End of training
Browse files- README.md +54 -0
- adapter_config.json +28 -0
- adapter_model.bin +3 -0
- adapter_model.safetensors +3 -0
- all_results.json +7 -0
- args.txt +14 -0
- train_results.json +7 -0
- trainer_state.json +1948 -0
- training_args.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: llama2
|
| 3 |
+
base_model: lmsys/vicuna-7b-v1.5
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
model-index:
|
| 7 |
+
- name: arc_cot_64
|
| 8 |
+
results: []
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 12 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 13 |
+
|
| 14 |
+
# arc_cot_64
|
| 15 |
+
|
| 16 |
+
This model is a fine-tuned version of [lmsys/vicuna-7b-v1.5](https://huggingface.co/lmsys/vicuna-7b-v1.5) on an unknown dataset.
|
| 17 |
+
|
| 18 |
+
## Model description
|
| 19 |
+
|
| 20 |
+
More information needed
|
| 21 |
+
|
| 22 |
+
## Intended uses & limitations
|
| 23 |
+
|
| 24 |
+
More information needed
|
| 25 |
+
|
| 26 |
+
## Training and evaluation data
|
| 27 |
+
|
| 28 |
+
More information needed
|
| 29 |
+
|
| 30 |
+
## Training procedure
|
| 31 |
+
|
| 32 |
+
### Training hyperparameters
|
| 33 |
+
|
| 34 |
+
The following hyperparameters were used during training:
|
| 35 |
+
- learning_rate: 0.0001
|
| 36 |
+
- train_batch_size: 4
|
| 37 |
+
- eval_batch_size: 8
|
| 38 |
+
- seed: 42
|
| 39 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 40 |
+
- lr_scheduler_type: linear
|
| 41 |
+
- lr_scheduler_warmup_steps: 5
|
| 42 |
+
- num_epochs: 20
|
| 43 |
+
- mixed_precision_training: Native AMP
|
| 44 |
+
|
| 45 |
+
### Training results
|
| 46 |
+
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
### Framework versions
|
| 50 |
+
|
| 51 |
+
- Transformers 4.36.0.dev0
|
| 52 |
+
- Pytorch 2.1.0+cu121
|
| 53 |
+
- Datasets 2.13.1
|
| 54 |
+
- Tokenizers 0.14.1
|
adapter_config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "lmsys/vicuna-7b-v1.5",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"fan_in_fan_out": false,
|
| 7 |
+
"inference_mode": true,
|
| 8 |
+
"init_lora_weights": true,
|
| 9 |
+
"layers_pattern": null,
|
| 10 |
+
"layers_to_transform": null,
|
| 11 |
+
"lora_alpha": 64,
|
| 12 |
+
"lora_dropout": 0.1,
|
| 13 |
+
"modules_to_save": null,
|
| 14 |
+
"peft_type": "LORA",
|
| 15 |
+
"r": 16,
|
| 16 |
+
"rank_pattern": {},
|
| 17 |
+
"revision": null,
|
| 18 |
+
"target_modules": [
|
| 19 |
+
"o_proj",
|
| 20 |
+
"up_proj",
|
| 21 |
+
"q_proj",
|
| 22 |
+
"v_proj",
|
| 23 |
+
"down_proj",
|
| 24 |
+
"gate_proj",
|
| 25 |
+
"k_proj"
|
| 26 |
+
],
|
| 27 |
+
"task_type": "CAUSAL_LM"
|
| 28 |
+
}
|
adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:38c95b1fd1c9134ed992e040ce784ce49950b4b4e1e1d60512078216e5de8785
|
| 3 |
+
size 160069834
|
adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c786aa809854b200fa557dbcf847d21a2935375619f688596a398b19834d785a
|
| 3 |
+
size 159967880
|
all_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"train_loss": 0.26344227014342325,
|
| 4 |
+
"train_runtime": 536.4097,
|
| 5 |
+
"train_samples_per_second": 2.386,
|
| 6 |
+
"train_steps_per_second": 0.597
|
| 7 |
+
}
|
args.txt
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
base_model_name: lmsys/vicuna-7b-v1.5
|
| 2 |
+
batch_size: 4
|
| 3 |
+
cot: true
|
| 4 |
+
dataset_name: BENBENBENb/ARC1000COT
|
| 5 |
+
epochs: 20
|
| 6 |
+
eval_strategy: epoch
|
| 7 |
+
evaluation: false
|
| 8 |
+
learning_rate: 0.0001
|
| 9 |
+
logging_steps: 1
|
| 10 |
+
output_dir: brettbbb/arc_cot_64
|
| 11 |
+
seed: 42
|
| 12 |
+
skip_example: true
|
| 13 |
+
train_size: 64
|
| 14 |
+
warmup_steps: 5
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"train_loss": 0.26344227014342325,
|
| 4 |
+
"train_runtime": 536.4097,
|
| 5 |
+
"train_samples_per_second": 2.386,
|
| 6 |
+
"train_steps_per_second": 0.597
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1948 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 20.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 320,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.06,
|
| 13 |
+
"learning_rate": 2e-05,
|
| 14 |
+
"loss": 1.9389,
|
| 15 |
+
"step": 1
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.12,
|
| 19 |
+
"learning_rate": 4e-05,
|
| 20 |
+
"loss": 2.5146,
|
| 21 |
+
"step": 2
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.19,
|
| 25 |
+
"learning_rate": 6e-05,
|
| 26 |
+
"loss": 2.2218,
|
| 27 |
+
"step": 3
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.25,
|
| 31 |
+
"learning_rate": 8e-05,
|
| 32 |
+
"loss": 2.0735,
|
| 33 |
+
"step": 4
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.31,
|
| 37 |
+
"learning_rate": 0.0001,
|
| 38 |
+
"loss": 2.3033,
|
| 39 |
+
"step": 5
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.38,
|
| 43 |
+
"learning_rate": 9.968253968253969e-05,
|
| 44 |
+
"loss": 1.9486,
|
| 45 |
+
"step": 6
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.44,
|
| 49 |
+
"learning_rate": 9.936507936507938e-05,
|
| 50 |
+
"loss": 2.0072,
|
| 51 |
+
"step": 7
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.5,
|
| 55 |
+
"learning_rate": 9.904761904761905e-05,
|
| 56 |
+
"loss": 1.9243,
|
| 57 |
+
"step": 8
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.56,
|
| 61 |
+
"learning_rate": 9.873015873015874e-05,
|
| 62 |
+
"loss": 1.7337,
|
| 63 |
+
"step": 9
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.62,
|
| 67 |
+
"learning_rate": 9.841269841269841e-05,
|
| 68 |
+
"loss": 1.4964,
|
| 69 |
+
"step": 10
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.69,
|
| 73 |
+
"learning_rate": 9.80952380952381e-05,
|
| 74 |
+
"loss": 1.8149,
|
| 75 |
+
"step": 11
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.75,
|
| 79 |
+
"learning_rate": 9.777777777777778e-05,
|
| 80 |
+
"loss": 1.3444,
|
| 81 |
+
"step": 12
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.81,
|
| 85 |
+
"learning_rate": 9.746031746031747e-05,
|
| 86 |
+
"loss": 1.5002,
|
| 87 |
+
"step": 13
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.88,
|
| 91 |
+
"learning_rate": 9.714285714285715e-05,
|
| 92 |
+
"loss": 1.2595,
|
| 93 |
+
"step": 14
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.94,
|
| 97 |
+
"learning_rate": 9.682539682539682e-05,
|
| 98 |
+
"loss": 1.5071,
|
| 99 |
+
"step": 15
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 1.0,
|
| 103 |
+
"learning_rate": 9.650793650793651e-05,
|
| 104 |
+
"loss": 1.392,
|
| 105 |
+
"step": 16
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.06,
|
| 109 |
+
"learning_rate": 9.61904761904762e-05,
|
| 110 |
+
"loss": 1.3785,
|
| 111 |
+
"step": 17
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 1.12,
|
| 115 |
+
"learning_rate": 9.587301587301588e-05,
|
| 116 |
+
"loss": 1.3027,
|
| 117 |
+
"step": 18
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 1.19,
|
| 121 |
+
"learning_rate": 9.555555555555557e-05,
|
| 122 |
+
"loss": 1.4329,
|
| 123 |
+
"step": 19
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 1.25,
|
| 127 |
+
"learning_rate": 9.523809523809524e-05,
|
| 128 |
+
"loss": 1.1877,
|
| 129 |
+
"step": 20
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 1.31,
|
| 133 |
+
"learning_rate": 9.492063492063493e-05,
|
| 134 |
+
"loss": 1.3216,
|
| 135 |
+
"step": 21
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 1.38,
|
| 139 |
+
"learning_rate": 9.46031746031746e-05,
|
| 140 |
+
"loss": 1.1671,
|
| 141 |
+
"step": 22
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 1.44,
|
| 145 |
+
"learning_rate": 9.428571428571429e-05,
|
| 146 |
+
"loss": 1.0724,
|
| 147 |
+
"step": 23
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 1.5,
|
| 151 |
+
"learning_rate": 9.396825396825397e-05,
|
| 152 |
+
"loss": 1.148,
|
| 153 |
+
"step": 24
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 1.56,
|
| 157 |
+
"learning_rate": 9.365079365079366e-05,
|
| 158 |
+
"loss": 1.2077,
|
| 159 |
+
"step": 25
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 1.62,
|
| 163 |
+
"learning_rate": 9.333333333333334e-05,
|
| 164 |
+
"loss": 1.0365,
|
| 165 |
+
"step": 26
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 1.69,
|
| 169 |
+
"learning_rate": 9.301587301587303e-05,
|
| 170 |
+
"loss": 1.1894,
|
| 171 |
+
"step": 27
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 1.75,
|
| 175 |
+
"learning_rate": 9.26984126984127e-05,
|
| 176 |
+
"loss": 0.9611,
|
| 177 |
+
"step": 28
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 1.81,
|
| 181 |
+
"learning_rate": 9.238095238095239e-05,
|
| 182 |
+
"loss": 0.9976,
|
| 183 |
+
"step": 29
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 1.88,
|
| 187 |
+
"learning_rate": 9.206349206349206e-05,
|
| 188 |
+
"loss": 1.0812,
|
| 189 |
+
"step": 30
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.94,
|
| 193 |
+
"learning_rate": 9.174603174603175e-05,
|
| 194 |
+
"loss": 1.1406,
|
| 195 |
+
"step": 31
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 2.0,
|
| 199 |
+
"learning_rate": 9.142857142857143e-05,
|
| 200 |
+
"loss": 0.94,
|
| 201 |
+
"step": 32
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 2.06,
|
| 205 |
+
"learning_rate": 9.111111111111112e-05,
|
| 206 |
+
"loss": 0.9377,
|
| 207 |
+
"step": 33
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 2.12,
|
| 211 |
+
"learning_rate": 9.079365079365079e-05,
|
| 212 |
+
"loss": 0.9567,
|
| 213 |
+
"step": 34
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 2.19,
|
| 217 |
+
"learning_rate": 9.047619047619048e-05,
|
| 218 |
+
"loss": 0.9207,
|
| 219 |
+
"step": 35
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 2.25,
|
| 223 |
+
"learning_rate": 9.015873015873016e-05,
|
| 224 |
+
"loss": 0.7693,
|
| 225 |
+
"step": 36
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 2.31,
|
| 229 |
+
"learning_rate": 8.984126984126985e-05,
|
| 230 |
+
"loss": 0.8674,
|
| 231 |
+
"step": 37
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 2.38,
|
| 235 |
+
"learning_rate": 8.952380952380953e-05,
|
| 236 |
+
"loss": 0.7837,
|
| 237 |
+
"step": 38
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 2.44,
|
| 241 |
+
"learning_rate": 8.920634920634922e-05,
|
| 242 |
+
"loss": 0.946,
|
| 243 |
+
"step": 39
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 2.5,
|
| 247 |
+
"learning_rate": 8.888888888888889e-05,
|
| 248 |
+
"loss": 0.9637,
|
| 249 |
+
"step": 40
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 2.56,
|
| 253 |
+
"learning_rate": 8.857142857142857e-05,
|
| 254 |
+
"loss": 0.9053,
|
| 255 |
+
"step": 41
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 2.62,
|
| 259 |
+
"learning_rate": 8.825396825396825e-05,
|
| 260 |
+
"loss": 0.8995,
|
| 261 |
+
"step": 42
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 2.69,
|
| 265 |
+
"learning_rate": 8.793650793650794e-05,
|
| 266 |
+
"loss": 0.8295,
|
| 267 |
+
"step": 43
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 2.75,
|
| 271 |
+
"learning_rate": 8.761904761904762e-05,
|
| 272 |
+
"loss": 0.7863,
|
| 273 |
+
"step": 44
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 2.81,
|
| 277 |
+
"learning_rate": 8.730158730158731e-05,
|
| 278 |
+
"loss": 0.8187,
|
| 279 |
+
"step": 45
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 2.88,
|
| 283 |
+
"learning_rate": 8.6984126984127e-05,
|
| 284 |
+
"loss": 0.7932,
|
| 285 |
+
"step": 46
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 2.94,
|
| 289 |
+
"learning_rate": 8.666666666666667e-05,
|
| 290 |
+
"loss": 0.8956,
|
| 291 |
+
"step": 47
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 3.0,
|
| 295 |
+
"learning_rate": 8.634920634920635e-05,
|
| 296 |
+
"loss": 0.646,
|
| 297 |
+
"step": 48
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 3.06,
|
| 301 |
+
"learning_rate": 8.603174603174604e-05,
|
| 302 |
+
"loss": 0.602,
|
| 303 |
+
"step": 49
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 3.12,
|
| 307 |
+
"learning_rate": 8.571428571428571e-05,
|
| 308 |
+
"loss": 0.6327,
|
| 309 |
+
"step": 50
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 3.19,
|
| 313 |
+
"learning_rate": 8.53968253968254e-05,
|
| 314 |
+
"loss": 0.5247,
|
| 315 |
+
"step": 51
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 3.25,
|
| 319 |
+
"learning_rate": 8.507936507936508e-05,
|
| 320 |
+
"loss": 0.5231,
|
| 321 |
+
"step": 52
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 3.31,
|
| 325 |
+
"learning_rate": 8.476190476190477e-05,
|
| 326 |
+
"loss": 0.3324,
|
| 327 |
+
"step": 53
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 3.38,
|
| 331 |
+
"learning_rate": 8.444444444444444e-05,
|
| 332 |
+
"loss": 0.287,
|
| 333 |
+
"step": 54
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 3.44,
|
| 337 |
+
"learning_rate": 8.412698412698413e-05,
|
| 338 |
+
"loss": 0.3745,
|
| 339 |
+
"step": 55
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 3.5,
|
| 343 |
+
"learning_rate": 8.380952380952382e-05,
|
| 344 |
+
"loss": 0.4717,
|
| 345 |
+
"step": 56
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 3.56,
|
| 349 |
+
"learning_rate": 8.34920634920635e-05,
|
| 350 |
+
"loss": 0.3668,
|
| 351 |
+
"step": 57
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 3.62,
|
| 355 |
+
"learning_rate": 8.317460317460319e-05,
|
| 356 |
+
"loss": 0.4103,
|
| 357 |
+
"step": 58
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 3.69,
|
| 361 |
+
"learning_rate": 8.285714285714287e-05,
|
| 362 |
+
"loss": 0.4191,
|
| 363 |
+
"step": 59
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 3.75,
|
| 367 |
+
"learning_rate": 8.253968253968255e-05,
|
| 368 |
+
"loss": 0.3856,
|
| 369 |
+
"step": 60
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 3.81,
|
| 373 |
+
"learning_rate": 8.222222222222222e-05,
|
| 374 |
+
"loss": 0.3812,
|
| 375 |
+
"step": 61
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 3.88,
|
| 379 |
+
"learning_rate": 8.19047619047619e-05,
|
| 380 |
+
"loss": 0.4362,
|
| 381 |
+
"step": 62
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 3.94,
|
| 385 |
+
"learning_rate": 8.158730158730159e-05,
|
| 386 |
+
"loss": 0.5308,
|
| 387 |
+
"step": 63
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 4.0,
|
| 391 |
+
"learning_rate": 8.126984126984128e-05,
|
| 392 |
+
"loss": 0.4531,
|
| 393 |
+
"step": 64
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 4.06,
|
| 397 |
+
"learning_rate": 8.095238095238096e-05,
|
| 398 |
+
"loss": 0.3031,
|
| 399 |
+
"step": 65
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 4.12,
|
| 403 |
+
"learning_rate": 8.063492063492063e-05,
|
| 404 |
+
"loss": 0.2105,
|
| 405 |
+
"step": 66
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 4.19,
|
| 409 |
+
"learning_rate": 8.031746031746032e-05,
|
| 410 |
+
"loss": 0.2167,
|
| 411 |
+
"step": 67
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 4.25,
|
| 415 |
+
"learning_rate": 8e-05,
|
| 416 |
+
"loss": 0.2035,
|
| 417 |
+
"step": 68
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 4.31,
|
| 421 |
+
"learning_rate": 7.968253968253968e-05,
|
| 422 |
+
"loss": 0.1989,
|
| 423 |
+
"step": 69
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 4.38,
|
| 427 |
+
"learning_rate": 7.936507936507937e-05,
|
| 428 |
+
"loss": 0.2009,
|
| 429 |
+
"step": 70
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 4.44,
|
| 433 |
+
"learning_rate": 7.904761904761905e-05,
|
| 434 |
+
"loss": 0.2053,
|
| 435 |
+
"step": 71
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 4.5,
|
| 439 |
+
"learning_rate": 7.873015873015874e-05,
|
| 440 |
+
"loss": 0.2292,
|
| 441 |
+
"step": 72
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 4.56,
|
| 445 |
+
"learning_rate": 7.841269841269841e-05,
|
| 446 |
+
"loss": 0.1514,
|
| 447 |
+
"step": 73
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 4.62,
|
| 451 |
+
"learning_rate": 7.80952380952381e-05,
|
| 452 |
+
"loss": 0.1214,
|
| 453 |
+
"step": 74
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 4.69,
|
| 457 |
+
"learning_rate": 7.777777777777778e-05,
|
| 458 |
+
"loss": 0.1856,
|
| 459 |
+
"step": 75
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 4.75,
|
| 463 |
+
"learning_rate": 7.746031746031747e-05,
|
| 464 |
+
"loss": 0.1447,
|
| 465 |
+
"step": 76
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 4.81,
|
| 469 |
+
"learning_rate": 7.714285714285715e-05,
|
| 470 |
+
"loss": 0.2344,
|
| 471 |
+
"step": 77
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 4.88,
|
| 475 |
+
"learning_rate": 7.682539682539684e-05,
|
| 476 |
+
"loss": 0.1582,
|
| 477 |
+
"step": 78
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 4.94,
|
| 481 |
+
"learning_rate": 7.650793650793651e-05,
|
| 482 |
+
"loss": 0.1825,
|
| 483 |
+
"step": 79
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 5.0,
|
| 487 |
+
"learning_rate": 7.619047619047618e-05,
|
| 488 |
+
"loss": 0.2036,
|
| 489 |
+
"step": 80
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 5.06,
|
| 493 |
+
"learning_rate": 7.587301587301587e-05,
|
| 494 |
+
"loss": 0.1085,
|
| 495 |
+
"step": 81
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 5.12,
|
| 499 |
+
"learning_rate": 7.555555555555556e-05,
|
| 500 |
+
"loss": 0.0881,
|
| 501 |
+
"step": 82
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 5.19,
|
| 505 |
+
"learning_rate": 7.523809523809524e-05,
|
| 506 |
+
"loss": 0.1532,
|
| 507 |
+
"step": 83
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 5.25,
|
| 511 |
+
"learning_rate": 7.492063492063493e-05,
|
| 512 |
+
"loss": 0.0944,
|
| 513 |
+
"step": 84
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 5.31,
|
| 517 |
+
"learning_rate": 7.460317460317461e-05,
|
| 518 |
+
"loss": 0.1049,
|
| 519 |
+
"step": 85
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 5.38,
|
| 523 |
+
"learning_rate": 7.428571428571429e-05,
|
| 524 |
+
"loss": 0.1145,
|
| 525 |
+
"step": 86
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 5.44,
|
| 529 |
+
"learning_rate": 7.396825396825397e-05,
|
| 530 |
+
"loss": 0.1323,
|
| 531 |
+
"step": 87
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 5.5,
|
| 535 |
+
"learning_rate": 7.365079365079366e-05,
|
| 536 |
+
"loss": 0.095,
|
| 537 |
+
"step": 88
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 5.56,
|
| 541 |
+
"learning_rate": 7.333333333333333e-05,
|
| 542 |
+
"loss": 0.1388,
|
| 543 |
+
"step": 89
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 5.62,
|
| 547 |
+
"learning_rate": 7.301587301587302e-05,
|
| 548 |
+
"loss": 0.1003,
|
| 549 |
+
"step": 90
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 5.69,
|
| 553 |
+
"learning_rate": 7.26984126984127e-05,
|
| 554 |
+
"loss": 0.1175,
|
| 555 |
+
"step": 91
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 5.75,
|
| 559 |
+
"learning_rate": 7.238095238095238e-05,
|
| 560 |
+
"loss": 0.0839,
|
| 561 |
+
"step": 92
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 5.81,
|
| 565 |
+
"learning_rate": 7.206349206349206e-05,
|
| 566 |
+
"loss": 0.0986,
|
| 567 |
+
"step": 93
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 5.88,
|
| 571 |
+
"learning_rate": 7.174603174603175e-05,
|
| 572 |
+
"loss": 0.132,
|
| 573 |
+
"step": 94
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 5.94,
|
| 577 |
+
"learning_rate": 7.142857142857143e-05,
|
| 578 |
+
"loss": 0.1225,
|
| 579 |
+
"step": 95
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 6.0,
|
| 583 |
+
"learning_rate": 7.111111111111112e-05,
|
| 584 |
+
"loss": 0.1174,
|
| 585 |
+
"step": 96
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 6.06,
|
| 589 |
+
"learning_rate": 7.07936507936508e-05,
|
| 590 |
+
"loss": 0.0935,
|
| 591 |
+
"step": 97
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 6.12,
|
| 595 |
+
"learning_rate": 7.047619047619048e-05,
|
| 596 |
+
"loss": 0.0843,
|
| 597 |
+
"step": 98
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 6.19,
|
| 601 |
+
"learning_rate": 7.015873015873015e-05,
|
| 602 |
+
"loss": 0.0702,
|
| 603 |
+
"step": 99
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 6.25,
|
| 607 |
+
"learning_rate": 6.984126984126984e-05,
|
| 608 |
+
"loss": 0.0918,
|
| 609 |
+
"step": 100
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 6.31,
|
| 613 |
+
"learning_rate": 6.952380952380952e-05,
|
| 614 |
+
"loss": 0.0982,
|
| 615 |
+
"step": 101
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 6.38,
|
| 619 |
+
"learning_rate": 6.920634920634921e-05,
|
| 620 |
+
"loss": 0.1065,
|
| 621 |
+
"step": 102
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 6.44,
|
| 625 |
+
"learning_rate": 6.88888888888889e-05,
|
| 626 |
+
"loss": 0.0687,
|
| 627 |
+
"step": 103
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 6.5,
|
| 631 |
+
"learning_rate": 6.857142857142858e-05,
|
| 632 |
+
"loss": 0.0924,
|
| 633 |
+
"step": 104
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 6.56,
|
| 637 |
+
"learning_rate": 6.825396825396825e-05,
|
| 638 |
+
"loss": 0.0966,
|
| 639 |
+
"step": 105
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 6.62,
|
| 643 |
+
"learning_rate": 6.793650793650794e-05,
|
| 644 |
+
"loss": 0.0852,
|
| 645 |
+
"step": 106
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 6.69,
|
| 649 |
+
"learning_rate": 6.761904761904763e-05,
|
| 650 |
+
"loss": 0.0814,
|
| 651 |
+
"step": 107
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 6.75,
|
| 655 |
+
"learning_rate": 6.730158730158731e-05,
|
| 656 |
+
"loss": 0.0883,
|
| 657 |
+
"step": 108
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 6.81,
|
| 661 |
+
"learning_rate": 6.698412698412698e-05,
|
| 662 |
+
"loss": 0.0916,
|
| 663 |
+
"step": 109
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 6.88,
|
| 667 |
+
"learning_rate": 6.666666666666667e-05,
|
| 668 |
+
"loss": 0.1194,
|
| 669 |
+
"step": 110
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 6.94,
|
| 673 |
+
"learning_rate": 6.634920634920636e-05,
|
| 674 |
+
"loss": 0.0714,
|
| 675 |
+
"step": 111
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 7.0,
|
| 679 |
+
"learning_rate": 6.603174603174603e-05,
|
| 680 |
+
"loss": 0.0731,
|
| 681 |
+
"step": 112
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 7.06,
|
| 685 |
+
"learning_rate": 6.571428571428571e-05,
|
| 686 |
+
"loss": 0.0711,
|
| 687 |
+
"step": 113
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 7.12,
|
| 691 |
+
"learning_rate": 6.53968253968254e-05,
|
| 692 |
+
"loss": 0.048,
|
| 693 |
+
"step": 114
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 7.19,
|
| 697 |
+
"learning_rate": 6.507936507936509e-05,
|
| 698 |
+
"loss": 0.0566,
|
| 699 |
+
"step": 115
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 7.25,
|
| 703 |
+
"learning_rate": 6.476190476190477e-05,
|
| 704 |
+
"loss": 0.0635,
|
| 705 |
+
"step": 116
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 7.31,
|
| 709 |
+
"learning_rate": 6.444444444444446e-05,
|
| 710 |
+
"loss": 0.0508,
|
| 711 |
+
"step": 117
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 7.38,
|
| 715 |
+
"learning_rate": 6.412698412698413e-05,
|
| 716 |
+
"loss": 0.0587,
|
| 717 |
+
"step": 118
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 7.44,
|
| 721 |
+
"learning_rate": 6.38095238095238e-05,
|
| 722 |
+
"loss": 0.0809,
|
| 723 |
+
"step": 119
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 7.5,
|
| 727 |
+
"learning_rate": 6.349206349206349e-05,
|
| 728 |
+
"loss": 0.0664,
|
| 729 |
+
"step": 120
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 7.56,
|
| 733 |
+
"learning_rate": 6.317460317460318e-05,
|
| 734 |
+
"loss": 0.0717,
|
| 735 |
+
"step": 121
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 7.62,
|
| 739 |
+
"learning_rate": 6.285714285714286e-05,
|
| 740 |
+
"loss": 0.073,
|
| 741 |
+
"step": 122
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 7.69,
|
| 745 |
+
"learning_rate": 6.253968253968255e-05,
|
| 746 |
+
"loss": 0.0626,
|
| 747 |
+
"step": 123
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 7.75,
|
| 751 |
+
"learning_rate": 6.222222222222222e-05,
|
| 752 |
+
"loss": 0.1023,
|
| 753 |
+
"step": 124
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 7.81,
|
| 757 |
+
"learning_rate": 6.19047619047619e-05,
|
| 758 |
+
"loss": 0.0574,
|
| 759 |
+
"step": 125
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 7.88,
|
| 763 |
+
"learning_rate": 6.158730158730159e-05,
|
| 764 |
+
"loss": 0.0722,
|
| 765 |
+
"step": 126
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 7.94,
|
| 769 |
+
"learning_rate": 6.126984126984128e-05,
|
| 770 |
+
"loss": 0.0915,
|
| 771 |
+
"step": 127
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 8.0,
|
| 775 |
+
"learning_rate": 6.0952380952380964e-05,
|
| 776 |
+
"loss": 0.0778,
|
| 777 |
+
"step": 128
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 8.06,
|
| 781 |
+
"learning_rate": 6.0634920634920636e-05,
|
| 782 |
+
"loss": 0.0495,
|
| 783 |
+
"step": 129
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 8.12,
|
| 787 |
+
"learning_rate": 6.0317460317460316e-05,
|
| 788 |
+
"loss": 0.0833,
|
| 789 |
+
"step": 130
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 8.19,
|
| 793 |
+
"learning_rate": 6e-05,
|
| 794 |
+
"loss": 0.0365,
|
| 795 |
+
"step": 131
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 8.25,
|
| 799 |
+
"learning_rate": 5.968253968253969e-05,
|
| 800 |
+
"loss": 0.0563,
|
| 801 |
+
"step": 132
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 8.31,
|
| 805 |
+
"learning_rate": 5.936507936507937e-05,
|
| 806 |
+
"loss": 0.079,
|
| 807 |
+
"step": 133
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 8.38,
|
| 811 |
+
"learning_rate": 5.904761904761905e-05,
|
| 812 |
+
"loss": 0.0581,
|
| 813 |
+
"step": 134
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 8.44,
|
| 817 |
+
"learning_rate": 5.873015873015873e-05,
|
| 818 |
+
"loss": 0.05,
|
| 819 |
+
"step": 135
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 8.5,
|
| 823 |
+
"learning_rate": 5.841269841269842e-05,
|
| 824 |
+
"loss": 0.0539,
|
| 825 |
+
"step": 136
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 8.56,
|
| 829 |
+
"learning_rate": 5.8095238095238104e-05,
|
| 830 |
+
"loss": 0.073,
|
| 831 |
+
"step": 137
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 8.62,
|
| 835 |
+
"learning_rate": 5.7777777777777776e-05,
|
| 836 |
+
"loss": 0.0594,
|
| 837 |
+
"step": 138
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 8.69,
|
| 841 |
+
"learning_rate": 5.7460317460317456e-05,
|
| 842 |
+
"loss": 0.0843,
|
| 843 |
+
"step": 139
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 8.75,
|
| 847 |
+
"learning_rate": 5.714285714285714e-05,
|
| 848 |
+
"loss": 0.0694,
|
| 849 |
+
"step": 140
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 8.81,
|
| 853 |
+
"learning_rate": 5.682539682539683e-05,
|
| 854 |
+
"loss": 0.0384,
|
| 855 |
+
"step": 141
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 8.88,
|
| 859 |
+
"learning_rate": 5.650793650793651e-05,
|
| 860 |
+
"loss": 0.0473,
|
| 861 |
+
"step": 142
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 8.94,
|
| 865 |
+
"learning_rate": 5.619047619047619e-05,
|
| 866 |
+
"loss": 0.0721,
|
| 867 |
+
"step": 143
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 9.0,
|
| 871 |
+
"learning_rate": 5.587301587301588e-05,
|
| 872 |
+
"loss": 0.0604,
|
| 873 |
+
"step": 144
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 9.06,
|
| 877 |
+
"learning_rate": 5.555555555555556e-05,
|
| 878 |
+
"loss": 0.0413,
|
| 879 |
+
"step": 145
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 9.12,
|
| 883 |
+
"learning_rate": 5.5238095238095244e-05,
|
| 884 |
+
"loss": 0.0695,
|
| 885 |
+
"step": 146
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 9.19,
|
| 889 |
+
"learning_rate": 5.492063492063493e-05,
|
| 890 |
+
"loss": 0.0511,
|
| 891 |
+
"step": 147
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 9.25,
|
| 895 |
+
"learning_rate": 5.46031746031746e-05,
|
| 896 |
+
"loss": 0.041,
|
| 897 |
+
"step": 148
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 9.31,
|
| 901 |
+
"learning_rate": 5.428571428571428e-05,
|
| 902 |
+
"loss": 0.0432,
|
| 903 |
+
"step": 149
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 9.38,
|
| 907 |
+
"learning_rate": 5.396825396825397e-05,
|
| 908 |
+
"loss": 0.0419,
|
| 909 |
+
"step": 150
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 9.44,
|
| 913 |
+
"learning_rate": 5.3650793650793654e-05,
|
| 914 |
+
"loss": 0.0483,
|
| 915 |
+
"step": 151
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 9.5,
|
| 919 |
+
"learning_rate": 5.333333333333333e-05,
|
| 920 |
+
"loss": 0.0418,
|
| 921 |
+
"step": 152
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 9.56,
|
| 925 |
+
"learning_rate": 5.301587301587302e-05,
|
| 926 |
+
"loss": 0.0538,
|
| 927 |
+
"step": 153
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 9.62,
|
| 931 |
+
"learning_rate": 5.2698412698412705e-05,
|
| 932 |
+
"loss": 0.0672,
|
| 933 |
+
"step": 154
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 9.69,
|
| 937 |
+
"learning_rate": 5.2380952380952384e-05,
|
| 938 |
+
"loss": 0.0441,
|
| 939 |
+
"step": 155
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 9.75,
|
| 943 |
+
"learning_rate": 5.206349206349207e-05,
|
| 944 |
+
"loss": 0.0504,
|
| 945 |
+
"step": 156
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 9.81,
|
| 949 |
+
"learning_rate": 5.1746031746031756e-05,
|
| 950 |
+
"loss": 0.0423,
|
| 951 |
+
"step": 157
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 9.88,
|
| 955 |
+
"learning_rate": 5.142857142857143e-05,
|
| 956 |
+
"loss": 0.1098,
|
| 957 |
+
"step": 158
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 9.94,
|
| 961 |
+
"learning_rate": 5.111111111111111e-05,
|
| 962 |
+
"loss": 0.0623,
|
| 963 |
+
"step": 159
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 10.0,
|
| 967 |
+
"learning_rate": 5.0793650793650794e-05,
|
| 968 |
+
"loss": 0.0499,
|
| 969 |
+
"step": 160
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 10.06,
|
| 973 |
+
"learning_rate": 5.047619047619048e-05,
|
| 974 |
+
"loss": 0.0472,
|
| 975 |
+
"step": 161
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 10.12,
|
| 979 |
+
"learning_rate": 5.015873015873016e-05,
|
| 980 |
+
"loss": 0.0422,
|
| 981 |
+
"step": 162
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 10.19,
|
| 985 |
+
"learning_rate": 4.9841269841269845e-05,
|
| 986 |
+
"loss": 0.0383,
|
| 987 |
+
"step": 163
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 10.25,
|
| 991 |
+
"learning_rate": 4.9523809523809525e-05,
|
| 992 |
+
"loss": 0.0509,
|
| 993 |
+
"step": 164
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 10.31,
|
| 997 |
+
"learning_rate": 4.9206349206349204e-05,
|
| 998 |
+
"loss": 0.0473,
|
| 999 |
+
"step": 165
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 10.38,
|
| 1003 |
+
"learning_rate": 4.888888888888889e-05,
|
| 1004 |
+
"loss": 0.0486,
|
| 1005 |
+
"step": 166
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 10.44,
|
| 1009 |
+
"learning_rate": 4.8571428571428576e-05,
|
| 1010 |
+
"loss": 0.0557,
|
| 1011 |
+
"step": 167
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 10.5,
|
| 1015 |
+
"learning_rate": 4.8253968253968255e-05,
|
| 1016 |
+
"loss": 0.0484,
|
| 1017 |
+
"step": 168
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 10.56,
|
| 1021 |
+
"learning_rate": 4.793650793650794e-05,
|
| 1022 |
+
"loss": 0.038,
|
| 1023 |
+
"step": 169
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 10.62,
|
| 1027 |
+
"learning_rate": 4.761904761904762e-05,
|
| 1028 |
+
"loss": 0.0443,
|
| 1029 |
+
"step": 170
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 10.69,
|
| 1033 |
+
"learning_rate": 4.73015873015873e-05,
|
| 1034 |
+
"loss": 0.04,
|
| 1035 |
+
"step": 171
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 10.75,
|
| 1039 |
+
"learning_rate": 4.6984126984126986e-05,
|
| 1040 |
+
"loss": 0.0626,
|
| 1041 |
+
"step": 172
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 10.81,
|
| 1045 |
+
"learning_rate": 4.666666666666667e-05,
|
| 1046 |
+
"loss": 0.0474,
|
| 1047 |
+
"step": 173
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 10.88,
|
| 1051 |
+
"learning_rate": 4.634920634920635e-05,
|
| 1052 |
+
"loss": 0.076,
|
| 1053 |
+
"step": 174
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 10.94,
|
| 1057 |
+
"learning_rate": 4.603174603174603e-05,
|
| 1058 |
+
"loss": 0.0391,
|
| 1059 |
+
"step": 175
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 11.0,
|
| 1063 |
+
"learning_rate": 4.5714285714285716e-05,
|
| 1064 |
+
"loss": 0.0483,
|
| 1065 |
+
"step": 176
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 11.06,
|
| 1069 |
+
"learning_rate": 4.5396825396825395e-05,
|
| 1070 |
+
"loss": 0.0516,
|
| 1071 |
+
"step": 177
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 11.12,
|
| 1075 |
+
"learning_rate": 4.507936507936508e-05,
|
| 1076 |
+
"loss": 0.043,
|
| 1077 |
+
"step": 178
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 11.19,
|
| 1081 |
+
"learning_rate": 4.476190476190477e-05,
|
| 1082 |
+
"loss": 0.0537,
|
| 1083 |
+
"step": 179
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 11.25,
|
| 1087 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 1088 |
+
"loss": 0.0629,
|
| 1089 |
+
"step": 180
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 11.31,
|
| 1093 |
+
"learning_rate": 4.4126984126984126e-05,
|
| 1094 |
+
"loss": 0.0626,
|
| 1095 |
+
"step": 181
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 11.38,
|
| 1099 |
+
"learning_rate": 4.380952380952381e-05,
|
| 1100 |
+
"loss": 0.0295,
|
| 1101 |
+
"step": 182
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 11.44,
|
| 1105 |
+
"learning_rate": 4.34920634920635e-05,
|
| 1106 |
+
"loss": 0.0398,
|
| 1107 |
+
"step": 183
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 11.5,
|
| 1111 |
+
"learning_rate": 4.317460317460318e-05,
|
| 1112 |
+
"loss": 0.0386,
|
| 1113 |
+
"step": 184
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 11.56,
|
| 1117 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 1118 |
+
"loss": 0.0548,
|
| 1119 |
+
"step": 185
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 11.62,
|
| 1123 |
+
"learning_rate": 4.253968253968254e-05,
|
| 1124 |
+
"loss": 0.0427,
|
| 1125 |
+
"step": 186
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 11.69,
|
| 1129 |
+
"learning_rate": 4.222222222222222e-05,
|
| 1130 |
+
"loss": 0.0478,
|
| 1131 |
+
"step": 187
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 11.75,
|
| 1135 |
+
"learning_rate": 4.190476190476191e-05,
|
| 1136 |
+
"loss": 0.0326,
|
| 1137 |
+
"step": 188
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 11.81,
|
| 1141 |
+
"learning_rate": 4.1587301587301594e-05,
|
| 1142 |
+
"loss": 0.0614,
|
| 1143 |
+
"step": 189
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 11.88,
|
| 1147 |
+
"learning_rate": 4.126984126984127e-05,
|
| 1148 |
+
"loss": 0.0685,
|
| 1149 |
+
"step": 190
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 11.94,
|
| 1153 |
+
"learning_rate": 4.095238095238095e-05,
|
| 1154 |
+
"loss": 0.0332,
|
| 1155 |
+
"step": 191
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 12.0,
|
| 1159 |
+
"learning_rate": 4.063492063492064e-05,
|
| 1160 |
+
"loss": 0.0407,
|
| 1161 |
+
"step": 192
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 12.06,
|
| 1165 |
+
"learning_rate": 4.031746031746032e-05,
|
| 1166 |
+
"loss": 0.0373,
|
| 1167 |
+
"step": 193
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 12.12,
|
| 1171 |
+
"learning_rate": 4e-05,
|
| 1172 |
+
"loss": 0.044,
|
| 1173 |
+
"step": 194
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 12.19,
|
| 1177 |
+
"learning_rate": 3.968253968253968e-05,
|
| 1178 |
+
"loss": 0.041,
|
| 1179 |
+
"step": 195
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 12.25,
|
| 1183 |
+
"learning_rate": 3.936507936507937e-05,
|
| 1184 |
+
"loss": 0.0313,
|
| 1185 |
+
"step": 196
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 12.31,
|
| 1189 |
+
"learning_rate": 3.904761904761905e-05,
|
| 1190 |
+
"loss": 0.0442,
|
| 1191 |
+
"step": 197
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 12.38,
|
| 1195 |
+
"learning_rate": 3.8730158730158734e-05,
|
| 1196 |
+
"loss": 0.0487,
|
| 1197 |
+
"step": 198
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 12.44,
|
| 1201 |
+
"learning_rate": 3.841269841269842e-05,
|
| 1202 |
+
"loss": 0.0267,
|
| 1203 |
+
"step": 199
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 12.5,
|
| 1207 |
+
"learning_rate": 3.809523809523809e-05,
|
| 1208 |
+
"loss": 0.0461,
|
| 1209 |
+
"step": 200
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 12.56,
|
| 1213 |
+
"learning_rate": 3.777777777777778e-05,
|
| 1214 |
+
"loss": 0.0459,
|
| 1215 |
+
"step": 201
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 12.62,
|
| 1219 |
+
"learning_rate": 3.7460317460317464e-05,
|
| 1220 |
+
"loss": 0.0386,
|
| 1221 |
+
"step": 202
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 12.69,
|
| 1225 |
+
"learning_rate": 3.7142857142857143e-05,
|
| 1226 |
+
"loss": 0.0512,
|
| 1227 |
+
"step": 203
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 12.75,
|
| 1231 |
+
"learning_rate": 3.682539682539683e-05,
|
| 1232 |
+
"loss": 0.0416,
|
| 1233 |
+
"step": 204
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 12.81,
|
| 1237 |
+
"learning_rate": 3.650793650793651e-05,
|
| 1238 |
+
"loss": 0.0505,
|
| 1239 |
+
"step": 205
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 12.88,
|
| 1243 |
+
"learning_rate": 3.619047619047619e-05,
|
| 1244 |
+
"loss": 0.0422,
|
| 1245 |
+
"step": 206
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 12.94,
|
| 1249 |
+
"learning_rate": 3.5873015873015874e-05,
|
| 1250 |
+
"loss": 0.0374,
|
| 1251 |
+
"step": 207
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 13.0,
|
| 1255 |
+
"learning_rate": 3.555555555555556e-05,
|
| 1256 |
+
"loss": 0.0518,
|
| 1257 |
+
"step": 208
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 13.06,
|
| 1261 |
+
"learning_rate": 3.523809523809524e-05,
|
| 1262 |
+
"loss": 0.0418,
|
| 1263 |
+
"step": 209
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 13.12,
|
| 1267 |
+
"learning_rate": 3.492063492063492e-05,
|
| 1268 |
+
"loss": 0.038,
|
| 1269 |
+
"step": 210
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 13.19,
|
| 1273 |
+
"learning_rate": 3.4603174603174604e-05,
|
| 1274 |
+
"loss": 0.0456,
|
| 1275 |
+
"step": 211
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 13.25,
|
| 1279 |
+
"learning_rate": 3.428571428571429e-05,
|
| 1280 |
+
"loss": 0.0465,
|
| 1281 |
+
"step": 212
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 13.31,
|
| 1285 |
+
"learning_rate": 3.396825396825397e-05,
|
| 1286 |
+
"loss": 0.0329,
|
| 1287 |
+
"step": 213
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 13.38,
|
| 1291 |
+
"learning_rate": 3.3650793650793656e-05,
|
| 1292 |
+
"loss": 0.0734,
|
| 1293 |
+
"step": 214
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 13.44,
|
| 1297 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 1298 |
+
"loss": 0.0298,
|
| 1299 |
+
"step": 215
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 13.5,
|
| 1303 |
+
"learning_rate": 3.3015873015873014e-05,
|
| 1304 |
+
"loss": 0.0321,
|
| 1305 |
+
"step": 216
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 13.56,
|
| 1309 |
+
"learning_rate": 3.26984126984127e-05,
|
| 1310 |
+
"loss": 0.0361,
|
| 1311 |
+
"step": 217
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 13.62,
|
| 1315 |
+
"learning_rate": 3.2380952380952386e-05,
|
| 1316 |
+
"loss": 0.0334,
|
| 1317 |
+
"step": 218
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 13.69,
|
| 1321 |
+
"learning_rate": 3.2063492063492065e-05,
|
| 1322 |
+
"loss": 0.0495,
|
| 1323 |
+
"step": 219
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 13.75,
|
| 1327 |
+
"learning_rate": 3.1746031746031745e-05,
|
| 1328 |
+
"loss": 0.0491,
|
| 1329 |
+
"step": 220
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 13.81,
|
| 1333 |
+
"learning_rate": 3.142857142857143e-05,
|
| 1334 |
+
"loss": 0.0348,
|
| 1335 |
+
"step": 221
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 13.88,
|
| 1339 |
+
"learning_rate": 3.111111111111111e-05,
|
| 1340 |
+
"loss": 0.0547,
|
| 1341 |
+
"step": 222
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 13.94,
|
| 1345 |
+
"learning_rate": 3.0793650793650796e-05,
|
| 1346 |
+
"loss": 0.0476,
|
| 1347 |
+
"step": 223
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 14.0,
|
| 1351 |
+
"learning_rate": 3.0476190476190482e-05,
|
| 1352 |
+
"loss": 0.0452,
|
| 1353 |
+
"step": 224
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 14.06,
|
| 1357 |
+
"learning_rate": 3.0158730158730158e-05,
|
| 1358 |
+
"loss": 0.0356,
|
| 1359 |
+
"step": 225
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 14.12,
|
| 1363 |
+
"learning_rate": 2.9841269841269844e-05,
|
| 1364 |
+
"loss": 0.032,
|
| 1365 |
+
"step": 226
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 14.19,
|
| 1369 |
+
"learning_rate": 2.9523809523809526e-05,
|
| 1370 |
+
"loss": 0.0379,
|
| 1371 |
+
"step": 227
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 14.25,
|
| 1375 |
+
"learning_rate": 2.920634920634921e-05,
|
| 1376 |
+
"loss": 0.0371,
|
| 1377 |
+
"step": 228
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 14.31,
|
| 1381 |
+
"learning_rate": 2.8888888888888888e-05,
|
| 1382 |
+
"loss": 0.033,
|
| 1383 |
+
"step": 229
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 14.38,
|
| 1387 |
+
"learning_rate": 2.857142857142857e-05,
|
| 1388 |
+
"loss": 0.0368,
|
| 1389 |
+
"step": 230
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 14.44,
|
| 1393 |
+
"learning_rate": 2.8253968253968253e-05,
|
| 1394 |
+
"loss": 0.0413,
|
| 1395 |
+
"step": 231
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 14.5,
|
| 1399 |
+
"learning_rate": 2.793650793650794e-05,
|
| 1400 |
+
"loss": 0.0316,
|
| 1401 |
+
"step": 232
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 14.56,
|
| 1405 |
+
"learning_rate": 2.7619047619047622e-05,
|
| 1406 |
+
"loss": 0.0365,
|
| 1407 |
+
"step": 233
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 14.62,
|
| 1411 |
+
"learning_rate": 2.73015873015873e-05,
|
| 1412 |
+
"loss": 0.0407,
|
| 1413 |
+
"step": 234
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 14.69,
|
| 1417 |
+
"learning_rate": 2.6984126984126984e-05,
|
| 1418 |
+
"loss": 0.0377,
|
| 1419 |
+
"step": 235
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 14.75,
|
| 1423 |
+
"learning_rate": 2.6666666666666667e-05,
|
| 1424 |
+
"loss": 0.037,
|
| 1425 |
+
"step": 236
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 14.81,
|
| 1429 |
+
"learning_rate": 2.6349206349206353e-05,
|
| 1430 |
+
"loss": 0.0366,
|
| 1431 |
+
"step": 237
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 14.88,
|
| 1435 |
+
"learning_rate": 2.6031746031746035e-05,
|
| 1436 |
+
"loss": 0.0517,
|
| 1437 |
+
"step": 238
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 14.94,
|
| 1441 |
+
"learning_rate": 2.5714285714285714e-05,
|
| 1442 |
+
"loss": 0.0553,
|
| 1443 |
+
"step": 239
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 15.0,
|
| 1447 |
+
"learning_rate": 2.5396825396825397e-05,
|
| 1448 |
+
"loss": 0.0557,
|
| 1449 |
+
"step": 240
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 15.06,
|
| 1453 |
+
"learning_rate": 2.507936507936508e-05,
|
| 1454 |
+
"loss": 0.0314,
|
| 1455 |
+
"step": 241
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 15.12,
|
| 1459 |
+
"learning_rate": 2.4761904761904762e-05,
|
| 1460 |
+
"loss": 0.0343,
|
| 1461 |
+
"step": 242
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 15.19,
|
| 1465 |
+
"learning_rate": 2.4444444444444445e-05,
|
| 1466 |
+
"loss": 0.0274,
|
| 1467 |
+
"step": 243
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 15.25,
|
| 1471 |
+
"learning_rate": 2.4126984126984128e-05,
|
| 1472 |
+
"loss": 0.0409,
|
| 1473 |
+
"step": 244
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 15.31,
|
| 1477 |
+
"learning_rate": 2.380952380952381e-05,
|
| 1478 |
+
"loss": 0.0409,
|
| 1479 |
+
"step": 245
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 15.38,
|
| 1483 |
+
"learning_rate": 2.3492063492063493e-05,
|
| 1484 |
+
"loss": 0.0316,
|
| 1485 |
+
"step": 246
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 15.44,
|
| 1489 |
+
"learning_rate": 2.3174603174603175e-05,
|
| 1490 |
+
"loss": 0.0514,
|
| 1491 |
+
"step": 247
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 15.5,
|
| 1495 |
+
"learning_rate": 2.2857142857142858e-05,
|
| 1496 |
+
"loss": 0.0413,
|
| 1497 |
+
"step": 248
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 15.56,
|
| 1501 |
+
"learning_rate": 2.253968253968254e-05,
|
| 1502 |
+
"loss": 0.0423,
|
| 1503 |
+
"step": 249
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 15.62,
|
| 1507 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 1508 |
+
"loss": 0.0387,
|
| 1509 |
+
"step": 250
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 15.69,
|
| 1513 |
+
"learning_rate": 2.1904761904761906e-05,
|
| 1514 |
+
"loss": 0.0399,
|
| 1515 |
+
"step": 251
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 15.75,
|
| 1519 |
+
"learning_rate": 2.158730158730159e-05,
|
| 1520 |
+
"loss": 0.0294,
|
| 1521 |
+
"step": 252
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 15.81,
|
| 1525 |
+
"learning_rate": 2.126984126984127e-05,
|
| 1526 |
+
"loss": 0.056,
|
| 1527 |
+
"step": 253
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 15.88,
|
| 1531 |
+
"learning_rate": 2.0952380952380954e-05,
|
| 1532 |
+
"loss": 0.0473,
|
| 1533 |
+
"step": 254
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 15.94,
|
| 1537 |
+
"learning_rate": 2.0634920634920636e-05,
|
| 1538 |
+
"loss": 0.036,
|
| 1539 |
+
"step": 255
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 16.0,
|
| 1543 |
+
"learning_rate": 2.031746031746032e-05,
|
| 1544 |
+
"loss": 0.0387,
|
| 1545 |
+
"step": 256
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 16.06,
|
| 1549 |
+
"learning_rate": 2e-05,
|
| 1550 |
+
"loss": 0.0308,
|
| 1551 |
+
"step": 257
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 16.12,
|
| 1555 |
+
"learning_rate": 1.9682539682539684e-05,
|
| 1556 |
+
"loss": 0.0344,
|
| 1557 |
+
"step": 258
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 16.19,
|
| 1561 |
+
"learning_rate": 1.9365079365079367e-05,
|
| 1562 |
+
"loss": 0.0401,
|
| 1563 |
+
"step": 259
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 16.25,
|
| 1567 |
+
"learning_rate": 1.9047619047619046e-05,
|
| 1568 |
+
"loss": 0.0473,
|
| 1569 |
+
"step": 260
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 16.31,
|
| 1573 |
+
"learning_rate": 1.8730158730158732e-05,
|
| 1574 |
+
"loss": 0.0296,
|
| 1575 |
+
"step": 261
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 16.38,
|
| 1579 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 1580 |
+
"loss": 0.0331,
|
| 1581 |
+
"step": 262
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 16.44,
|
| 1585 |
+
"learning_rate": 1.8095238095238094e-05,
|
| 1586 |
+
"loss": 0.0347,
|
| 1587 |
+
"step": 263
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 16.5,
|
| 1591 |
+
"learning_rate": 1.777777777777778e-05,
|
| 1592 |
+
"loss": 0.036,
|
| 1593 |
+
"step": 264
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 16.56,
|
| 1597 |
+
"learning_rate": 1.746031746031746e-05,
|
| 1598 |
+
"loss": 0.0507,
|
| 1599 |
+
"step": 265
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 16.62,
|
| 1603 |
+
"learning_rate": 1.7142857142857145e-05,
|
| 1604 |
+
"loss": 0.0368,
|
| 1605 |
+
"step": 266
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 16.69,
|
| 1609 |
+
"learning_rate": 1.6825396825396828e-05,
|
| 1610 |
+
"loss": 0.0448,
|
| 1611 |
+
"step": 267
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 16.75,
|
| 1615 |
+
"learning_rate": 1.6507936507936507e-05,
|
| 1616 |
+
"loss": 0.0451,
|
| 1617 |
+
"step": 268
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 16.81,
|
| 1621 |
+
"learning_rate": 1.6190476190476193e-05,
|
| 1622 |
+
"loss": 0.0355,
|
| 1623 |
+
"step": 269
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 16.88,
|
| 1627 |
+
"learning_rate": 1.5873015873015872e-05,
|
| 1628 |
+
"loss": 0.0269,
|
| 1629 |
+
"step": 270
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 16.94,
|
| 1633 |
+
"learning_rate": 1.5555555555555555e-05,
|
| 1634 |
+
"loss": 0.0572,
|
| 1635 |
+
"step": 271
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 17.0,
|
| 1639 |
+
"learning_rate": 1.5238095238095241e-05,
|
| 1640 |
+
"loss": 0.0426,
|
| 1641 |
+
"step": 272
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 17.06,
|
| 1645 |
+
"learning_rate": 1.4920634920634922e-05,
|
| 1646 |
+
"loss": 0.0356,
|
| 1647 |
+
"step": 273
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 17.12,
|
| 1651 |
+
"learning_rate": 1.4603174603174605e-05,
|
| 1652 |
+
"loss": 0.0246,
|
| 1653 |
+
"step": 274
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 17.19,
|
| 1657 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 1658 |
+
"loss": 0.0361,
|
| 1659 |
+
"step": 275
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 17.25,
|
| 1663 |
+
"learning_rate": 1.396825396825397e-05,
|
| 1664 |
+
"loss": 0.0358,
|
| 1665 |
+
"step": 276
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 17.31,
|
| 1669 |
+
"learning_rate": 1.365079365079365e-05,
|
| 1670 |
+
"loss": 0.0434,
|
| 1671 |
+
"step": 277
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 17.38,
|
| 1675 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 1676 |
+
"loss": 0.0338,
|
| 1677 |
+
"step": 278
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 17.44,
|
| 1681 |
+
"learning_rate": 1.3015873015873018e-05,
|
| 1682 |
+
"loss": 0.0411,
|
| 1683 |
+
"step": 279
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 17.5,
|
| 1687 |
+
"learning_rate": 1.2698412698412699e-05,
|
| 1688 |
+
"loss": 0.0282,
|
| 1689 |
+
"step": 280
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 17.56,
|
| 1693 |
+
"learning_rate": 1.2380952380952381e-05,
|
| 1694 |
+
"loss": 0.0276,
|
| 1695 |
+
"step": 281
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 17.62,
|
| 1699 |
+
"learning_rate": 1.2063492063492064e-05,
|
| 1700 |
+
"loss": 0.0341,
|
| 1701 |
+
"step": 282
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 17.69,
|
| 1705 |
+
"learning_rate": 1.1746031746031746e-05,
|
| 1706 |
+
"loss": 0.0343,
|
| 1707 |
+
"step": 283
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 17.75,
|
| 1711 |
+
"learning_rate": 1.1428571428571429e-05,
|
| 1712 |
+
"loss": 0.0507,
|
| 1713 |
+
"step": 284
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 17.81,
|
| 1717 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 1718 |
+
"loss": 0.0556,
|
| 1719 |
+
"step": 285
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 17.88,
|
| 1723 |
+
"learning_rate": 1.0793650793650794e-05,
|
| 1724 |
+
"loss": 0.049,
|
| 1725 |
+
"step": 286
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 17.94,
|
| 1729 |
+
"learning_rate": 1.0476190476190477e-05,
|
| 1730 |
+
"loss": 0.0456,
|
| 1731 |
+
"step": 287
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 18.0,
|
| 1735 |
+
"learning_rate": 1.015873015873016e-05,
|
| 1736 |
+
"loss": 0.0447,
|
| 1737 |
+
"step": 288
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 18.06,
|
| 1741 |
+
"learning_rate": 9.841269841269842e-06,
|
| 1742 |
+
"loss": 0.0294,
|
| 1743 |
+
"step": 289
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 18.12,
|
| 1747 |
+
"learning_rate": 9.523809523809523e-06,
|
| 1748 |
+
"loss": 0.0337,
|
| 1749 |
+
"step": 290
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 18.19,
|
| 1753 |
+
"learning_rate": 9.206349206349207e-06,
|
| 1754 |
+
"loss": 0.0397,
|
| 1755 |
+
"step": 291
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 18.25,
|
| 1759 |
+
"learning_rate": 8.88888888888889e-06,
|
| 1760 |
+
"loss": 0.0349,
|
| 1761 |
+
"step": 292
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 18.31,
|
| 1765 |
+
"learning_rate": 8.571428571428573e-06,
|
| 1766 |
+
"loss": 0.0537,
|
| 1767 |
+
"step": 293
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 18.38,
|
| 1771 |
+
"learning_rate": 8.253968253968254e-06,
|
| 1772 |
+
"loss": 0.0335,
|
| 1773 |
+
"step": 294
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 18.44,
|
| 1777 |
+
"learning_rate": 7.936507936507936e-06,
|
| 1778 |
+
"loss": 0.0399,
|
| 1779 |
+
"step": 295
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 18.5,
|
| 1783 |
+
"learning_rate": 7.6190476190476205e-06,
|
| 1784 |
+
"loss": 0.0381,
|
| 1785 |
+
"step": 296
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 18.56,
|
| 1789 |
+
"learning_rate": 7.301587301587302e-06,
|
| 1790 |
+
"loss": 0.031,
|
| 1791 |
+
"step": 297
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 18.62,
|
| 1795 |
+
"learning_rate": 6.984126984126985e-06,
|
| 1796 |
+
"loss": 0.034,
|
| 1797 |
+
"step": 298
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 18.69,
|
| 1801 |
+
"learning_rate": 6.666666666666667e-06,
|
| 1802 |
+
"loss": 0.0373,
|
| 1803 |
+
"step": 299
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 18.75,
|
| 1807 |
+
"learning_rate": 6.349206349206349e-06,
|
| 1808 |
+
"loss": 0.0473,
|
| 1809 |
+
"step": 300
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 18.81,
|
| 1813 |
+
"learning_rate": 6.031746031746032e-06,
|
| 1814 |
+
"loss": 0.0331,
|
| 1815 |
+
"step": 301
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 18.88,
|
| 1819 |
+
"learning_rate": 5.7142857142857145e-06,
|
| 1820 |
+
"loss": 0.0402,
|
| 1821 |
+
"step": 302
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 18.94,
|
| 1825 |
+
"learning_rate": 5.396825396825397e-06,
|
| 1826 |
+
"loss": 0.03,
|
| 1827 |
+
"step": 303
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 19.0,
|
| 1831 |
+
"learning_rate": 5.07936507936508e-06,
|
| 1832 |
+
"loss": 0.0408,
|
| 1833 |
+
"step": 304
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 19.06,
|
| 1837 |
+
"learning_rate": 4.7619047619047615e-06,
|
| 1838 |
+
"loss": 0.0406,
|
| 1839 |
+
"step": 305
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 19.12,
|
| 1843 |
+
"learning_rate": 4.444444444444445e-06,
|
| 1844 |
+
"loss": 0.032,
|
| 1845 |
+
"step": 306
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 19.19,
|
| 1849 |
+
"learning_rate": 4.126984126984127e-06,
|
| 1850 |
+
"loss": 0.0371,
|
| 1851 |
+
"step": 307
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 19.25,
|
| 1855 |
+
"learning_rate": 3.8095238095238102e-06,
|
| 1856 |
+
"loss": 0.0419,
|
| 1857 |
+
"step": 308
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 19.31,
|
| 1861 |
+
"learning_rate": 3.4920634920634924e-06,
|
| 1862 |
+
"loss": 0.0281,
|
| 1863 |
+
"step": 309
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 19.38,
|
| 1867 |
+
"learning_rate": 3.1746031746031746e-06,
|
| 1868 |
+
"loss": 0.0402,
|
| 1869 |
+
"step": 310
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 19.44,
|
| 1873 |
+
"learning_rate": 2.8571428571428573e-06,
|
| 1874 |
+
"loss": 0.0422,
|
| 1875 |
+
"step": 311
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 19.5,
|
| 1879 |
+
"learning_rate": 2.53968253968254e-06,
|
| 1880 |
+
"loss": 0.0348,
|
| 1881 |
+
"step": 312
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 19.56,
|
| 1885 |
+
"learning_rate": 2.2222222222222225e-06,
|
| 1886 |
+
"loss": 0.0302,
|
| 1887 |
+
"step": 313
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 19.62,
|
| 1891 |
+
"learning_rate": 1.9047619047619051e-06,
|
| 1892 |
+
"loss": 0.0539,
|
| 1893 |
+
"step": 314
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 19.69,
|
| 1897 |
+
"learning_rate": 1.5873015873015873e-06,
|
| 1898 |
+
"loss": 0.0294,
|
| 1899 |
+
"step": 315
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 19.75,
|
| 1903 |
+
"learning_rate": 1.26984126984127e-06,
|
| 1904 |
+
"loss": 0.0446,
|
| 1905 |
+
"step": 316
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 19.81,
|
| 1909 |
+
"learning_rate": 9.523809523809526e-07,
|
| 1910 |
+
"loss": 0.0281,
|
| 1911 |
+
"step": 317
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 19.88,
|
| 1915 |
+
"learning_rate": 6.34920634920635e-07,
|
| 1916 |
+
"loss": 0.0406,
|
| 1917 |
+
"step": 318
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 19.94,
|
| 1921 |
+
"learning_rate": 3.174603174603175e-07,
|
| 1922 |
+
"loss": 0.0353,
|
| 1923 |
+
"step": 319
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 20.0,
|
| 1927 |
+
"learning_rate": 0.0,
|
| 1928 |
+
"loss": 0.0348,
|
| 1929 |
+
"step": 320
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 20.0,
|
| 1933 |
+
"step": 320,
|
| 1934 |
+
"total_flos": 1.63364550475776e+16,
|
| 1935 |
+
"train_loss": 0.26344227014342325,
|
| 1936 |
+
"train_runtime": 536.4097,
|
| 1937 |
+
"train_samples_per_second": 2.386,
|
| 1938 |
+
"train_steps_per_second": 0.597
|
| 1939 |
+
}
|
| 1940 |
+
],
|
| 1941 |
+
"logging_steps": 1,
|
| 1942 |
+
"max_steps": 320,
|
| 1943 |
+
"num_train_epochs": 20,
|
| 1944 |
+
"save_steps": 500,
|
| 1945 |
+
"total_flos": 1.63364550475776e+16,
|
| 1946 |
+
"trial_name": null,
|
| 1947 |
+
"trial_params": null
|
| 1948 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cf16b6ac265d8304ccac61f4a46372f6b16a6546ad521b72a027f09d643f6d2b
|
| 3 |
+
size 4600
|