
"loss": 0.9513,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.7360406091370558,
|
| 2246 |
+
"grad_norm": 1.7329896835019194,
|
| 2247 |
+
"learning_rate": 3.954008851376252e-06,
|
| 2248 |
+
"loss": 0.9464,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.7383479464697739,
|
| 2253 |
+
"grad_norm": 1.6668720129339225,
|
| 2254 |
+
"learning_rate": 3.890042304708758e-06,
|
| 2255 |
+
"loss": 0.9349,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.7406552838024919,
|
| 2260 |
+
"grad_norm": 1.6612958616670062,
|
| 2261 |
+
"learning_rate": 3.826472224440202e-06,
|
| 2262 |
+
"loss": 0.9753,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.7429626211352099,
|
| 2267 |
+
"grad_norm": 1.689937062434287,
|
| 2268 |
+
"learning_rate": 3.763302735541987e-06,
|
| 2269 |
+
"loss": 0.9755,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.745269958467928,
|
| 2274 |
+
"grad_norm": 1.8524303075498816,
|
| 2275 |
+
"learning_rate": 3.700537936991733e-06,
|
| 2276 |
+
"loss": 0.9919,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.747577295800646,
|
| 2281 |
+
"grad_norm": 1.7330330880413027,
|
| 2282 |
+
"learning_rate": 3.6381819015072652e-06,
|
| 2283 |
+
"loss": 0.9968,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.7498846331333641,
|
| 2288 |
+
"grad_norm": 1.732375990079818,
|
| 2289 |
+
"learning_rate": 3.5762386752823643e-06,
|
| 2290 |
+
"loss": 0.9598,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.7521919704660821,
|
| 2295 |
+
"grad_norm": 1.6723382538398348,
|
| 2296 |
+
"learning_rate": 3.5147122777242203e-06,
|
| 2297 |
+
"loss": 0.9826,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.7544993077988001,
|
| 2302 |
+
"grad_norm": 1.698445076932435,
|
| 2303 |
+
"learning_rate": 3.4536067011925945e-06,
|
| 2304 |
+
"loss": 0.975,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.7568066451315182,
|
| 2309 |
+
"grad_norm": 1.6756544799204833,
|
| 2310 |
+
"learning_rate": 3.3929259107407785e-06,
|
| 2311 |
+
"loss": 0.9596,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.7591139824642362,
|
| 2316 |
+
"grad_norm": 1.7323687941815844,
|
| 2317 |
+
"learning_rate": 3.3326738438583116e-06,
|
| 2318 |
+
"loss": 0.9471,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.7614213197969543,
|
| 2323 |
+
"grad_norm": 1.6841658818773522,
|
| 2324 |
+
"learning_rate": 3.272854410215467e-06,
|
| 2325 |
+
"loss": 0.9478,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.7637286571296723,
|
| 2330 |
+
"grad_norm": 1.7258401397718819,
|
| 2331 |
+
"learning_rate": 3.213471491409568e-06,
|
| 2332 |
+
"loss": 0.9545,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.7660359944623903,
|
| 2337 |
+
"grad_norm": 1.754250495342998,
|
| 2338 |
+
"learning_rate": 3.1545289407131128e-06,
|
| 2339 |
+
"loss": 0.9557,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.7683433317951085,
|
| 2344 |
+
"grad_norm": 1.7109892895946872,
|
| 2345 |
+
"learning_rate": 3.0960305828237568e-06,
|
| 2346 |
+
"loss": 0.9649,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.7706506691278265,
|
| 2351 |
+
"grad_norm": 1.8321981237158624,
|
| 2352 |
+
"learning_rate": 3.0379802136161073e-06,
|
| 2353 |
+
"loss": 0.9612,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.7729580064605446,
|
| 2358 |
+
"grad_norm": 1.6838996787097582,
|
| 2359 |
+
"learning_rate": 2.9803815998954334e-06,
|
| 2360 |
+
"loss": 0.9701,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.7752653437932626,
|
| 2365 |
+
"grad_norm": 1.7069776476837635,
|
| 2366 |
+
"learning_rate": 2.9232384791532377e-06,
|
| 2367 |
+
"loss": 0.9724,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.7775726811259807,
|
| 2372 |
+
"grad_norm": 1.6302606016182208,
|
| 2373 |
+
"learning_rate": 2.866554559324731e-06,
|
| 2374 |
+
"loss": 0.9441,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.7798800184586987,
|
| 2379 |
+
"grad_norm": 1.6451304542174006,
|
| 2380 |
+
"learning_rate": 2.810333518548246e-06,
|
| 2381 |
+
"loss": 0.9337,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.7821873557914167,
|
| 2386 |
+
"grad_norm": 1.763422741436097,
|
| 2387 |
+
"learning_rate": 2.7545790049265506e-06,
|
| 2388 |
+
"loss": 0.9542,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.7844946931241348,
|
| 2393 |
+
"grad_norm": 1.6780739225283752,
|
| 2394 |
+
"learning_rate": 2.699294636290134e-06,
|
| 2395 |
+
"loss": 0.9468,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.7868020304568528,
|
| 2400 |
+
"grad_norm": 1.7170452157015115,
|
| 2401 |
+
"learning_rate": 2.6444839999624496e-06,
|
| 2402 |
+
"loss": 0.9333,
|
| 2403 |
+
"step": 1705
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.7891093677895709,
|
| 2407 |
+
"grad_norm": 1.6339639009427172,
|
| 2408 |
+
"learning_rate": 2.5901506525271424e-06,
|
| 2409 |
+
"loss": 0.9656,
|
| 2410 |
+
"step": 1710
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.7914167051222889,
|
| 2414 |
+
"grad_norm": 1.7076055984466658,
|
| 2415 |
+
"learning_rate": 2.5362981195972627e-06,
|
| 2416 |
+
"loss": 0.9292,
|
| 2417 |
+
"step": 1715
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.7937240424550069,
|
| 2421 |
+
"grad_norm": 1.694775381636099,
|
| 2422 |
+
"learning_rate": 2.4829298955865022e-06,
|
| 2423 |
+
"loss": 0.9621,
|
| 2424 |
+
"step": 1720
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.796031379787725,
|
| 2428 |
+
"grad_norm": 1.6448376757312444,
|
| 2429 |
+
"learning_rate": 2.4300494434824373e-06,
|
| 2430 |
+
"loss": 0.9323,
|
| 2431 |
+
"step": 1725
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.798338717120443,
|
| 2435 |
+
"grad_norm": 1.7190721711044321,
|
| 2436 |
+
"learning_rate": 2.3776601946218225e-06,
|
| 2437 |
+
"loss": 0.9536,
|
| 2438 |
+
"step": 1730
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.8006460544531611,
|
| 2442 |
+
"grad_norm": 1.7369056682520372,
|
| 2443 |
+
"learning_rate": 2.3257655484679376e-06,
|
| 2444 |
+
"loss": 0.9474,
|
| 2445 |
+
"step": 1735
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.8029533917858791,
|
| 2449 |
+
"grad_norm": 1.9277048905233987,
|
| 2450 |
+
"learning_rate": 2.274368872390009e-06,
|
| 2451 |
+
"loss": 0.953,
|
| 2452 |
+
"step": 1740
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.8052607291185971,
|
| 2456 |
+
"grad_norm": 1.6988058479966548,
|
| 2457 |
+
"learning_rate": 2.2234735014446905e-06,
|
| 2458 |
+
"loss": 0.9546,
|
| 2459 |
+
"step": 1745
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.8075680664513152,
|
| 2463 |
+
"grad_norm": 1.708926311661711,
|
| 2464 |
+
"learning_rate": 2.1730827381596643e-06,
|
| 2465 |
+
"loss": 0.9442,
|
| 2466 |
+
"step": 1750
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.8098754037840332,
|
| 2470 |
+
"grad_norm": 1.7384137551353784,
|
| 2471 |
+
"learning_rate": 2.123199852319352e-06,
|
| 2472 |
+
"loss": 0.9415,
|
| 2473 |
+
"step": 1755
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.8121827411167513,
|
| 2477 |
+
"grad_norm": 1.706865871432203,
|
| 2478 |
+
"learning_rate": 2.073828080752728e-06,
|
| 2479 |
+
"loss": 0.9514,
|
| 2480 |
+
"step": 1760
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.8144900784494693,
|
| 2484 |
+
"grad_norm": 1.631591345911517,
|
| 2485 |
+
"learning_rate": 2.024970627123295e-06,
|
| 2486 |
+
"loss": 0.9593,
|
| 2487 |
+
"step": 1765
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.8167974157821873,
|
| 2491 |
+
"grad_norm": 1.7250014472303201,
|
| 2492 |
+
"learning_rate": 1.976630661721207e-06,
|
| 2493 |
+
"loss": 0.9312,
|
| 2494 |
+
"step": 1770
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.8191047531149054,
|
| 2498 |
+
"grad_norm": 1.7176929983275837,
|
| 2499 |
+
"learning_rate": 1.9288113212575454e-06,
|
| 2500 |
+
"loss": 0.9392,
|
| 2501 |
+
"step": 1775
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.8214120904476234,
|
| 2505 |
+
"grad_norm": 1.7515517183747666,
|
| 2506 |
+
"learning_rate": 1.8815157086607826e-06,
|
| 2507 |
+
"loss": 0.9461,
|
| 2508 |
+
"step": 1780
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.8237194277803415,
|
| 2512 |
+
"grad_norm": 1.7256573401014044,
|
| 2513 |
+
"learning_rate": 1.8347468928754408e-06,
|
| 2514 |
+
"loss": 0.9625,
|
| 2515 |
+
"step": 1785
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.8260267651130595,
|
| 2519 |
+
"grad_norm": 1.6925351828448565,
|
| 2520 |
+
"learning_rate": 1.7885079086629598e-06,
|
| 2521 |
+
"loss": 0.9618,
|
| 2522 |
+
"step": 1790
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.8283341024457775,
|
| 2526 |
+
"grad_norm": 1.6848423059711715,
|
| 2527 |
+
"learning_rate": 1.7428017564047594e-06,
|
| 2528 |
+
"loss": 0.957,
|
| 2529 |
+
"step": 1795
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.8306414397784956,
|
| 2533 |
+
"grad_norm": 1.685378392680059,
|
| 2534 |
+
"learning_rate": 1.697631401907559e-06,
|
| 2535 |
+
"loss": 0.9405,
|
| 2536 |
+
"step": 1800
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.8329487771112136,
|
| 2540 |
+
"grad_norm": 1.779147836504438,
|
| 2541 |
+
"learning_rate": 1.6529997762109319e-06,
|
| 2542 |
+
"loss": 0.9475,
|
| 2543 |
+
"step": 1805
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.8352561144439317,
|
| 2547 |
+
"grad_norm": 1.6862245640499274,
|
| 2548 |
+
"learning_rate": 1.6089097753971061e-06,
|
| 2549 |
+
"loss": 0.9433,
|
| 2550 |
+
"step": 1810
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.8375634517766497,
|
| 2554 |
+
"grad_norm": 1.6387699919494911,
|
| 2555 |
+
"learning_rate": 1.565364260403055e-06,
|
| 2556 |
+
"loss": 0.9393,
|
| 2557 |
+
"step": 1815
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.8398707891093677,
|
| 2561 |
+
"grad_norm": 1.7088328181524817,
|
| 2562 |
+
"learning_rate": 1.522366056834844e-06,
|
| 2563 |
+
"loss": 0.9322,
|
| 2564 |
+
"step": 1820
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.8421781264420858,
|
| 2568 |
+
"grad_norm": 1.717214554275275,
|
| 2569 |
+
"learning_rate": 1.4799179547842823e-06,
|
| 2570 |
+
"loss": 0.9393,
|
| 2571 |
+
"step": 1825
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.8444854637748038,
|
| 2575 |
+
"grad_norm": 1.717143943169584,
|
| 2576 |
+
"learning_rate": 1.4380227086478816e-06,
|
| 2577 |
+
"loss": 0.96,
|
| 2578 |
+
"step": 1830
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.846792801107522,
|
| 2582 |
+
"grad_norm": 1.7390583641873172,
|
| 2583 |
+
"learning_rate": 1.3966830369481231e-06,
|
| 2584 |
+
"loss": 0.9487,
|
| 2585 |
+
"step": 1835
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.84910013844024,
|
| 2589 |
+
"grad_norm": 1.6483510037137357,
|
| 2590 |
+
"learning_rate": 1.3559016221570663e-06,
|
| 2591 |
+
"loss": 0.9315,
|
| 2592 |
+
"step": 1840
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.8514074757729581,
|
| 2596 |
+
"grad_norm": 1.7716145374153562,
|
| 2597 |
+
"learning_rate": 1.3156811105222723e-06,
|
| 2598 |
+
"loss": 0.9375,
|
| 2599 |
+
"step": 1845
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.8537148131056761,
|
| 2603 |
+
"grad_norm": 1.7136369284376767,
|
| 2604 |
+
"learning_rate": 1.276024111895101e-06,
|
| 2605 |
+
"loss": 0.9592,
|
| 2606 |
+
"step": 1850
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.8560221504383941,
|
| 2610 |
+
"grad_norm": 1.659445261277345,
|
| 2611 |
+
"learning_rate": 1.2369331995613664e-06,
|
| 2612 |
+
"loss": 0.9466,
|
| 2613 |
+
"step": 1855
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.8583294877711122,
|
| 2617 |
+
"grad_norm": 1.7022849150801465,
|
| 2618 |
+
"learning_rate": 1.1984109100743445e-06,
|
| 2619 |
+
"loss": 0.934,
|
| 2620 |
+
"step": 1860
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.8606368251038302,
|
| 2624 |
+
"grad_norm": 1.7296616243070897,
|
| 2625 |
+
"learning_rate": 1.1604597430902032e-06,
|
| 2626 |
+
"loss": 0.9413,
|
| 2627 |
+
"step": 1865
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.8629441624365483,
|
| 2631 |
+
"grad_norm": 1.70400411645417,
|
| 2632 |
+
"learning_rate": 1.123082161205775e-06,
|
| 2633 |
+
"loss": 0.9192,
|
| 2634 |
+
"step": 1870
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.8652514997692663,
|
| 2638 |
+
"grad_norm": 1.7716704719549248,
|
| 2639 |
+
"learning_rate": 1.0862805897987894e-06,
|
| 2640 |
+
"loss": 0.9313,
|
| 2641 |
+
"step": 1875
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.8675588371019843,
|
| 2645 |
+
"grad_norm": 1.6028791905428008,
|
| 2646 |
+
"learning_rate": 1.0500574168704746e-06,
|
| 2647 |
+
"loss": 0.9222,
|
| 2648 |
+
"step": 1880
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.8698661744347024,
|
| 2652 |
+
"grad_norm": 1.6811903757943443,
|
| 2653 |
+
"learning_rate": 1.014414992890611e-06,
|
| 2654 |
+
"loss": 0.9613,
|
| 2655 |
+
"step": 1885
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.8721735117674204,
|
| 2659 |
+
"grad_norm": 1.6582769629121996,
|
| 2660 |
+
"learning_rate": 9.793556306450125e-07,
|
| 2661 |
+
"loss": 0.9397,
|
| 2662 |
+
"step": 1890
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.8744808491001385,
|
| 2666 |
+
"grad_norm": 1.7329477389715338,
|
| 2667 |
+
"learning_rate": 9.448816050854559e-07,
|
| 2668 |
+
"loss": 0.9456,
|
| 2669 |
+
"step": 1895
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.8767881864328565,
|
| 2673 |
+
"grad_norm": 1.6579278218600528,
|
| 2674 |
+
"learning_rate": 9.10995153182056e-07,
|
| 2675 |
+
"loss": 0.9622,
|
| 2676 |
+
"step": 1900
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.8790955237655745,
|
| 2680 |
+
"grad_norm": 1.7059466024322731,
|
| 2681 |
+
"learning_rate": 8.776984737781135e-07,
|
| 2682 |
+
"loss": 0.9247,
|
| 2683 |
+
"step": 1905
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.8814028610982926,
|
| 2687 |
+
"grad_norm": 1.6587934531932846,
|
| 2688 |
+
"learning_rate": 8.449937274474396e-07,
|
| 2689 |
+
"loss": 0.9287,
|
| 2690 |
+
"step": 1910
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.8837101984310106,
|
| 2694 |
+
"grad_norm": 1.7000421816452764,
|
| 2695 |
+
"learning_rate": 8.128830363541574e-07,
|
| 2696 |
+
"loss": 0.9579,
|
| 2697 |
+
"step": 1915
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.8860175357637287,
|
| 2701 |
+
"grad_norm": 1.5982289727068608,
|
| 2702 |
+
"learning_rate": 7.81368484114996e-07,
|
| 2703 |
+
"loss": 0.9252,
|
| 2704 |
+
"step": 1920
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.8883248730964467,
|
| 2708 |
+
"grad_norm": 1.6595039047537794,
|
| 2709 |
+
"learning_rate": 7.504521156640854e-07,
|
| 2710 |
+
"loss": 0.9535,
|
| 2711 |
+
"step": 1925
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.8906322104291647,
|
| 2715 |
+
"grad_norm": 1.67142320517366,
|
| 2716 |
+
"learning_rate": 7.201359371202698e-07,
|
| 2717 |
+
"loss": 0.9342,
|
| 2718 |
+
"step": 1930
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.8929395477618828,
|
| 2722 |
+
"grad_norm": 1.724210373217245,
|
| 2723 |
+
"learning_rate": 6.904219156569325e-07,
|
| 2724 |
+
"loss": 0.9537,
|
| 2725 |
+
"step": 1935
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.8952468850946008,
|
| 2729 |
+
"grad_norm": 1.775137829218875,
|
| 2730 |
+
"learning_rate": 6.613119793743428e-07,
|
| 2731 |
+
"loss": 0.9407,
|
| 2732 |
+
"step": 1940
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.8975542224273189,
|
| 2736 |
+
"grad_norm": 1.6829966494434596,
|
| 2737 |
+
"learning_rate": 6.32808017174551e-07,
|
| 2738 |
+
"loss": 0.9271,
|
| 2739 |
+
"step": 1945
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.8998615597600369,
|
| 2743 |
+
"grad_norm": 1.7666732440040138,
|
| 2744 |
+
"learning_rate": 6.049118786388153e-07,
|
| 2745 |
+
"loss": 0.9299,
|
| 2746 |
+
"step": 1950
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.9021688970927549,
|
| 2750 |
+
"grad_norm": 1.7328446243608238,
|
| 2751 |
+
"learning_rate": 5.776253739075887e-07,
|
| 2752 |
+
"loss": 0.9368,
|
| 2753 |
+
"step": 1955
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.904476234425473,
|
| 2757 |
+
"grad_norm": 1.794453478228994,
|
| 2758 |
+
"learning_rate": 5.509502735630601e-07,
|
| 2759 |
+
"loss": 0.9584,
|
| 2760 |
+
"step": 1960
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.906783571758191,
|
| 2764 |
+
"grad_norm": 1.8595416734645607,
|
| 2765 |
+
"learning_rate": 5.248883085142653e-07,
|
| 2766 |
+
"loss": 0.9278,
|
| 2767 |
+
"step": 1965
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.9090909090909091,
|
| 2771 |
+
"grad_norm": 1.8064599818138745,
|
| 2772 |
+
"learning_rate": 4.994411698847668e-07,
|
| 2773 |
+
"loss": 0.9521,
|
| 2774 |
+
"step": 1970
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.9113982464236271,
|
| 2778 |
+
"grad_norm": 11.97761580494739,
|
| 2779 |
+
"learning_rate": 4.746105089029229e-07,
|
| 2780 |
+
"loss": 0.9353,
|
| 2781 |
+
"step": 1975
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.9137055837563451,
|
| 2785 |
+
"grad_norm": 1.91129591540662,
|
| 2786 |
+
"learning_rate": 4.50397936794742e-07,
|
| 2787 |
+
"loss": 0.9518,
|
| 2788 |
+
"step": 1980
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.9160129210890632,
|
| 2792 |
+
"grad_norm": 1.68892958160335,
|
| 2793 |
+
"learning_rate": 4.268050246793276e-07,
|
| 2794 |
+
"loss": 0.9417,
|
| 2795 |
+
"step": 1985
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.9183202584217812,
|
| 2799 |
+
"grad_norm": 1.7682216031642473,
|
| 2800 |
+
"learning_rate": 4.038333034669406e-07,
|
| 2801 |
+
"loss": 0.9575,
|
| 2802 |
+
"step": 1990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.9206275957544993,
|
| 2806 |
+
"grad_norm": 1.6202959145852942,
|
| 2807 |
+
"learning_rate": 3.814842637596483e-07,
|
| 2808 |
+
"loss": 0.9202,
|
| 2809 |
+
"step": 1995
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.9229349330872173,
|
| 2813 |
+
"grad_norm": 1.7355964322786217,
|
| 2814 |
+
"learning_rate": 3.5975935575461083e-07,
|
| 2815 |
+
"loss": 0.9408,
|
| 2816 |
+
"step": 2000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.9252422704199353,
|
| 2820 |
+
"grad_norm": 1.784833888232325,
|
| 2821 |
+
"learning_rate": 3.3865998914997645e-07,
|
| 2822 |
+
"loss": 0.9451,
|
| 2823 |
+
"step": 2005
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.9275496077526535,
|
| 2827 |
+
"grad_norm": 1.7347540917405002,
|
| 2828 |
+
"learning_rate": 3.1818753305340566e-07,
|
| 2829 |
+
"loss": 0.9503,
|
| 2830 |
+
"step": 2010
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.9298569450853715,
|
| 2834 |
+
"grad_norm": 1.7098932522181238,
|
| 2835 |
+
"learning_rate": 2.9834331589323697e-07,
|
| 2836 |
+
"loss": 0.9648,
|
| 2837 |
+
"step": 2015
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.9321642824180896,
|
| 2841 |
+
"grad_norm": 1.7375733049730988,
|
| 2842 |
+
"learning_rate": 2.791286253322856e-07,
|
| 2843 |
+
"loss": 0.9325,
|
| 2844 |
+
"step": 2020
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.9344716197508076,
|
| 2848 |
+
"grad_norm": 1.7376811511996035,
|
| 2849 |
+
"learning_rate": 2.605447081842838e-07,
|
| 2850 |
+
"loss": 0.9236,
|
| 2851 |
+
"step": 2025
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.9367789570835257,
|
| 2855 |
+
"grad_norm": 1.6285382014408727,
|
| 2856 |
+
"learning_rate": 2.425927703329856e-07,
|
| 2857 |
+
"loss": 0.9374,
|
| 2858 |
+
"step": 2030
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.9390862944162437,
|
| 2862 |
+
"grad_norm": 1.7187090366001978,
|
| 2863 |
+
"learning_rate": 2.2527397665391026e-07,
|
| 2864 |
+
"loss": 0.9408,
|
| 2865 |
+
"step": 2035
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.9413936317489617,
|
| 2869 |
+
"grad_norm": 1.6018424242699771,
|
| 2870 |
+
"learning_rate": 2.0858945093876315e-07,
|
| 2871 |
+
"loss": 0.9255,
|
| 2872 |
+
"step": 2040
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.9437009690816798,
|
| 2876 |
+
"grad_norm": 1.7403672056926338,
|
| 2877 |
+
"learning_rate": 1.9254027582250588e-07,
|
| 2878 |
+
"loss": 0.9386,
|
| 2879 |
+
"step": 2045
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.9460083064143978,
|
| 2883 |
+
"grad_norm": 1.6413885110477684,
|
| 2884 |
+
"learning_rate": 1.7712749271311392e-07,
|
| 2885 |
+
"loss": 0.9463,
|
| 2886 |
+
"step": 2050
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.9483156437471159,
|
| 2890 |
+
"grad_norm": 1.6584652712298515,
|
| 2891 |
+
"learning_rate": 1.6235210172399373e-07,
|
| 2892 |
+
"loss": 0.9197,
|
| 2893 |
+
"step": 2055
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.9506229810798339,
|
| 2897 |
+
"grad_norm": 1.7146073811212112,
|
| 2898 |
+
"learning_rate": 1.4821506160909492e-07,
|
| 2899 |
+
"loss": 0.9325,
|
| 2900 |
+
"step": 2060
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.9529303184125519,
|
| 2904 |
+
"grad_norm": 1.7862356926674563,
|
| 2905 |
+
"learning_rate": 1.3471728970068986e-07,
|
| 2906 |
+
"loss": 0.9415,
|
| 2907 |
+
"step": 2065
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.95523765574527,
|
| 2911 |
+
"grad_norm": 1.7900835474706864,
|
| 2912 |
+
"learning_rate": 1.2185966184985687e-07,
|
| 2913 |
+
"loss": 0.9516,
|
| 2914 |
+
"step": 2070
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.957544993077988,
|
| 2918 |
+
"grad_norm": 1.61965861822062,
|
| 2919 |
+
"learning_rate": 1.0964301236963904e-07,
|
| 2920 |
+
"loss": 0.9272,
|
| 2921 |
+
"step": 2075
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.9598523304107061,
|
| 2925 |
+
"grad_norm": 1.752477339334271,
|
| 2926 |
+
"learning_rate": 9.806813398091419e-08,
|
| 2927 |
+
"loss": 0.9231,
|
| 2928 |
+
"step": 2080
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.9621596677434241,
|
| 2932 |
+
"grad_norm": 1.6566621651012274,
|
| 2933 |
+
"learning_rate": 8.713577776095494e-08,
|
| 2934 |
+
"loss": 0.9293,
|
| 2935 |
+
"step": 2085
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.9644670050761421,
|
| 2939 |
+
"grad_norm": 1.637994937785989,
|
| 2940 |
+
"learning_rate": 7.684665309468875e-08,
|
| 2941 |
+
"loss": 0.9539,
|
| 2942 |
+
"step": 2090
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.9667743424088602,
|
| 2946 |
+
"grad_norm": 1.6112435656292767,
|
| 2947 |
+
"learning_rate": 6.720142762867032e-08,
|
| 2948 |
+
"loss": 0.9558,
|
| 2949 |
+
"step": 2095
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.9690816797415782,
|
| 2953 |
+
"grad_norm": 1.7629733611663445,
|
| 2954 |
+
"learning_rate": 5.820072722775849e-08,
|
| 2955 |
+
"loss": 0.9441,
|
| 2956 |
+
"step": 2100
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.9713890170742963,
|
| 2960 |
+
"grad_norm": 1.6502417380976349,
|
| 2961 |
+
"learning_rate": 4.984513593450424e-08,
|
| 2962 |
+
"loss": 0.9527,
|
| 2963 |
+
"step": 2105
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.9736963544070143,
|
| 2967 |
+
"grad_norm": 1.697689686578805,
|
| 2968 |
+
"learning_rate": 4.2135195931249925e-08,
|
| 2969 |
+
"loss": 0.9468,
|
| 2970 |
+
"step": 2110
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.9760036917397323,
|
| 2974 |
+
"grad_norm": 1.7391293446601994,
|
| 2975 |
+
"learning_rate": 3.50714075049563e-08,
|
| 2976 |
+
"loss": 0.932,
|
| 2977 |
+
"step": 2115
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.9783110290724504,
|
| 2981 |
+
"grad_norm": 1.7092970910189305,
|
| 2982 |
+
"learning_rate": 2.8654229014730694e-08,
|
| 2983 |
+
"loss": 0.9377,
|
| 2984 |
+
"step": 2120
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.9806183664051684,
|
| 2988 |
+
"grad_norm": 1.6908979053958657,
|
| 2989 |
+
"learning_rate": 2.2884076862089712e-08,
|
| 2990 |
+
"loss": 0.9238,
|
| 2991 |
+
"step": 2125
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.9829257037378865,
|
| 2995 |
+
"grad_norm": 1.634296694613385,
|
| 2996 |
+
"learning_rate": 1.7761325463937495e-08,
|
| 2997 |
+
"loss": 0.9473,
|
| 2998 |
+
"step": 2130
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.9852330410706045,
|
| 3002 |
+
"grad_norm": 1.700195091519778,
|
| 3003 |
+
"learning_rate": 1.3286307228269623e-08,
|
| 3004 |
+
"loss": 0.9491,
|
| 3005 |
+
"step": 2135
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.9875403784033225,
|
| 3009 |
+
"grad_norm": 1.6899977441806657,
|
| 3010 |
+
"learning_rate": 9.459312532608122e-09,
|
| 3011 |
+
"loss": 0.9393,
|
| 3012 |
+
"step": 2140
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.9898477157360406,
|
| 3016 |
+
"grad_norm": 1.6459431767651571,
|
| 3017 |
+
"learning_rate": 6.280589705153217e-09,
|
| 3018 |
+
"loss": 0.9316,
|
| 3019 |
+
"step": 2145
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.9921550530687586,
|
| 3023 |
+
"grad_norm": 1.6824622254831971,
|
| 3024 |
+
"learning_rate": 3.750345008675105e-09,
|
| 3025 |
+
"loss": 0.9455,
|
| 3026 |
+
"step": 2150
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.9944623904014767,
|
| 3030 |
+
"grad_norm": 1.6711402676574085,
|
| 3031 |
+
"learning_rate": 1.8687426271246646e-09,
|
| 3032 |
+
"loss": 0.9299,
|
| 3033 |
+
"step": 2155
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.9967697277341947,
|
| 3037 |
+
"grad_norm": 1.7394983815714609,
|
| 3038 |
+
"learning_rate": 6.359046549864189e-10,
|
| 3039 |
+
"loss": 0.9262,
|
| 3040 |
+
"step": 2160
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.9990770650669127,
|
| 3044 |
+
"grad_norm": 1.9029748936658466,
|
| 3045 |
+
"learning_rate": 5.1911089347100876e-11,
|
| 3046 |
+
"loss": 0.973,
|
| 3047 |
+
"step": 2165
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 1.0,
|
| 3051 |
+
"eval_loss": 0.9539673924446106,
|
| 3052 |
+
"eval_runtime": 317.6397,
|
| 3053 |
+
"eval_samples_per_second": 48.325,
|
| 3054 |
+
"eval_steps_per_second": 0.756,
|
| 3055 |
+
"step": 2167
|
| 3056 |
+
},
|
| 3057 |
+
{
|
| 3058 |
+
"epoch": 1.0,
|
| 3059 |
+
"step": 2167,
|
| 3060 |
+
"total_flos": 453725713858560.0,
|
| 3061 |
+
"train_loss": 1.0076359760722753,
|
| 3062 |
+
"train_runtime": 13620.7718,
|
| 3063 |
+
"train_samples_per_second": 10.182,
|
| 3064 |
+
"train_steps_per_second": 0.159
|
| 3065 |
+
}
|
| 3066 |
+
],
|
| 3067 |
+
"logging_steps": 5,
|
| 3068 |
+
"max_steps": 2167,
|
| 3069 |
+
"num_input_tokens_seen": 0,
|
| 3070 |
+
"num_train_epochs": 1,
|
| 3071 |
+
"save_steps": 100,
|
| 3072 |
+
"stateful_callbacks": {
|
| 3073 |
+
"TrainerControl": {
|
| 3074 |
+
"args": {
|
| 3075 |
+
"should_epoch_stop": false,
|
| 3076 |
+
"should_evaluate": false,
|
| 3077 |
+
"should_log": false,
|
| 3078 |
+
"should_save": true,
|
| 3079 |
+
"should_training_stop": true
|
| 3080 |
+
},
|
| 3081 |
+
"attributes": {}
|
| 3082 |
+
}
|
| 3083 |
+
},
|
| 3084 |
+
"total_flos": 453725713858560.0,
|
| 3085 |
+
"train_batch_size": 8,
|
| 3086 |
+
"trial_name": null,
|
| 3087 |
+
"trial_params": null
|
| 3088 |
+
}
|