ANTOUN Wissam
commited on
Commit
•
79ed695
0
Parent(s):
added files
Browse files- .gitattributes +34 -0
- README.md +83 -0
- all_results.json +22 -0
- config.json +38 -0
- eval_results.json +16 -0
- predict_results_full_multi_1.0_0.5_0.5.txt +0 -0
- pytorch_model.bin +3 -0
- sentencepiece.bpe.model +3 -0
- special_tokens_map.json +15 -0
- tokenizer.json +0 -0
- tokenizer_config.json +20 -0
- train_results.json +8 -0
- trainer_state.json +2626 -0
- training_args.bin +3 -0
.gitattributes
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 8 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 9 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 10 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 11 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 12 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
| 13 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 14 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 15 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 16 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 17 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
| 18 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
| 19 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
| 20 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
| 21 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
| 22 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
| 23 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
| 24 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
| 25 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
| 26 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
| 27 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
| 28 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
| 29 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
| 30 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
| 31 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
| 32 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 33 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 34 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,83 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en,fr
|
| 4 |
+
license: mit
|
| 5 |
+
tags:
|
| 6 |
+
- lm-detection
|
| 7 |
+
datasets:
|
| 8 |
+
- hc3_multi_custom_ms_hg
|
| 9 |
+
metrics:
|
| 10 |
+
- f1
|
| 11 |
+
model-index:
|
| 12 |
+
- name: xlmr-chatgptdetect-noisy
|
| 13 |
+
results:
|
| 14 |
+
- task:
|
| 15 |
+
name: Text Classification
|
| 16 |
+
type: text-classification
|
| 17 |
+
dataset:
|
| 18 |
+
name: HC3 FULL_MULTI_1.0_0.5_0.5
|
| 19 |
+
type: glue
|
| 20 |
+
config: full_multi_1.0_0.5_0.5
|
| 21 |
+
split: vsl
|
| 22 |
+
args: full_multi_1.0_0.5_0.5
|
| 23 |
+
metrics:
|
| 24 |
+
- name: F1
|
| 25 |
+
type: f1
|
| 26 |
+
value: 0.963274059512108
|
| 27 |
+
---
|
| 28 |
+
|
| 29 |
+
# xlmr-chatgptdetect-noisy
|
| 30 |
+
|
| 31 |
+
Multilingual ChatGPT detection model from [Towards a Robust Detection of Language Model-Generated Text: Is ChatGPT that easy to detect?](TODO:)
|
| 32 |
+
|
| 33 |
+
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the HC3 FULL_MULTI_1.0_0.5_0.5 dataset with noise added.
|
| 34 |
+
It achieves the following results on the evaluation set:
|
| 35 |
+
- Loss: 0.1573
|
| 36 |
+
- F1: 0.9633
|
| 37 |
+
|
| 38 |
+
## Model description
|
| 39 |
+
|
| 40 |
+
This a model trained to detect text created by ChatGPT in French.
|
| 41 |
+
The training data is the combination of the `hc3_fr_full` and `hc3_en_full` subsets of [almanach/hc3_multi](https://huggingface.co/almanach/hc3_multi), but with added misspelling and homoglyph attacks.
|
| 42 |
+
|
| 43 |
+
## Intended uses & limitations
|
| 44 |
+
|
| 45 |
+
This model is for research purposes only.
|
| 46 |
+
It is not intended to be used in production as we said in our paper:
|
| 47 |
+
|
| 48 |
+
**We would like to emphasize that our study does not claim to have produced an universally accurate detector. Our strong results are based on in-domain testing and, unsurprisingly, do not generalize in out-of-domain scenarios. This is even more so when used on text specifically designed to fool language model detectors and on text intentionally stylistically similar to ChatGPT-generated text, especially instructional text.**
|
| 49 |
+
|
| 50 |
+
## Training procedure
|
| 51 |
+
|
| 52 |
+
### Training hyperparameters
|
| 53 |
+
|
| 54 |
+
The following hyperparameters were used during training:
|
| 55 |
+
- learning_rate: 2e-05
|
| 56 |
+
- train_batch_size: 8
|
| 57 |
+
- eval_batch_size: 8
|
| 58 |
+
- seed: 1
|
| 59 |
+
- gradient_accumulation_steps: 4
|
| 60 |
+
- total_train_batch_size: 32
|
| 61 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 62 |
+
- lr_scheduler_type: linear
|
| 63 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 64 |
+
- num_epochs: 5.0
|
| 65 |
+
- mixed_precision_training: Native AMP
|
| 66 |
+
|
| 67 |
+
### Training results
|
| 68 |
+
|
| 69 |
+
| Training Loss | Epoch | Step | Validation Loss | F1 |
|
| 70 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 71 |
+
| 0.0317 | 1.0 | 8538 | 0.1732 | 0.9492 |
|
| 72 |
+
| 0.008 | 2.0 | 17076 | 0.3541 | 0.9270 |
|
| 73 |
+
| 0.0085 | 3.0 | 25614 | 0.1161 | 0.9726 |
|
| 74 |
+
| 0.0015 | 4.0 | 34152 | 0.2557 | 0.9516 |
|
| 75 |
+
| 0.0 | 5.0 | 42690 | 0.2286 | 0.9650 |
|
| 76 |
+
|
| 77 |
+
|
| 78 |
+
### Framework versions
|
| 79 |
+
|
| 80 |
+
- Transformers 4.26.1
|
| 81 |
+
- Pytorch 1.11.0+cu115
|
| 82 |
+
- Datasets 2.8.0
|
| 83 |
+
- Tokenizers 0.13.2
|
all_results.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"epoch_full_en_multi": 5.0,
|
| 4 |
+
"epoch_full_fr_multi": 5.0,
|
| 5 |
+
"eval_f1_full_en_multi": 0.9820795051396344,
|
| 6 |
+
"eval_f1_full_fr_multi": 0.963274059512108,
|
| 7 |
+
"eval_loss_full_en_multi": 0.07490870356559753,
|
| 8 |
+
"eval_loss_full_fr_multi": 0.15729032456874847,
|
| 9 |
+
"eval_runtime_full_en_multi": 23.6848,
|
| 10 |
+
"eval_runtime_full_fr_multi": 24.6114,
|
| 11 |
+
"eval_samples_full_en_multi": 17114,
|
| 12 |
+
"eval_samples_full_fr_multi": 17107,
|
| 13 |
+
"eval_samples_per_second_full_en_multi": 722.575,
|
| 14 |
+
"eval_samples_per_second_full_fr_multi": 695.085,
|
| 15 |
+
"eval_steps_per_second_full_en_multi": 90.353,
|
| 16 |
+
"eval_steps_per_second_full_fr_multi": 86.911,
|
| 17 |
+
"train_loss": 0.023754316836012614,
|
| 18 |
+
"train_runtime": 9371.3538,
|
| 19 |
+
"train_samples": 273234,
|
| 20 |
+
"train_samples_per_second": 145.781,
|
| 21 |
+
"train_steps_per_second": 4.555
|
| 22 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "xlm-roberta-base-chatgptdetect-noisy",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"finetuning_task": "full_multi_1.0_0.5_0.5",
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.1,
|
| 13 |
+
"hidden_size": 768,
|
| 14 |
+
"id2label": {
|
| 15 |
+
"0": "HUMAN",
|
| 16 |
+
"1": "CHATGPT"
|
| 17 |
+
},
|
| 18 |
+
"initializer_range": 0.02,
|
| 19 |
+
"intermediate_size": 3072,
|
| 20 |
+
"label2id": {
|
| 21 |
+
"CHATGPT": 1,
|
| 22 |
+
"HUMAN": 0
|
| 23 |
+
},
|
| 24 |
+
"layer_norm_eps": 1e-05,
|
| 25 |
+
"max_position_embeddings": 514,
|
| 26 |
+
"model_type": "xlm-roberta",
|
| 27 |
+
"num_attention_heads": 12,
|
| 28 |
+
"num_hidden_layers": 12,
|
| 29 |
+
"output_past": true,
|
| 30 |
+
"pad_token_id": 1,
|
| 31 |
+
"position_embedding_type": "absolute",
|
| 32 |
+
"problem_type": "single_label_classification",
|
| 33 |
+
"torch_dtype": "float32",
|
| 34 |
+
"transformers_version": "4.26.1",
|
| 35 |
+
"type_vocab_size": 1,
|
| 36 |
+
"use_cache": true,
|
| 37 |
+
"vocab_size": 250002
|
| 38 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch_full_en_multi": 5.0,
|
| 3 |
+
"epoch_full_fr_multi": 5.0,
|
| 4 |
+
"eval_f1_full_en_multi": 0.9820795051396344,
|
| 5 |
+
"eval_f1_full_fr_multi": 0.963274059512108,
|
| 6 |
+
"eval_loss_full_en_multi": 0.07490870356559753,
|
| 7 |
+
"eval_loss_full_fr_multi": 0.15729032456874847,
|
| 8 |
+
"eval_runtime_full_en_multi": 23.6848,
|
| 9 |
+
"eval_runtime_full_fr_multi": 24.6114,
|
| 10 |
+
"eval_samples_full_en_multi": 17114,
|
| 11 |
+
"eval_samples_full_fr_multi": 17107,
|
| 12 |
+
"eval_samples_per_second_full_en_multi": 722.575,
|
| 13 |
+
"eval_samples_per_second_full_fr_multi": 695.085,
|
| 14 |
+
"eval_steps_per_second_full_en_multi": 90.353,
|
| 15 |
+
"eval_steps_per_second_full_fr_multi": 86.911
|
| 16 |
+
}
|
predict_results_full_multi_1.0_0.5_0.5.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b0861dd42c470c9320d2d09611cf583984f39ab6867b83717c7cf6ce4528a871
|
| 3 |
+
size 1112252397
|
sentencepiece.bpe.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
|
| 3 |
+
size 5069051
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": {
|
| 6 |
+
"content": "<mask>",
|
| 7 |
+
"lstrip": true,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false
|
| 11 |
+
},
|
| 12 |
+
"pad_token": "<pad>",
|
| 13 |
+
"sep_token": "</s>",
|
| 14 |
+
"unk_token": "<unk>"
|
| 15 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": {
|
| 6 |
+
"__type": "AddedToken",
|
| 7 |
+
"content": "<mask>",
|
| 8 |
+
"lstrip": true,
|
| 9 |
+
"normalized": true,
|
| 10 |
+
"rstrip": false,
|
| 11 |
+
"single_word": false
|
| 12 |
+
},
|
| 13 |
+
"model_max_length": 512,
|
| 14 |
+
"name_or_path": "xlm-roberta-base",
|
| 15 |
+
"pad_token": "<pad>",
|
| 16 |
+
"sep_token": "</s>",
|
| 17 |
+
"special_tokens_map_file": null,
|
| 18 |
+
"tokenizer_class": "XLMRobertaTokenizer",
|
| 19 |
+
"unk_token": "<unk>"
|
| 20 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"train_loss": 0.023754316836012614,
|
| 4 |
+
"train_runtime": 9371.3538,
|
| 5 |
+
"train_samples": 273234,
|
| 6 |
+
"train_samples_per_second": 145.781,
|
| 7 |
+
"train_steps_per_second": 4.555
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2626 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9725928597187162,
|
| 3 |
+
"best_model_checkpoint": "/scratch_results/multi/full-1.0_0.5_0.5/xlm-roberta/gradient_accumulation_steps-4-learning_rate-2e-05-epochs-5-lr_scheduler-linear-warmup_steps-0.1/SEED-1/checkpoint-25614",
|
| 4 |
+
"epoch": 4.999912165129556,
|
| 5 |
+
"global_step": 42690,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 4.6849379245724996e-07,
|
| 13 |
+
"loss": 0.6875,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.02,
|
| 18 |
+
"learning_rate": 9.369875849144999e-07,
|
| 19 |
+
"loss": 0.6518,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.04,
|
| 24 |
+
"learning_rate": 1.4007964394471774e-06,
|
| 25 |
+
"loss": 0.5616,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.05,
|
| 30 |
+
"learning_rate": 1.8646052939798549e-06,
|
| 31 |
+
"loss": 0.4713,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.06,
|
| 36 |
+
"learning_rate": 2.333099086437105e-06,
|
| 37 |
+
"loss": 0.4066,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.07,
|
| 42 |
+
"learning_rate": 2.7969079409697823e-06,
|
| 43 |
+
"loss": 0.325,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.08,
|
| 48 |
+
"learning_rate": 3.26071679550246e-06,
|
| 49 |
+
"loss": 0.2452,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.09,
|
| 54 |
+
"learning_rate": 3.7292105879597097e-06,
|
| 55 |
+
"loss": 0.1638,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.11,
|
| 60 |
+
"learning_rate": 4.1977043804169594e-06,
|
| 61 |
+
"loss": 0.1392,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.12,
|
| 66 |
+
"learning_rate": 4.66619817287421e-06,
|
| 67 |
+
"loss": 0.1285,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.13,
|
| 72 |
+
"learning_rate": 5.13469196533146e-06,
|
| 73 |
+
"loss": 0.1207,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.14,
|
| 78 |
+
"learning_rate": 5.6031857577887095e-06,
|
| 79 |
+
"loss": 0.1013,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.15,
|
| 84 |
+
"learning_rate": 6.07167955024596e-06,
|
| 85 |
+
"loss": 0.12,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.16,
|
| 90 |
+
"learning_rate": 6.535488404778637e-06,
|
| 91 |
+
"loss": 0.0914,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.18,
|
| 96 |
+
"learning_rate": 7.003982197235887e-06,
|
| 97 |
+
"loss": 0.1162,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.19,
|
| 102 |
+
"learning_rate": 7.472475989693137e-06,
|
| 103 |
+
"loss": 0.1136,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.2,
|
| 108 |
+
"learning_rate": 7.940969782150387e-06,
|
| 109 |
+
"loss": 0.0809,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.21,
|
| 114 |
+
"learning_rate": 8.409463574607638e-06,
|
| 115 |
+
"loss": 0.0871,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.22,
|
| 120 |
+
"learning_rate": 8.877957367064887e-06,
|
| 121 |
+
"loss": 0.088,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.23,
|
| 126 |
+
"learning_rate": 9.346451159522137e-06,
|
| 127 |
+
"loss": 0.0985,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.25,
|
| 132 |
+
"learning_rate": 9.814944951979388e-06,
|
| 133 |
+
"loss": 0.0847,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.26,
|
| 138 |
+
"learning_rate": 1.0283438744436638e-05,
|
| 139 |
+
"loss": 0.0807,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.27,
|
| 144 |
+
"learning_rate": 1.0751932536893887e-05,
|
| 145 |
+
"loss": 0.092,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.28,
|
| 150 |
+
"learning_rate": 1.1220426329351138e-05,
|
| 151 |
+
"loss": 0.1086,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.29,
|
| 156 |
+
"learning_rate": 1.1688920121808387e-05,
|
| 157 |
+
"loss": 0.071,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.3,
|
| 162 |
+
"learning_rate": 1.2157413914265637e-05,
|
| 163 |
+
"loss": 0.105,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.32,
|
| 168 |
+
"learning_rate": 1.2625907706722886e-05,
|
| 169 |
+
"loss": 0.0757,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.33,
|
| 174 |
+
"learning_rate": 1.3094401499180137e-05,
|
| 175 |
+
"loss": 0.0639,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.34,
|
| 180 |
+
"learning_rate": 1.3562895291637387e-05,
|
| 181 |
+
"loss": 0.0778,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.35,
|
| 186 |
+
"learning_rate": 1.4031389084094638e-05,
|
| 187 |
+
"loss": 0.0774,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.36,
|
| 192 |
+
"learning_rate": 1.4499882876551887e-05,
|
| 193 |
+
"loss": 0.0616,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.37,
|
| 198 |
+
"learning_rate": 1.4968376669009137e-05,
|
| 199 |
+
"loss": 0.0723,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.39,
|
| 204 |
+
"learning_rate": 1.5436870461466388e-05,
|
| 205 |
+
"loss": 0.0645,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.4,
|
| 210 |
+
"learning_rate": 1.5905364253923635e-05,
|
| 211 |
+
"loss": 0.0496,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.41,
|
| 216 |
+
"learning_rate": 1.6373858046380886e-05,
|
| 217 |
+
"loss": 0.0579,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.42,
|
| 222 |
+
"learning_rate": 1.6842351838838136e-05,
|
| 223 |
+
"loss": 0.0513,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.43,
|
| 228 |
+
"learning_rate": 1.7310845631295387e-05,
|
| 229 |
+
"loss": 0.0463,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.45,
|
| 234 |
+
"learning_rate": 1.7779339423752638e-05,
|
| 235 |
+
"loss": 0.0603,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.46,
|
| 240 |
+
"learning_rate": 1.8247833216209888e-05,
|
| 241 |
+
"loss": 0.0506,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.47,
|
| 246 |
+
"learning_rate": 1.8716327008667135e-05,
|
| 247 |
+
"loss": 0.0607,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.48,
|
| 252 |
+
"learning_rate": 1.9184820801124386e-05,
|
| 253 |
+
"loss": 0.0588,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.49,
|
| 258 |
+
"learning_rate": 1.9653314593581637e-05,
|
| 259 |
+
"loss": 0.0486,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.5,
|
| 264 |
+
"learning_rate": 1.998646573488457e-05,
|
| 265 |
+
"loss": 0.0644,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.52,
|
| 270 |
+
"learning_rate": 1.9934410869055988e-05,
|
| 271 |
+
"loss": 0.0433,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.53,
|
| 276 |
+
"learning_rate": 1.9882356003227403e-05,
|
| 277 |
+
"loss": 0.0462,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.54,
|
| 282 |
+
"learning_rate": 1.983030113739882e-05,
|
| 283 |
+
"loss": 0.049,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.55,
|
| 288 |
+
"learning_rate": 1.9778246271570237e-05,
|
| 289 |
+
"loss": 0.0394,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.56,
|
| 294 |
+
"learning_rate": 1.9726191405741653e-05,
|
| 295 |
+
"loss": 0.0394,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.57,
|
| 300 |
+
"learning_rate": 1.967413653991307e-05,
|
| 301 |
+
"loss": 0.0501,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.59,
|
| 306 |
+
"learning_rate": 1.9622081674084487e-05,
|
| 307 |
+
"loss": 0.0421,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.6,
|
| 312 |
+
"learning_rate": 1.9570026808255903e-05,
|
| 313 |
+
"loss": 0.0514,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.61,
|
| 318 |
+
"learning_rate": 1.951797194242732e-05,
|
| 319 |
+
"loss": 0.0364,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.62,
|
| 324 |
+
"learning_rate": 1.9465917076598737e-05,
|
| 325 |
+
"loss": 0.0476,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.63,
|
| 330 |
+
"learning_rate": 1.9413862210770153e-05,
|
| 331 |
+
"loss": 0.0499,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.64,
|
| 336 |
+
"learning_rate": 1.936180734494157e-05,
|
| 337 |
+
"loss": 0.0449,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.66,
|
| 342 |
+
"learning_rate": 1.9309752479112987e-05,
|
| 343 |
+
"loss": 0.0413,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.67,
|
| 348 |
+
"learning_rate": 1.9257697613284402e-05,
|
| 349 |
+
"loss": 0.0274,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.68,
|
| 354 |
+
"learning_rate": 1.9205642747455818e-05,
|
| 355 |
+
"loss": 0.0368,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.69,
|
| 360 |
+
"learning_rate": 1.9154108430285523e-05,
|
| 361 |
+
"loss": 0.0424,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.7,
|
| 366 |
+
"learning_rate": 1.9102053564456938e-05,
|
| 367 |
+
"loss": 0.0557,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.71,
|
| 372 |
+
"learning_rate": 1.9049998698628357e-05,
|
| 373 |
+
"loss": 0.0355,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.73,
|
| 378 |
+
"learning_rate": 1.8997943832799772e-05,
|
| 379 |
+
"loss": 0.0469,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.74,
|
| 384 |
+
"learning_rate": 1.8945888966971188e-05,
|
| 385 |
+
"loss": 0.0576,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.75,
|
| 390 |
+
"learning_rate": 1.8893834101142607e-05,
|
| 391 |
+
"loss": 0.033,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.76,
|
| 396 |
+
"learning_rate": 1.8841779235314022e-05,
|
| 397 |
+
"loss": 0.0379,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.77,
|
| 402 |
+
"learning_rate": 1.8789724369485438e-05,
|
| 403 |
+
"loss": 0.0386,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.78,
|
| 408 |
+
"learning_rate": 1.8737669503656856e-05,
|
| 409 |
+
"loss": 0.0474,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.8,
|
| 414 |
+
"learning_rate": 1.8685614637828272e-05,
|
| 415 |
+
"loss": 0.0352,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.81,
|
| 420 |
+
"learning_rate": 1.863355977199969e-05,
|
| 421 |
+
"loss": 0.0332,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.82,
|
| 426 |
+
"learning_rate": 1.8581504906171106e-05,
|
| 427 |
+
"loss": 0.0357,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.83,
|
| 432 |
+
"learning_rate": 1.8529450040342522e-05,
|
| 433 |
+
"loss": 0.0373,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.84,
|
| 438 |
+
"learning_rate": 1.847739517451394e-05,
|
| 439 |
+
"loss": 0.0273,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.85,
|
| 444 |
+
"learning_rate": 1.8425340308685356e-05,
|
| 445 |
+
"loss": 0.0299,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.87,
|
| 450 |
+
"learning_rate": 1.837328544285677e-05,
|
| 451 |
+
"loss": 0.0279,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.88,
|
| 456 |
+
"learning_rate": 1.8321230577028187e-05,
|
| 457 |
+
"loss": 0.0321,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.89,
|
| 462 |
+
"learning_rate": 1.8269175711199606e-05,
|
| 463 |
+
"loss": 0.0257,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.9,
|
| 468 |
+
"learning_rate": 1.8217120845371025e-05,
|
| 469 |
+
"loss": 0.0223,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.91,
|
| 474 |
+
"learning_rate": 1.816506597954244e-05,
|
| 475 |
+
"loss": 0.0312,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.93,
|
| 480 |
+
"learning_rate": 1.8113011113713856e-05,
|
| 481 |
+
"loss": 0.0243,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.94,
|
| 486 |
+
"learning_rate": 1.8060956247885275e-05,
|
| 487 |
+
"loss": 0.0266,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.95,
|
| 492 |
+
"learning_rate": 1.800890138205669e-05,
|
| 493 |
+
"loss": 0.0311,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.96,
|
| 498 |
+
"learning_rate": 1.7956846516228105e-05,
|
| 499 |
+
"loss": 0.0175,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.97,
|
| 504 |
+
"learning_rate": 1.790479165039952e-05,
|
| 505 |
+
"loss": 0.0357,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.98,
|
| 510 |
+
"learning_rate": 1.785273678457094e-05,
|
| 511 |
+
"loss": 0.0252,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.0,
|
| 516 |
+
"learning_rate": 1.7800681918742355e-05,
|
| 517 |
+
"loss": 0.0317,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 1.0,
|
| 522 |
+
"eval_f1": 0.9491644678979771,
|
| 523 |
+
"eval_loss": 0.17317870259284973,
|
| 524 |
+
"eval_runtime": 47.425,
|
| 525 |
+
"eval_samples_per_second": 721.582,
|
| 526 |
+
"eval_steps_per_second": 90.206,
|
| 527 |
+
"step": 8538
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 1.01,
|
| 531 |
+
"learning_rate": 1.774862705291377e-05,
|
| 532 |
+
"loss": 0.0238,
|
| 533 |
+
"step": 8600
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 1.02,
|
| 537 |
+
"learning_rate": 1.769657218708519e-05,
|
| 538 |
+
"loss": 0.0224,
|
| 539 |
+
"step": 8700
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 1.03,
|
| 543 |
+
"learning_rate": 1.7644517321256605e-05,
|
| 544 |
+
"loss": 0.015,
|
| 545 |
+
"step": 8800
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 1.04,
|
| 549 |
+
"learning_rate": 1.7592462455428024e-05,
|
| 550 |
+
"loss": 0.0261,
|
| 551 |
+
"step": 8900
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 1.05,
|
| 555 |
+
"learning_rate": 1.754040758959944e-05,
|
| 556 |
+
"loss": 0.0234,
|
| 557 |
+
"step": 9000
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 1.07,
|
| 561 |
+
"learning_rate": 1.7488352723770855e-05,
|
| 562 |
+
"loss": 0.0261,
|
| 563 |
+
"step": 9100
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 1.08,
|
| 567 |
+
"learning_rate": 1.7436297857942274e-05,
|
| 568 |
+
"loss": 0.0168,
|
| 569 |
+
"step": 9200
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 1.09,
|
| 573 |
+
"learning_rate": 1.738424299211369e-05,
|
| 574 |
+
"loss": 0.0154,
|
| 575 |
+
"step": 9300
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 1.1,
|
| 579 |
+
"learning_rate": 1.7332708674943394e-05,
|
| 580 |
+
"loss": 0.0279,
|
| 581 |
+
"step": 9400
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 1.11,
|
| 585 |
+
"learning_rate": 1.728065380911481e-05,
|
| 586 |
+
"loss": 0.0219,
|
| 587 |
+
"step": 9500
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 1.12,
|
| 591 |
+
"learning_rate": 1.7228598943286225e-05,
|
| 592 |
+
"loss": 0.0265,
|
| 593 |
+
"step": 9600
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 1.14,
|
| 597 |
+
"learning_rate": 1.717654407745764e-05,
|
| 598 |
+
"loss": 0.0183,
|
| 599 |
+
"step": 9700
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 1.15,
|
| 603 |
+
"learning_rate": 1.712448921162906e-05,
|
| 604 |
+
"loss": 0.0254,
|
| 605 |
+
"step": 9800
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 1.16,
|
| 609 |
+
"learning_rate": 1.7072434345800475e-05,
|
| 610 |
+
"loss": 0.0167,
|
| 611 |
+
"step": 9900
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 1.17,
|
| 615 |
+
"learning_rate": 1.702037947997189e-05,
|
| 616 |
+
"loss": 0.0143,
|
| 617 |
+
"step": 10000
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 1.18,
|
| 621 |
+
"learning_rate": 1.696832461414331e-05,
|
| 622 |
+
"loss": 0.0197,
|
| 623 |
+
"step": 10100
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 1.19,
|
| 627 |
+
"learning_rate": 1.6916269748314725e-05,
|
| 628 |
+
"loss": 0.0304,
|
| 629 |
+
"step": 10200
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 1.21,
|
| 633 |
+
"learning_rate": 1.686421488248614e-05,
|
| 634 |
+
"loss": 0.0105,
|
| 635 |
+
"step": 10300
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 1.22,
|
| 639 |
+
"learning_rate": 1.681216001665756e-05,
|
| 640 |
+
"loss": 0.033,
|
| 641 |
+
"step": 10400
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 1.23,
|
| 645 |
+
"learning_rate": 1.6760105150828974e-05,
|
| 646 |
+
"loss": 0.0195,
|
| 647 |
+
"step": 10500
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 1.24,
|
| 651 |
+
"learning_rate": 1.6708050285000393e-05,
|
| 652 |
+
"loss": 0.0314,
|
| 653 |
+
"step": 10600
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 1.25,
|
| 657 |
+
"learning_rate": 1.665599541917181e-05,
|
| 658 |
+
"loss": 0.0163,
|
| 659 |
+
"step": 10700
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 1.26,
|
| 663 |
+
"learning_rate": 1.6603940553343224e-05,
|
| 664 |
+
"loss": 0.0156,
|
| 665 |
+
"step": 10800
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 1.28,
|
| 669 |
+
"learning_rate": 1.6551885687514643e-05,
|
| 670 |
+
"loss": 0.014,
|
| 671 |
+
"step": 10900
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 1.29,
|
| 675 |
+
"learning_rate": 1.649983082168606e-05,
|
| 676 |
+
"loss": 0.0128,
|
| 677 |
+
"step": 11000
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 1.3,
|
| 681 |
+
"learning_rate": 1.6447775955857474e-05,
|
| 682 |
+
"loss": 0.0167,
|
| 683 |
+
"step": 11100
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 1.31,
|
| 687 |
+
"learning_rate": 1.639572109002889e-05,
|
| 688 |
+
"loss": 0.0174,
|
| 689 |
+
"step": 11200
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 1.32,
|
| 693 |
+
"learning_rate": 1.6343666224200308e-05,
|
| 694 |
+
"loss": 0.0218,
|
| 695 |
+
"step": 11300
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 1.34,
|
| 699 |
+
"learning_rate": 1.6291611358371727e-05,
|
| 700 |
+
"loss": 0.0072,
|
| 701 |
+
"step": 11400
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 1.35,
|
| 705 |
+
"learning_rate": 1.6239556492543143e-05,
|
| 706 |
+
"loss": 0.0199,
|
| 707 |
+
"step": 11500
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 1.36,
|
| 711 |
+
"learning_rate": 1.6187501626714558e-05,
|
| 712 |
+
"loss": 0.0197,
|
| 713 |
+
"step": 11600
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 1.37,
|
| 717 |
+
"learning_rate": 1.6135446760885977e-05,
|
| 718 |
+
"loss": 0.0166,
|
| 719 |
+
"step": 11700
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 1.38,
|
| 723 |
+
"learning_rate": 1.6083391895057392e-05,
|
| 724 |
+
"loss": 0.0173,
|
| 725 |
+
"step": 11800
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 1.39,
|
| 729 |
+
"learning_rate": 1.6031337029228808e-05,
|
| 730 |
+
"loss": 0.0157,
|
| 731 |
+
"step": 11900
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 1.41,
|
| 735 |
+
"learning_rate": 1.5979282163400223e-05,
|
| 736 |
+
"loss": 0.0207,
|
| 737 |
+
"step": 12000
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 1.42,
|
| 741 |
+
"learning_rate": 1.5927227297571642e-05,
|
| 742 |
+
"loss": 0.0155,
|
| 743 |
+
"step": 12100
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 1.43,
|
| 747 |
+
"learning_rate": 1.5875172431743058e-05,
|
| 748 |
+
"loss": 0.0095,
|
| 749 |
+
"step": 12200
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 1.44,
|
| 753 |
+
"learning_rate": 1.5823117565914473e-05,
|
| 754 |
+
"loss": 0.0172,
|
| 755 |
+
"step": 12300
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 1.45,
|
| 759 |
+
"learning_rate": 1.5771062700085892e-05,
|
| 760 |
+
"loss": 0.0194,
|
| 761 |
+
"step": 12400
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 1.46,
|
| 765 |
+
"learning_rate": 1.571900783425731e-05,
|
| 766 |
+
"loss": 0.0145,
|
| 767 |
+
"step": 12500
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 1.48,
|
| 771 |
+
"learning_rate": 1.5666952968428726e-05,
|
| 772 |
+
"loss": 0.0188,
|
| 773 |
+
"step": 12600
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 1.49,
|
| 777 |
+
"learning_rate": 1.5614898102600142e-05,
|
| 778 |
+
"loss": 0.0149,
|
| 779 |
+
"step": 12700
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 1.5,
|
| 783 |
+
"learning_rate": 1.5562843236771557e-05,
|
| 784 |
+
"loss": 0.0169,
|
| 785 |
+
"step": 12800
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 1.51,
|
| 789 |
+
"learning_rate": 1.5510788370942976e-05,
|
| 790 |
+
"loss": 0.0167,
|
| 791 |
+
"step": 12900
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 1.52,
|
| 795 |
+
"learning_rate": 1.545873350511439e-05,
|
| 796 |
+
"loss": 0.0122,
|
| 797 |
+
"step": 13000
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 1.53,
|
| 801 |
+
"learning_rate": 1.5406678639285807e-05,
|
| 802 |
+
"loss": 0.0163,
|
| 803 |
+
"step": 13100
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 1.55,
|
| 807 |
+
"learning_rate": 1.5355144322115512e-05,
|
| 808 |
+
"loss": 0.0205,
|
| 809 |
+
"step": 13200
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 1.56,
|
| 813 |
+
"learning_rate": 1.5303089456286927e-05,
|
| 814 |
+
"loss": 0.0157,
|
| 815 |
+
"step": 13300
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 1.57,
|
| 819 |
+
"learning_rate": 1.5251034590458344e-05,
|
| 820 |
+
"loss": 0.0134,
|
| 821 |
+
"step": 13400
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 1.58,
|
| 825 |
+
"learning_rate": 1.5198979724629762e-05,
|
| 826 |
+
"loss": 0.0227,
|
| 827 |
+
"step": 13500
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 1.59,
|
| 831 |
+
"learning_rate": 1.5146924858801177e-05,
|
| 832 |
+
"loss": 0.0219,
|
| 833 |
+
"step": 13600
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 1.6,
|
| 837 |
+
"learning_rate": 1.5094869992972594e-05,
|
| 838 |
+
"loss": 0.013,
|
| 839 |
+
"step": 13700
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 1.62,
|
| 843 |
+
"learning_rate": 1.504281512714401e-05,
|
| 844 |
+
"loss": 0.0135,
|
| 845 |
+
"step": 13800
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 1.63,
|
| 849 |
+
"learning_rate": 1.4990760261315427e-05,
|
| 850 |
+
"loss": 0.0207,
|
| 851 |
+
"step": 13900
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 1.64,
|
| 855 |
+
"learning_rate": 1.4938705395486844e-05,
|
| 856 |
+
"loss": 0.0174,
|
| 857 |
+
"step": 14000
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 1.65,
|
| 861 |
+
"learning_rate": 1.4886650529658261e-05,
|
| 862 |
+
"loss": 0.006,
|
| 863 |
+
"step": 14100
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 1.66,
|
| 867 |
+
"learning_rate": 1.4834595663829678e-05,
|
| 868 |
+
"loss": 0.0108,
|
| 869 |
+
"step": 14200
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 1.67,
|
| 873 |
+
"learning_rate": 1.4782540798001096e-05,
|
| 874 |
+
"loss": 0.0123,
|
| 875 |
+
"step": 14300
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 1.69,
|
| 879 |
+
"learning_rate": 1.4730485932172511e-05,
|
| 880 |
+
"loss": 0.0156,
|
| 881 |
+
"step": 14400
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 1.7,
|
| 885 |
+
"learning_rate": 1.4678431066343928e-05,
|
| 886 |
+
"loss": 0.0262,
|
| 887 |
+
"step": 14500
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 1.71,
|
| 891 |
+
"learning_rate": 1.4626376200515344e-05,
|
| 892 |
+
"loss": 0.0136,
|
| 893 |
+
"step": 14600
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 1.72,
|
| 897 |
+
"learning_rate": 1.457432133468676e-05,
|
| 898 |
+
"loss": 0.0098,
|
| 899 |
+
"step": 14700
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 1.73,
|
| 903 |
+
"learning_rate": 1.4522266468858178e-05,
|
| 904 |
+
"loss": 0.0138,
|
| 905 |
+
"step": 14800
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 1.75,
|
| 909 |
+
"learning_rate": 1.4470211603029593e-05,
|
| 910 |
+
"loss": 0.0132,
|
| 911 |
+
"step": 14900
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 1.76,
|
| 915 |
+
"learning_rate": 1.441815673720101e-05,
|
| 916 |
+
"loss": 0.0106,
|
| 917 |
+
"step": 15000
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 1.77,
|
| 921 |
+
"learning_rate": 1.436610187137243e-05,
|
| 922 |
+
"loss": 0.015,
|
| 923 |
+
"step": 15100
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 1.78,
|
| 927 |
+
"learning_rate": 1.4314047005543845e-05,
|
| 928 |
+
"loss": 0.0167,
|
| 929 |
+
"step": 15200
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 1.79,
|
| 933 |
+
"learning_rate": 1.4261992139715262e-05,
|
| 934 |
+
"loss": 0.0084,
|
| 935 |
+
"step": 15300
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 1.8,
|
| 939 |
+
"learning_rate": 1.4209937273886678e-05,
|
| 940 |
+
"loss": 0.0188,
|
| 941 |
+
"step": 15400
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 1.82,
|
| 945 |
+
"learning_rate": 1.4157882408058095e-05,
|
| 946 |
+
"loss": 0.0088,
|
| 947 |
+
"step": 15500
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 1.83,
|
| 951 |
+
"learning_rate": 1.4106348090887796e-05,
|
| 952 |
+
"loss": 0.0171,
|
| 953 |
+
"step": 15600
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 1.84,
|
| 957 |
+
"learning_rate": 1.4054293225059215e-05,
|
| 958 |
+
"loss": 0.0168,
|
| 959 |
+
"step": 15700
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 1.85,
|
| 963 |
+
"learning_rate": 1.400223835923063e-05,
|
| 964 |
+
"loss": 0.009,
|
| 965 |
+
"step": 15800
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 1.86,
|
| 969 |
+
"learning_rate": 1.3950183493402048e-05,
|
| 970 |
+
"loss": 0.0078,
|
| 971 |
+
"step": 15900
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 1.87,
|
| 975 |
+
"learning_rate": 1.3898128627573465e-05,
|
| 976 |
+
"loss": 0.0108,
|
| 977 |
+
"step": 16000
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 1.89,
|
| 981 |
+
"learning_rate": 1.3846594310403166e-05,
|
| 982 |
+
"loss": 0.0072,
|
| 983 |
+
"step": 16100
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 1.9,
|
| 987 |
+
"learning_rate": 1.3794539444574582e-05,
|
| 988 |
+
"loss": 0.0089,
|
| 989 |
+
"step": 16200
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 1.91,
|
| 993 |
+
"learning_rate": 1.3742484578746e-05,
|
| 994 |
+
"loss": 0.0055,
|
| 995 |
+
"step": 16300
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 1.92,
|
| 999 |
+
"learning_rate": 1.3690429712917418e-05,
|
| 1000 |
+
"loss": 0.0177,
|
| 1001 |
+
"step": 16400
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 1.93,
|
| 1005 |
+
"learning_rate": 1.3638374847088833e-05,
|
| 1006 |
+
"loss": 0.008,
|
| 1007 |
+
"step": 16500
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 1.94,
|
| 1011 |
+
"learning_rate": 1.358631998126025e-05,
|
| 1012 |
+
"loss": 0.0127,
|
| 1013 |
+
"step": 16600
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 1.96,
|
| 1017 |
+
"learning_rate": 1.3534265115431666e-05,
|
| 1018 |
+
"loss": 0.0069,
|
| 1019 |
+
"step": 16700
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 1.97,
|
| 1023 |
+
"learning_rate": 1.3482210249603083e-05,
|
| 1024 |
+
"loss": 0.0089,
|
| 1025 |
+
"step": 16800
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 1.98,
|
| 1029 |
+
"learning_rate": 1.3430155383774498e-05,
|
| 1030 |
+
"loss": 0.0094,
|
| 1031 |
+
"step": 16900
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 1.99,
|
| 1035 |
+
"learning_rate": 1.3378100517945916e-05,
|
| 1036 |
+
"loss": 0.008,
|
| 1037 |
+
"step": 17000
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 2.0,
|
| 1041 |
+
"eval_f1": 0.9269947608004809,
|
| 1042 |
+
"eval_loss": 0.3541407883167267,
|
| 1043 |
+
"eval_runtime": 56.6762,
|
| 1044 |
+
"eval_samples_per_second": 603.799,
|
| 1045 |
+
"eval_steps_per_second": 75.481,
|
| 1046 |
+
"step": 17076
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 2.0,
|
| 1050 |
+
"learning_rate": 1.3326045652117333e-05,
|
| 1051 |
+
"loss": 0.0091,
|
| 1052 |
+
"step": 17100
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 2.01,
|
| 1056 |
+
"learning_rate": 1.3273990786288748e-05,
|
| 1057 |
+
"loss": 0.0075,
|
| 1058 |
+
"step": 17200
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 2.03,
|
| 1062 |
+
"learning_rate": 1.3221935920460167e-05,
|
| 1063 |
+
"loss": 0.0084,
|
| 1064 |
+
"step": 17300
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 2.04,
|
| 1068 |
+
"learning_rate": 1.3169881054631584e-05,
|
| 1069 |
+
"loss": 0.0032,
|
| 1070 |
+
"step": 17400
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 2.05,
|
| 1074 |
+
"learning_rate": 1.3117826188803e-05,
|
| 1075 |
+
"loss": 0.0113,
|
| 1076 |
+
"step": 17500
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 2.06,
|
| 1080 |
+
"learning_rate": 1.3065771322974417e-05,
|
| 1081 |
+
"loss": 0.0053,
|
| 1082 |
+
"step": 17600
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 2.07,
|
| 1086 |
+
"learning_rate": 1.3013716457145832e-05,
|
| 1087 |
+
"loss": 0.0077,
|
| 1088 |
+
"step": 17700
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 2.08,
|
| 1092 |
+
"learning_rate": 1.296166159131725e-05,
|
| 1093 |
+
"loss": 0.0058,
|
| 1094 |
+
"step": 17800
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 2.1,
|
| 1098 |
+
"learning_rate": 1.2909606725488667e-05,
|
| 1099 |
+
"loss": 0.0085,
|
| 1100 |
+
"step": 17900
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 2.11,
|
| 1104 |
+
"learning_rate": 1.2857551859660082e-05,
|
| 1105 |
+
"loss": 0.0029,
|
| 1106 |
+
"step": 18000
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 2.12,
|
| 1110 |
+
"learning_rate": 1.28054969938315e-05,
|
| 1111 |
+
"loss": 0.0059,
|
| 1112 |
+
"step": 18100
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 2.13,
|
| 1116 |
+
"learning_rate": 1.2753442128002915e-05,
|
| 1117 |
+
"loss": 0.0079,
|
| 1118 |
+
"step": 18200
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 2.14,
|
| 1122 |
+
"learning_rate": 1.2701387262174332e-05,
|
| 1123 |
+
"loss": 0.004,
|
| 1124 |
+
"step": 18300
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 2.16,
|
| 1128 |
+
"learning_rate": 1.264933239634575e-05,
|
| 1129 |
+
"loss": 0.0096,
|
| 1130 |
+
"step": 18400
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 2.17,
|
| 1134 |
+
"learning_rate": 1.2597277530517166e-05,
|
| 1135 |
+
"loss": 0.0063,
|
| 1136 |
+
"step": 18500
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 2.18,
|
| 1140 |
+
"learning_rate": 1.2545222664688583e-05,
|
| 1141 |
+
"loss": 0.0029,
|
| 1142 |
+
"step": 18600
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 2.19,
|
| 1146 |
+
"learning_rate": 1.249316779886e-05,
|
| 1147 |
+
"loss": 0.0119,
|
| 1148 |
+
"step": 18700
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 2.2,
|
| 1152 |
+
"learning_rate": 1.2441112933031416e-05,
|
| 1153 |
+
"loss": 0.0088,
|
| 1154 |
+
"step": 18800
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 2.21,
|
| 1158 |
+
"learning_rate": 1.2389058067202833e-05,
|
| 1159 |
+
"loss": 0.004,
|
| 1160 |
+
"step": 18900
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 2.23,
|
| 1164 |
+
"learning_rate": 1.2337003201374249e-05,
|
| 1165 |
+
"loss": 0.0066,
|
| 1166 |
+
"step": 19000
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 2.24,
|
| 1170 |
+
"learning_rate": 1.2284948335545666e-05,
|
| 1171 |
+
"loss": 0.0058,
|
| 1172 |
+
"step": 19100
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 2.25,
|
| 1176 |
+
"learning_rate": 1.2232893469717081e-05,
|
| 1177 |
+
"loss": 0.005,
|
| 1178 |
+
"step": 19200
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 2.26,
|
| 1182 |
+
"learning_rate": 1.2180838603888498e-05,
|
| 1183 |
+
"loss": 0.0121,
|
| 1184 |
+
"step": 19300
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 2.27,
|
| 1188 |
+
"learning_rate": 1.2128783738059917e-05,
|
| 1189 |
+
"loss": 0.0039,
|
| 1190 |
+
"step": 19400
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 2.28,
|
| 1194 |
+
"learning_rate": 1.2076728872231333e-05,
|
| 1195 |
+
"loss": 0.0053,
|
| 1196 |
+
"step": 19500
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 2.3,
|
| 1200 |
+
"learning_rate": 1.202467400640275e-05,
|
| 1201 |
+
"loss": 0.0041,
|
| 1202 |
+
"step": 19600
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 2.31,
|
| 1206 |
+
"learning_rate": 1.1972619140574167e-05,
|
| 1207 |
+
"loss": 0.019,
|
| 1208 |
+
"step": 19700
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 2.32,
|
| 1212 |
+
"learning_rate": 1.1920564274745583e-05,
|
| 1213 |
+
"loss": 0.0065,
|
| 1214 |
+
"step": 19800
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 2.33,
|
| 1218 |
+
"learning_rate": 1.1868509408917e-05,
|
| 1219 |
+
"loss": 0.0099,
|
| 1220 |
+
"step": 19900
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 2.34,
|
| 1224 |
+
"learning_rate": 1.1816454543088415e-05,
|
| 1225 |
+
"loss": 0.0029,
|
| 1226 |
+
"step": 20000
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 2.35,
|
| 1230 |
+
"learning_rate": 1.1764399677259832e-05,
|
| 1231 |
+
"loss": 0.0045,
|
| 1232 |
+
"step": 20100
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 2.37,
|
| 1236 |
+
"learning_rate": 1.1712865360089535e-05,
|
| 1237 |
+
"loss": 0.013,
|
| 1238 |
+
"step": 20200
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 2.38,
|
| 1242 |
+
"learning_rate": 1.1660810494260953e-05,
|
| 1243 |
+
"loss": 0.003,
|
| 1244 |
+
"step": 20300
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 2.39,
|
| 1248 |
+
"learning_rate": 1.1608755628432368e-05,
|
| 1249 |
+
"loss": 0.0062,
|
| 1250 |
+
"step": 20400
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 2.4,
|
| 1254 |
+
"learning_rate": 1.1556700762603785e-05,
|
| 1255 |
+
"loss": 0.0054,
|
| 1256 |
+
"step": 20500
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 2.41,
|
| 1260 |
+
"learning_rate": 1.1504645896775202e-05,
|
| 1261 |
+
"loss": 0.0123,
|
| 1262 |
+
"step": 20600
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 2.42,
|
| 1266 |
+
"learning_rate": 1.1452591030946618e-05,
|
| 1267 |
+
"loss": 0.0095,
|
| 1268 |
+
"step": 20700
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 2.44,
|
| 1272 |
+
"learning_rate": 1.1400536165118035e-05,
|
| 1273 |
+
"loss": 0.009,
|
| 1274 |
+
"step": 20800
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 2.45,
|
| 1278 |
+
"learning_rate": 1.134848129928945e-05,
|
| 1279 |
+
"loss": 0.0061,
|
| 1280 |
+
"step": 20900
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 2.46,
|
| 1284 |
+
"learning_rate": 1.129642643346087e-05,
|
| 1285 |
+
"loss": 0.0005,
|
| 1286 |
+
"step": 21000
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 2.47,
|
| 1290 |
+
"learning_rate": 1.1244371567632287e-05,
|
| 1291 |
+
"loss": 0.0015,
|
| 1292 |
+
"step": 21100
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 2.48,
|
| 1296 |
+
"learning_rate": 1.1192316701803702e-05,
|
| 1297 |
+
"loss": 0.0019,
|
| 1298 |
+
"step": 21200
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 2.49,
|
| 1302 |
+
"learning_rate": 1.114026183597512e-05,
|
| 1303 |
+
"loss": 0.0027,
|
| 1304 |
+
"step": 21300
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 2.51,
|
| 1308 |
+
"learning_rate": 1.1088206970146536e-05,
|
| 1309 |
+
"loss": 0.004,
|
| 1310 |
+
"step": 21400
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 2.52,
|
| 1314 |
+
"learning_rate": 1.1036152104317952e-05,
|
| 1315 |
+
"loss": 0.0075,
|
| 1316 |
+
"step": 21500
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 2.53,
|
| 1320 |
+
"learning_rate": 1.0984097238489369e-05,
|
| 1321 |
+
"loss": 0.007,
|
| 1322 |
+
"step": 21600
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 2.54,
|
| 1326 |
+
"learning_rate": 1.0932042372660784e-05,
|
| 1327 |
+
"loss": 0.0039,
|
| 1328 |
+
"step": 21700
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 2.55,
|
| 1332 |
+
"learning_rate": 1.0879987506832202e-05,
|
| 1333 |
+
"loss": 0.0145,
|
| 1334 |
+
"step": 21800
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 2.56,
|
| 1338 |
+
"learning_rate": 1.0827932641003617e-05,
|
| 1339 |
+
"loss": 0.0071,
|
| 1340 |
+
"step": 21900
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 2.58,
|
| 1344 |
+
"learning_rate": 1.0775877775175034e-05,
|
| 1345 |
+
"loss": 0.0168,
|
| 1346 |
+
"step": 22000
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 2.59,
|
| 1350 |
+
"learning_rate": 1.0723822909346453e-05,
|
| 1351 |
+
"loss": 0.0095,
|
| 1352 |
+
"step": 22100
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 2.6,
|
| 1356 |
+
"learning_rate": 1.0671768043517869e-05,
|
| 1357 |
+
"loss": 0.0033,
|
| 1358 |
+
"step": 22200
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 2.61,
|
| 1362 |
+
"learning_rate": 1.0619713177689286e-05,
|
| 1363 |
+
"loss": 0.0022,
|
| 1364 |
+
"step": 22300
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 2.62,
|
| 1368 |
+
"learning_rate": 1.0567658311860703e-05,
|
| 1369 |
+
"loss": 0.0022,
|
| 1370 |
+
"step": 22400
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 2.64,
|
| 1374 |
+
"learning_rate": 1.0515603446032118e-05,
|
| 1375 |
+
"loss": 0.0082,
|
| 1376 |
+
"step": 22500
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 2.65,
|
| 1380 |
+
"learning_rate": 1.0463548580203536e-05,
|
| 1381 |
+
"loss": 0.007,
|
| 1382 |
+
"step": 22600
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 2.66,
|
| 1386 |
+
"learning_rate": 1.0411493714374951e-05,
|
| 1387 |
+
"loss": 0.0043,
|
| 1388 |
+
"step": 22700
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 2.67,
|
| 1392 |
+
"learning_rate": 1.0359438848546368e-05,
|
| 1393 |
+
"loss": 0.0109,
|
| 1394 |
+
"step": 22800
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 2.68,
|
| 1398 |
+
"learning_rate": 1.0307383982717785e-05,
|
| 1399 |
+
"loss": 0.0023,
|
| 1400 |
+
"step": 22900
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 2.69,
|
| 1404 |
+
"learning_rate": 1.02553291168892e-05,
|
| 1405 |
+
"loss": 0.0037,
|
| 1406 |
+
"step": 23000
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 2.71,
|
| 1410 |
+
"learning_rate": 1.0203794799718904e-05,
|
| 1411 |
+
"loss": 0.0072,
|
| 1412 |
+
"step": 23100
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 2.72,
|
| 1416 |
+
"learning_rate": 1.0151739933890321e-05,
|
| 1417 |
+
"loss": 0.0096,
|
| 1418 |
+
"step": 23200
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 2.73,
|
| 1422 |
+
"learning_rate": 1.0099685068061738e-05,
|
| 1423 |
+
"loss": 0.0072,
|
| 1424 |
+
"step": 23300
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 2.74,
|
| 1428 |
+
"learning_rate": 1.0047630202233154e-05,
|
| 1429 |
+
"loss": 0.0021,
|
| 1430 |
+
"step": 23400
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 2.75,
|
| 1434 |
+
"learning_rate": 9.99557533640457e-06,
|
| 1435 |
+
"loss": 0.0075,
|
| 1436 |
+
"step": 23500
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 2.76,
|
| 1440 |
+
"learning_rate": 9.943520470575988e-06,
|
| 1441 |
+
"loss": 0.0071,
|
| 1442 |
+
"step": 23600
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 2.78,
|
| 1446 |
+
"learning_rate": 9.891465604747405e-06,
|
| 1447 |
+
"loss": 0.0028,
|
| 1448 |
+
"step": 23700
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 2.79,
|
| 1452 |
+
"learning_rate": 9.83941073891882e-06,
|
| 1453 |
+
"loss": 0.0083,
|
| 1454 |
+
"step": 23800
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 2.8,
|
| 1458 |
+
"learning_rate": 9.787355873090238e-06,
|
| 1459 |
+
"loss": 0.0028,
|
| 1460 |
+
"step": 23900
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 2.81,
|
| 1464 |
+
"learning_rate": 9.735301007261655e-06,
|
| 1465 |
+
"loss": 0.0016,
|
| 1466 |
+
"step": 24000
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 2.82,
|
| 1470 |
+
"learning_rate": 9.683246141433072e-06,
|
| 1471 |
+
"loss": 0.0117,
|
| 1472 |
+
"step": 24100
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 2.83,
|
| 1476 |
+
"learning_rate": 9.631191275604488e-06,
|
| 1477 |
+
"loss": 0.0124,
|
| 1478 |
+
"step": 24200
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 2.85,
|
| 1482 |
+
"learning_rate": 9.579136409775905e-06,
|
| 1483 |
+
"loss": 0.0033,
|
| 1484 |
+
"step": 24300
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 2.86,
|
| 1488 |
+
"learning_rate": 9.52708154394732e-06,
|
| 1489 |
+
"loss": 0.005,
|
| 1490 |
+
"step": 24400
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 2.87,
|
| 1494 |
+
"learning_rate": 9.475026678118739e-06,
|
| 1495 |
+
"loss": 0.0001,
|
| 1496 |
+
"step": 24500
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 2.88,
|
| 1500 |
+
"learning_rate": 9.422971812290155e-06,
|
| 1501 |
+
"loss": 0.0025,
|
| 1502 |
+
"step": 24600
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 2.89,
|
| 1506 |
+
"learning_rate": 9.370916946461572e-06,
|
| 1507 |
+
"loss": 0.0158,
|
| 1508 |
+
"step": 24700
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 2.9,
|
| 1512 |
+
"learning_rate": 9.318862080632987e-06,
|
| 1513 |
+
"loss": 0.0042,
|
| 1514 |
+
"step": 24800
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 2.92,
|
| 1518 |
+
"learning_rate": 9.266807214804404e-06,
|
| 1519 |
+
"loss": 0.0035,
|
| 1520 |
+
"step": 24900
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 2.93,
|
| 1524 |
+
"learning_rate": 9.214752348975822e-06,
|
| 1525 |
+
"loss": 0.0024,
|
| 1526 |
+
"step": 25000
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 2.94,
|
| 1530 |
+
"learning_rate": 9.162697483147239e-06,
|
| 1531 |
+
"loss": 0.0053,
|
| 1532 |
+
"step": 25100
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 2.95,
|
| 1536 |
+
"learning_rate": 9.11116316597694e-06,
|
| 1537 |
+
"loss": 0.0064,
|
| 1538 |
+
"step": 25200
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 2.96,
|
| 1542 |
+
"learning_rate": 9.059108300148357e-06,
|
| 1543 |
+
"loss": 0.0013,
|
| 1544 |
+
"step": 25300
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 2.97,
|
| 1548 |
+
"learning_rate": 9.007053434319773e-06,
|
| 1549 |
+
"loss": 0.0058,
|
| 1550 |
+
"step": 25400
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 2.99,
|
| 1554 |
+
"learning_rate": 8.95499856849119e-06,
|
| 1555 |
+
"loss": 0.0029,
|
| 1556 |
+
"step": 25500
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 3.0,
|
| 1560 |
+
"learning_rate": 8.902943702662607e-06,
|
| 1561 |
+
"loss": 0.0085,
|
| 1562 |
+
"step": 25600
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 3.0,
|
| 1566 |
+
"eval_f1": 0.9725928597187162,
|
| 1567 |
+
"eval_loss": 0.11609108746051788,
|
| 1568 |
+
"eval_runtime": 47.4182,
|
| 1569 |
+
"eval_samples_per_second": 721.684,
|
| 1570 |
+
"eval_steps_per_second": 90.218,
|
| 1571 |
+
"step": 25614
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 3.01,
|
| 1575 |
+
"learning_rate": 8.850888836834024e-06,
|
| 1576 |
+
"loss": 0.0008,
|
| 1577 |
+
"step": 25700
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 3.02,
|
| 1581 |
+
"learning_rate": 8.79883397100544e-06,
|
| 1582 |
+
"loss": 0.0086,
|
| 1583 |
+
"step": 25800
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 3.03,
|
| 1587 |
+
"learning_rate": 8.746779105176857e-06,
|
| 1588 |
+
"loss": 0.0001,
|
| 1589 |
+
"step": 25900
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 3.05,
|
| 1593 |
+
"learning_rate": 8.694724239348274e-06,
|
| 1594 |
+
"loss": 0.0,
|
| 1595 |
+
"step": 26000
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 3.06,
|
| 1599 |
+
"learning_rate": 8.642669373519691e-06,
|
| 1600 |
+
"loss": 0.0,
|
| 1601 |
+
"step": 26100
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 3.07,
|
| 1605 |
+
"learning_rate": 8.590614507691107e-06,
|
| 1606 |
+
"loss": 0.0,
|
| 1607 |
+
"step": 26200
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 3.08,
|
| 1611 |
+
"learning_rate": 8.538559641862524e-06,
|
| 1612 |
+
"loss": 0.0027,
|
| 1613 |
+
"step": 26300
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 3.09,
|
| 1617 |
+
"learning_rate": 8.486504776033941e-06,
|
| 1618 |
+
"loss": 0.0,
|
| 1619 |
+
"step": 26400
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 3.1,
|
| 1623 |
+
"learning_rate": 8.434449910205356e-06,
|
| 1624 |
+
"loss": 0.0043,
|
| 1625 |
+
"step": 26500
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 3.12,
|
| 1629 |
+
"learning_rate": 8.382395044376774e-06,
|
| 1630 |
+
"loss": 0.004,
|
| 1631 |
+
"step": 26600
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 3.13,
|
| 1635 |
+
"learning_rate": 8.33034017854819e-06,
|
| 1636 |
+
"loss": 0.0,
|
| 1637 |
+
"step": 26700
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 3.14,
|
| 1641 |
+
"learning_rate": 8.278285312719608e-06,
|
| 1642 |
+
"loss": 0.0016,
|
| 1643 |
+
"step": 26800
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 3.15,
|
| 1647 |
+
"learning_rate": 8.226230446891023e-06,
|
| 1648 |
+
"loss": 0.0,
|
| 1649 |
+
"step": 26900
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 3.16,
|
| 1653 |
+
"learning_rate": 8.17417558106244e-06,
|
| 1654 |
+
"loss": 0.0117,
|
| 1655 |
+
"step": 27000
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 3.17,
|
| 1659 |
+
"learning_rate": 8.122120715233858e-06,
|
| 1660 |
+
"loss": 0.0046,
|
| 1661 |
+
"step": 27100
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 3.19,
|
| 1665 |
+
"learning_rate": 8.07058639806356e-06,
|
| 1666 |
+
"loss": 0.0025,
|
| 1667 |
+
"step": 27200
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 3.2,
|
| 1671 |
+
"learning_rate": 8.018531532234976e-06,
|
| 1672 |
+
"loss": 0.0022,
|
| 1673 |
+
"step": 27300
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 3.21,
|
| 1677 |
+
"learning_rate": 7.966476666406393e-06,
|
| 1678 |
+
"loss": 0.011,
|
| 1679 |
+
"step": 27400
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 3.22,
|
| 1683 |
+
"learning_rate": 7.914421800577809e-06,
|
| 1684 |
+
"loss": 0.007,
|
| 1685 |
+
"step": 27500
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 3.23,
|
| 1689 |
+
"learning_rate": 7.862366934749226e-06,
|
| 1690 |
+
"loss": 0.0081,
|
| 1691 |
+
"step": 27600
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 3.24,
|
| 1695 |
+
"learning_rate": 7.810312068920643e-06,
|
| 1696 |
+
"loss": 0.0032,
|
| 1697 |
+
"step": 27700
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 3.26,
|
| 1701 |
+
"learning_rate": 7.75825720309206e-06,
|
| 1702 |
+
"loss": 0.0038,
|
| 1703 |
+
"step": 27800
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 3.27,
|
| 1707 |
+
"learning_rate": 7.706202337263476e-06,
|
| 1708 |
+
"loss": 0.003,
|
| 1709 |
+
"step": 27900
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 3.28,
|
| 1713 |
+
"learning_rate": 7.654147471434893e-06,
|
| 1714 |
+
"loss": 0.007,
|
| 1715 |
+
"step": 28000
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 3.29,
|
| 1719 |
+
"learning_rate": 7.602092605606309e-06,
|
| 1720 |
+
"loss": 0.0004,
|
| 1721 |
+
"step": 28100
|
| 1722 |
+
},
|
| 1723 |
+
{
|
| 1724 |
+
"epoch": 3.3,
|
| 1725 |
+
"learning_rate": 7.550037739777726e-06,
|
| 1726 |
+
"loss": 0.0039,
|
| 1727 |
+
"step": 28200
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 3.31,
|
| 1731 |
+
"learning_rate": 7.497982873949144e-06,
|
| 1732 |
+
"loss": 0.0,
|
| 1733 |
+
"step": 28300
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 3.33,
|
| 1737 |
+
"learning_rate": 7.44592800812056e-06,
|
| 1738 |
+
"loss": 0.003,
|
| 1739 |
+
"step": 28400
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 3.34,
|
| 1743 |
+
"learning_rate": 7.393873142291976e-06,
|
| 1744 |
+
"loss": 0.0018,
|
| 1745 |
+
"step": 28500
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 3.35,
|
| 1749 |
+
"learning_rate": 7.341818276463393e-06,
|
| 1750 |
+
"loss": 0.0011,
|
| 1751 |
+
"step": 28600
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 3.36,
|
| 1755 |
+
"learning_rate": 7.289763410634809e-06,
|
| 1756 |
+
"loss": 0.0013,
|
| 1757 |
+
"step": 28700
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 3.37,
|
| 1761 |
+
"learning_rate": 7.237708544806227e-06,
|
| 1762 |
+
"loss": 0.0028,
|
| 1763 |
+
"step": 28800
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 3.38,
|
| 1767 |
+
"learning_rate": 7.185653678977643e-06,
|
| 1768 |
+
"loss": 0.0029,
|
| 1769 |
+
"step": 28900
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 3.4,
|
| 1773 |
+
"learning_rate": 7.13359881314906e-06,
|
| 1774 |
+
"loss": 0.0009,
|
| 1775 |
+
"step": 29000
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 3.41,
|
| 1779 |
+
"learning_rate": 7.081543947320476e-06,
|
| 1780 |
+
"loss": 0.0,
|
| 1781 |
+
"step": 29100
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 3.42,
|
| 1785 |
+
"learning_rate": 7.029489081491892e-06,
|
| 1786 |
+
"loss": 0.0,
|
| 1787 |
+
"step": 29200
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 3.43,
|
| 1791 |
+
"learning_rate": 6.97743421566331e-06,
|
| 1792 |
+
"loss": 0.0029,
|
| 1793 |
+
"step": 29300
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 3.44,
|
| 1797 |
+
"learning_rate": 6.9253793498347266e-06,
|
| 1798 |
+
"loss": 0.0,
|
| 1799 |
+
"step": 29400
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 3.46,
|
| 1803 |
+
"learning_rate": 6.873324484006143e-06,
|
| 1804 |
+
"loss": 0.0043,
|
| 1805 |
+
"step": 29500
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 3.47,
|
| 1809 |
+
"learning_rate": 6.821269618177559e-06,
|
| 1810 |
+
"loss": 0.0041,
|
| 1811 |
+
"step": 29600
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 3.48,
|
| 1815 |
+
"learning_rate": 6.769214752348976e-06,
|
| 1816 |
+
"loss": 0.0,
|
| 1817 |
+
"step": 29700
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 3.49,
|
| 1821 |
+
"learning_rate": 6.7171598865203935e-06,
|
| 1822 |
+
"loss": 0.0035,
|
| 1823 |
+
"step": 29800
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 3.5,
|
| 1827 |
+
"learning_rate": 6.66510502069181e-06,
|
| 1828 |
+
"loss": 0.0028,
|
| 1829 |
+
"step": 29900
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 3.51,
|
| 1833 |
+
"learning_rate": 6.613050154863226e-06,
|
| 1834 |
+
"loss": 0.0031,
|
| 1835 |
+
"step": 30000
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 3.53,
|
| 1839 |
+
"learning_rate": 6.5609952890346425e-06,
|
| 1840 |
+
"loss": 0.003,
|
| 1841 |
+
"step": 30100
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 3.54,
|
| 1845 |
+
"learning_rate": 6.50894042320606e-06,
|
| 1846 |
+
"loss": 0.0013,
|
| 1847 |
+
"step": 30200
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 3.55,
|
| 1851 |
+
"learning_rate": 6.456885557377477e-06,
|
| 1852 |
+
"loss": 0.0087,
|
| 1853 |
+
"step": 30300
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 3.56,
|
| 1857 |
+
"learning_rate": 6.404830691548893e-06,
|
| 1858 |
+
"loss": 0.0034,
|
| 1859 |
+
"step": 30400
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 3.57,
|
| 1863 |
+
"learning_rate": 6.3527758257203095e-06,
|
| 1864 |
+
"loss": 0.0015,
|
| 1865 |
+
"step": 30500
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 3.58,
|
| 1869 |
+
"learning_rate": 6.300720959891727e-06,
|
| 1870 |
+
"loss": 0.0015,
|
| 1871 |
+
"step": 30600
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 3.6,
|
| 1875 |
+
"learning_rate": 6.248666094063143e-06,
|
| 1876 |
+
"loss": 0.0014,
|
| 1877 |
+
"step": 30700
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 3.61,
|
| 1881 |
+
"learning_rate": 6.196611228234559e-06,
|
| 1882 |
+
"loss": 0.0017,
|
| 1883 |
+
"step": 30800
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 3.62,
|
| 1887 |
+
"learning_rate": 6.1445563624059764e-06,
|
| 1888 |
+
"loss": 0.0024,
|
| 1889 |
+
"step": 30900
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 3.63,
|
| 1893 |
+
"learning_rate": 6.092501496577394e-06,
|
| 1894 |
+
"loss": 0.0046,
|
| 1895 |
+
"step": 31000
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 3.64,
|
| 1899 |
+
"learning_rate": 6.04044663074881e-06,
|
| 1900 |
+
"loss": 0.0048,
|
| 1901 |
+
"step": 31100
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 3.65,
|
| 1905 |
+
"learning_rate": 5.988391764920226e-06,
|
| 1906 |
+
"loss": 0.0,
|
| 1907 |
+
"step": 31200
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 3.67,
|
| 1911 |
+
"learning_rate": 5.9368574477499284e-06,
|
| 1912 |
+
"loss": 0.0005,
|
| 1913 |
+
"step": 31300
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 3.68,
|
| 1917 |
+
"learning_rate": 5.8848025819213464e-06,
|
| 1918 |
+
"loss": 0.0,
|
| 1919 |
+
"step": 31400
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 3.69,
|
| 1923 |
+
"learning_rate": 5.832747716092763e-06,
|
| 1924 |
+
"loss": 0.0,
|
| 1925 |
+
"step": 31500
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 3.7,
|
| 1929 |
+
"learning_rate": 5.781213398922465e-06,
|
| 1930 |
+
"loss": 0.0054,
|
| 1931 |
+
"step": 31600
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 3.71,
|
| 1935 |
+
"learning_rate": 5.729158533093881e-06,
|
| 1936 |
+
"loss": 0.0066,
|
| 1937 |
+
"step": 31700
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 3.72,
|
| 1941 |
+
"learning_rate": 5.677103667265298e-06,
|
| 1942 |
+
"loss": 0.0017,
|
| 1943 |
+
"step": 31800
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 3.74,
|
| 1947 |
+
"learning_rate": 5.625048801436715e-06,
|
| 1948 |
+
"loss": 0.0017,
|
| 1949 |
+
"step": 31900
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 3.75,
|
| 1953 |
+
"learning_rate": 5.572993935608132e-06,
|
| 1954 |
+
"loss": 0.003,
|
| 1955 |
+
"step": 32000
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 3.76,
|
| 1959 |
+
"learning_rate": 5.520939069779548e-06,
|
| 1960 |
+
"loss": 0.0,
|
| 1961 |
+
"step": 32100
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 3.77,
|
| 1965 |
+
"learning_rate": 5.4688842039509646e-06,
|
| 1966 |
+
"loss": 0.0011,
|
| 1967 |
+
"step": 32200
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 3.78,
|
| 1971 |
+
"learning_rate": 5.417349886780668e-06,
|
| 1972 |
+
"loss": 0.0018,
|
| 1973 |
+
"step": 32300
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 3.79,
|
| 1977 |
+
"learning_rate": 5.365295020952084e-06,
|
| 1978 |
+
"loss": 0.0023,
|
| 1979 |
+
"step": 32400
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 3.81,
|
| 1983 |
+
"learning_rate": 5.3132401551235e-06,
|
| 1984 |
+
"loss": 0.0,
|
| 1985 |
+
"step": 32500
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 3.82,
|
| 1989 |
+
"learning_rate": 5.261185289294917e-06,
|
| 1990 |
+
"loss": 0.0017,
|
| 1991 |
+
"step": 32600
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 3.83,
|
| 1995 |
+
"learning_rate": 5.209130423466335e-06,
|
| 1996 |
+
"loss": 0.0,
|
| 1997 |
+
"step": 32700
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 3.84,
|
| 2001 |
+
"learning_rate": 5.157075557637751e-06,
|
| 2002 |
+
"loss": 0.0001,
|
| 2003 |
+
"step": 32800
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 3.85,
|
| 2007 |
+
"learning_rate": 5.105020691809167e-06,
|
| 2008 |
+
"loss": 0.0019,
|
| 2009 |
+
"step": 32900
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 3.86,
|
| 2013 |
+
"learning_rate": 5.0529658259805835e-06,
|
| 2014 |
+
"loss": 0.0042,
|
| 2015 |
+
"step": 33000
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 3.88,
|
| 2019 |
+
"learning_rate": 5.0009109601520015e-06,
|
| 2020 |
+
"loss": 0.0013,
|
| 2021 |
+
"step": 33100
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 3.89,
|
| 2025 |
+
"learning_rate": 4.948856094323417e-06,
|
| 2026 |
+
"loss": 0.0009,
|
| 2027 |
+
"step": 33200
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 3.9,
|
| 2031 |
+
"learning_rate": 4.896801228494834e-06,
|
| 2032 |
+
"loss": 0.0021,
|
| 2033 |
+
"step": 33300
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 3.91,
|
| 2037 |
+
"learning_rate": 4.8447463626662505e-06,
|
| 2038 |
+
"loss": 0.0031,
|
| 2039 |
+
"step": 33400
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 3.92,
|
| 2043 |
+
"learning_rate": 4.792691496837668e-06,
|
| 2044 |
+
"loss": 0.0019,
|
| 2045 |
+
"step": 33500
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 3.94,
|
| 2049 |
+
"learning_rate": 4.740636631009084e-06,
|
| 2050 |
+
"loss": 0.0,
|
| 2051 |
+
"step": 33600
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 3.95,
|
| 2055 |
+
"learning_rate": 4.6885817651805e-06,
|
| 2056 |
+
"loss": 0.0,
|
| 2057 |
+
"step": 33700
|
| 2058 |
+
},
|
| 2059 |
+
{
|
| 2060 |
+
"epoch": 3.96,
|
| 2061 |
+
"learning_rate": 4.6365268993519175e-06,
|
| 2062 |
+
"loss": 0.0,
|
| 2063 |
+
"step": 33800
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 3.97,
|
| 2067 |
+
"learning_rate": 4.584472033523334e-06,
|
| 2068 |
+
"loss": 0.0088,
|
| 2069 |
+
"step": 33900
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 3.98,
|
| 2073 |
+
"learning_rate": 4.53241716769475e-06,
|
| 2074 |
+
"loss": 0.0021,
|
| 2075 |
+
"step": 34000
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 3.99,
|
| 2079 |
+
"learning_rate": 4.480362301866167e-06,
|
| 2080 |
+
"loss": 0.0015,
|
| 2081 |
+
"step": 34100
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 4.0,
|
| 2085 |
+
"eval_f1": 0.9515538902358385,
|
| 2086 |
+
"eval_loss": 0.25566887855529785,
|
| 2087 |
+
"eval_runtime": 47.5026,
|
| 2088 |
+
"eval_samples_per_second": 720.403,
|
| 2089 |
+
"eval_steps_per_second": 90.058,
|
| 2090 |
+
"step": 34152
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 4.01,
|
| 2094 |
+
"learning_rate": 4.428307436037584e-06,
|
| 2095 |
+
"loss": 0.0,
|
| 2096 |
+
"step": 34200
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 4.02,
|
| 2100 |
+
"learning_rate": 4.376773118867287e-06,
|
| 2101 |
+
"loss": 0.0004,
|
| 2102 |
+
"step": 34300
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 4.03,
|
| 2106 |
+
"learning_rate": 4.324718253038703e-06,
|
| 2107 |
+
"loss": 0.0,
|
| 2108 |
+
"step": 34400
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 4.04,
|
| 2112 |
+
"learning_rate": 4.27266338721012e-06,
|
| 2113 |
+
"loss": 0.0,
|
| 2114 |
+
"step": 34500
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 4.05,
|
| 2118 |
+
"learning_rate": 4.2206085213815364e-06,
|
| 2119 |
+
"loss": 0.0,
|
| 2120 |
+
"step": 34600
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 4.06,
|
| 2124 |
+
"learning_rate": 4.168553655552954e-06,
|
| 2125 |
+
"loss": 0.0,
|
| 2126 |
+
"step": 34700
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 4.08,
|
| 2130 |
+
"learning_rate": 4.11649878972437e-06,
|
| 2131 |
+
"loss": 0.0,
|
| 2132 |
+
"step": 34800
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 4.09,
|
| 2136 |
+
"learning_rate": 4.064443923895787e-06,
|
| 2137 |
+
"loss": 0.0,
|
| 2138 |
+
"step": 34900
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 4.1,
|
| 2142 |
+
"learning_rate": 4.012389058067203e-06,
|
| 2143 |
+
"loss": 0.0006,
|
| 2144 |
+
"step": 35000
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 4.11,
|
| 2148 |
+
"learning_rate": 3.96033419223862e-06,
|
| 2149 |
+
"loss": 0.0024,
|
| 2150 |
+
"step": 35100
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 4.12,
|
| 2154 |
+
"learning_rate": 3.908279326410037e-06,
|
| 2155 |
+
"loss": 0.0024,
|
| 2156 |
+
"step": 35200
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 4.13,
|
| 2160 |
+
"learning_rate": 3.856224460581453e-06,
|
| 2161 |
+
"loss": 0.0,
|
| 2162 |
+
"step": 35300
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 4.15,
|
| 2166 |
+
"learning_rate": 3.80416959475287e-06,
|
| 2167 |
+
"loss": 0.0,
|
| 2168 |
+
"step": 35400
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 4.16,
|
| 2172 |
+
"learning_rate": 3.7521147289242867e-06,
|
| 2173 |
+
"loss": 0.0,
|
| 2174 |
+
"step": 35500
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 4.17,
|
| 2178 |
+
"learning_rate": 3.700059863095703e-06,
|
| 2179 |
+
"loss": 0.0,
|
| 2180 |
+
"step": 35600
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 4.18,
|
| 2184 |
+
"learning_rate": 3.64800499726712e-06,
|
| 2185 |
+
"loss": 0.0,
|
| 2186 |
+
"step": 35700
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 4.19,
|
| 2190 |
+
"learning_rate": 3.5959501314385365e-06,
|
| 2191 |
+
"loss": 0.0,
|
| 2192 |
+
"step": 35800
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 4.2,
|
| 2196 |
+
"learning_rate": 3.5438952656099537e-06,
|
| 2197 |
+
"loss": 0.0013,
|
| 2198 |
+
"step": 35900
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 4.22,
|
| 2202 |
+
"learning_rate": 3.49184039978137e-06,
|
| 2203 |
+
"loss": 0.002,
|
| 2204 |
+
"step": 36000
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 4.23,
|
| 2208 |
+
"learning_rate": 3.4397855339527863e-06,
|
| 2209 |
+
"loss": 0.0,
|
| 2210 |
+
"step": 36100
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 4.24,
|
| 2214 |
+
"learning_rate": 3.3877306681242035e-06,
|
| 2215 |
+
"loss": 0.0015,
|
| 2216 |
+
"step": 36200
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 4.25,
|
| 2220 |
+
"learning_rate": 3.3356758022956198e-06,
|
| 2221 |
+
"loss": 0.0,
|
| 2222 |
+
"step": 36300
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 4.26,
|
| 2226 |
+
"learning_rate": 3.283620936467036e-06,
|
| 2227 |
+
"loss": 0.0,
|
| 2228 |
+
"step": 36400
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 4.27,
|
| 2232 |
+
"learning_rate": 3.2315660706384533e-06,
|
| 2233 |
+
"loss": 0.0021,
|
| 2234 |
+
"step": 36500
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 4.29,
|
| 2238 |
+
"learning_rate": 3.1795112048098696e-06,
|
| 2239 |
+
"loss": 0.0,
|
| 2240 |
+
"step": 36600
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 4.3,
|
| 2244 |
+
"learning_rate": 3.1274563389812867e-06,
|
| 2245 |
+
"loss": 0.0,
|
| 2246 |
+
"step": 36700
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 4.31,
|
| 2250 |
+
"learning_rate": 3.075401473152703e-06,
|
| 2251 |
+
"loss": 0.0014,
|
| 2252 |
+
"step": 36800
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 4.32,
|
| 2256 |
+
"learning_rate": 3.02334660732412e-06,
|
| 2257 |
+
"loss": 0.0037,
|
| 2258 |
+
"step": 36900
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 4.33,
|
| 2262 |
+
"learning_rate": 2.9712917414955366e-06,
|
| 2263 |
+
"loss": 0.0,
|
| 2264 |
+
"step": 37000
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 4.35,
|
| 2268 |
+
"learning_rate": 2.9192368756669533e-06,
|
| 2269 |
+
"loss": 0.0013,
|
| 2270 |
+
"step": 37100
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 4.36,
|
| 2274 |
+
"learning_rate": 2.8677025584966555e-06,
|
| 2275 |
+
"loss": 0.0011,
|
| 2276 |
+
"step": 37200
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 4.37,
|
| 2280 |
+
"learning_rate": 2.8156476926680727e-06,
|
| 2281 |
+
"loss": 0.0031,
|
| 2282 |
+
"step": 37300
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 4.38,
|
| 2286 |
+
"learning_rate": 2.763592826839489e-06,
|
| 2287 |
+
"loss": 0.0012,
|
| 2288 |
+
"step": 37400
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 4.39,
|
| 2292 |
+
"learning_rate": 2.7115379610109053e-06,
|
| 2293 |
+
"loss": 0.0,
|
| 2294 |
+
"step": 37500
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 4.4,
|
| 2298 |
+
"learning_rate": 2.6594830951823225e-06,
|
| 2299 |
+
"loss": 0.0,
|
| 2300 |
+
"step": 37600
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 4.42,
|
| 2304 |
+
"learning_rate": 2.6074282293537388e-06,
|
| 2305 |
+
"loss": 0.0,
|
| 2306 |
+
"step": 37700
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 4.43,
|
| 2310 |
+
"learning_rate": 2.555373363525156e-06,
|
| 2311 |
+
"loss": 0.0012,
|
| 2312 |
+
"step": 37800
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 4.44,
|
| 2316 |
+
"learning_rate": 2.5033184976965723e-06,
|
| 2317 |
+
"loss": 0.0,
|
| 2318 |
+
"step": 37900
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 4.45,
|
| 2322 |
+
"learning_rate": 2.451263631867989e-06,
|
| 2323 |
+
"loss": 0.0033,
|
| 2324 |
+
"step": 38000
|
| 2325 |
+
},
|
| 2326 |
+
{
|
| 2327 |
+
"epoch": 4.46,
|
| 2328 |
+
"learning_rate": 2.3992087660394057e-06,
|
| 2329 |
+
"loss": 0.0,
|
| 2330 |
+
"step": 38100
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 4.47,
|
| 2334 |
+
"learning_rate": 2.3471539002108225e-06,
|
| 2335 |
+
"loss": 0.0,
|
| 2336 |
+
"step": 38200
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 4.49,
|
| 2340 |
+
"learning_rate": 2.2950990343822392e-06,
|
| 2341 |
+
"loss": 0.0019,
|
| 2342 |
+
"step": 38300
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 4.5,
|
| 2346 |
+
"learning_rate": 2.243044168553656e-06,
|
| 2347 |
+
"loss": 0.0,
|
| 2348 |
+
"step": 38400
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 4.51,
|
| 2352 |
+
"learning_rate": 2.1909893027250727e-06,
|
| 2353 |
+
"loss": 0.0005,
|
| 2354 |
+
"step": 38500
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 4.52,
|
| 2358 |
+
"learning_rate": 2.138934436896489e-06,
|
| 2359 |
+
"loss": 0.0018,
|
| 2360 |
+
"step": 38600
|
| 2361 |
+
},
|
| 2362 |
+
{
|
| 2363 |
+
"epoch": 4.53,
|
| 2364 |
+
"learning_rate": 2.0868795710679058e-06,
|
| 2365 |
+
"loss": 0.0,
|
| 2366 |
+
"step": 38700
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"epoch": 4.54,
|
| 2370 |
+
"learning_rate": 2.0348247052393225e-06,
|
| 2371 |
+
"loss": 0.0018,
|
| 2372 |
+
"step": 38800
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 4.56,
|
| 2376 |
+
"learning_rate": 1.9827698394107393e-06,
|
| 2377 |
+
"loss": 0.0,
|
| 2378 |
+
"step": 38900
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 4.57,
|
| 2382 |
+
"learning_rate": 1.9307149735821556e-06,
|
| 2383 |
+
"loss": 0.0,
|
| 2384 |
+
"step": 39000
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 4.58,
|
| 2388 |
+
"learning_rate": 1.8786601077535723e-06,
|
| 2389 |
+
"loss": 0.0,
|
| 2390 |
+
"step": 39100
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 4.59,
|
| 2394 |
+
"learning_rate": 1.826605241924989e-06,
|
| 2395 |
+
"loss": 0.0,
|
| 2396 |
+
"step": 39200
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 4.6,
|
| 2400 |
+
"learning_rate": 1.7745503760964058e-06,
|
| 2401 |
+
"loss": 0.0,
|
| 2402 |
+
"step": 39300
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 4.61,
|
| 2406 |
+
"learning_rate": 1.7224955102678223e-06,
|
| 2407 |
+
"loss": 0.0,
|
| 2408 |
+
"step": 39400
|
| 2409 |
+
},
|
| 2410 |
+
{
|
| 2411 |
+
"epoch": 4.63,
|
| 2412 |
+
"learning_rate": 1.670440644439239e-06,
|
| 2413 |
+
"loss": 0.0,
|
| 2414 |
+
"step": 39500
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 4.64,
|
| 2418 |
+
"learning_rate": 1.6189063272689415e-06,
|
| 2419 |
+
"loss": 0.0018,
|
| 2420 |
+
"step": 39600
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 4.65,
|
| 2424 |
+
"learning_rate": 1.5668514614403582e-06,
|
| 2425 |
+
"loss": 0.0,
|
| 2426 |
+
"step": 39700
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 4.66,
|
| 2430 |
+
"learning_rate": 1.514796595611775e-06,
|
| 2431 |
+
"loss": 0.0011,
|
| 2432 |
+
"step": 39800
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 4.67,
|
| 2436 |
+
"learning_rate": 1.4627417297831917e-06,
|
| 2437 |
+
"loss": 0.0,
|
| 2438 |
+
"step": 39900
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 4.68,
|
| 2442 |
+
"learning_rate": 1.4106868639546085e-06,
|
| 2443 |
+
"loss": 0.0,
|
| 2444 |
+
"step": 40000
|
| 2445 |
+
},
|
| 2446 |
+
{
|
| 2447 |
+
"epoch": 4.7,
|
| 2448 |
+
"learning_rate": 1.3586319981260248e-06,
|
| 2449 |
+
"loss": 0.0,
|
| 2450 |
+
"step": 40100
|
| 2451 |
+
},
|
| 2452 |
+
{
|
| 2453 |
+
"epoch": 4.71,
|
| 2454 |
+
"learning_rate": 1.3065771322974415e-06,
|
| 2455 |
+
"loss": 0.0013,
|
| 2456 |
+
"step": 40200
|
| 2457 |
+
},
|
| 2458 |
+
{
|
| 2459 |
+
"epoch": 4.72,
|
| 2460 |
+
"learning_rate": 1.2545222664688583e-06,
|
| 2461 |
+
"loss": 0.0,
|
| 2462 |
+
"step": 40300
|
| 2463 |
+
},
|
| 2464 |
+
{
|
| 2465 |
+
"epoch": 4.73,
|
| 2466 |
+
"learning_rate": 1.202467400640275e-06,
|
| 2467 |
+
"loss": 0.0,
|
| 2468 |
+
"step": 40400
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 4.74,
|
| 2472 |
+
"learning_rate": 1.1504125348116917e-06,
|
| 2473 |
+
"loss": 0.0,
|
| 2474 |
+
"step": 40500
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 4.76,
|
| 2478 |
+
"learning_rate": 1.0983576689831083e-06,
|
| 2479 |
+
"loss": 0.0,
|
| 2480 |
+
"step": 40600
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 4.77,
|
| 2484 |
+
"learning_rate": 1.046302803154525e-06,
|
| 2485 |
+
"loss": 0.0,
|
| 2486 |
+
"step": 40700
|
| 2487 |
+
},
|
| 2488 |
+
{
|
| 2489 |
+
"epoch": 4.78,
|
| 2490 |
+
"learning_rate": 9.942479373259415e-07,
|
| 2491 |
+
"loss": 0.0012,
|
| 2492 |
+
"step": 40800
|
| 2493 |
+
},
|
| 2494 |
+
{
|
| 2495 |
+
"epoch": 4.79,
|
| 2496 |
+
"learning_rate": 9.421930714973583e-07,
|
| 2497 |
+
"loss": 0.0012,
|
| 2498 |
+
"step": 40900
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 4.8,
|
| 2502 |
+
"learning_rate": 8.901382056687749e-07,
|
| 2503 |
+
"loss": 0.0,
|
| 2504 |
+
"step": 41000
|
| 2505 |
+
},
|
| 2506 |
+
{
|
| 2507 |
+
"epoch": 4.81,
|
| 2508 |
+
"learning_rate": 8.380833398401917e-07,
|
| 2509 |
+
"loss": 0.0013,
|
| 2510 |
+
"step": 41100
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 4.83,
|
| 2514 |
+
"learning_rate": 7.860284740116084e-07,
|
| 2515 |
+
"loss": 0.0,
|
| 2516 |
+
"step": 41200
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 4.84,
|
| 2520 |
+
"learning_rate": 7.339736081830249e-07,
|
| 2521 |
+
"loss": 0.0,
|
| 2522 |
+
"step": 41300
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 4.85,
|
| 2526 |
+
"learning_rate": 6.819187423544417e-07,
|
| 2527 |
+
"loss": 0.0,
|
| 2528 |
+
"step": 41400
|
| 2529 |
+
},
|
| 2530 |
+
{
|
| 2531 |
+
"epoch": 4.86,
|
| 2532 |
+
"learning_rate": 6.298638765258583e-07,
|
| 2533 |
+
"loss": 0.0001,
|
| 2534 |
+
"step": 41500
|
| 2535 |
+
},
|
| 2536 |
+
{
|
| 2537 |
+
"epoch": 4.87,
|
| 2538 |
+
"learning_rate": 5.77809010697275e-07,
|
| 2539 |
+
"loss": 0.0,
|
| 2540 |
+
"step": 41600
|
| 2541 |
+
},
|
| 2542 |
+
{
|
| 2543 |
+
"epoch": 4.88,
|
| 2544 |
+
"learning_rate": 5.257541448686916e-07,
|
| 2545 |
+
"loss": 0.0008,
|
| 2546 |
+
"step": 41700
|
| 2547 |
+
},
|
| 2548 |
+
{
|
| 2549 |
+
"epoch": 4.9,
|
| 2550 |
+
"learning_rate": 4.736992790401083e-07,
|
| 2551 |
+
"loss": 0.0,
|
| 2552 |
+
"step": 41800
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 4.91,
|
| 2556 |
+
"learning_rate": 4.21644413211525e-07,
|
| 2557 |
+
"loss": 0.0,
|
| 2558 |
+
"step": 41900
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 4.92,
|
| 2562 |
+
"learning_rate": 3.6958954738294165e-07,
|
| 2563 |
+
"loss": 0.0,
|
| 2564 |
+
"step": 42000
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 4.93,
|
| 2568 |
+
"learning_rate": 3.1753468155435834e-07,
|
| 2569 |
+
"loss": 0.0,
|
| 2570 |
+
"step": 42100
|
| 2571 |
+
},
|
| 2572 |
+
{
|
| 2573 |
+
"epoch": 4.94,
|
| 2574 |
+
"learning_rate": 2.65479815725775e-07,
|
| 2575 |
+
"loss": 0.0015,
|
| 2576 |
+
"step": 42200
|
| 2577 |
+
},
|
| 2578 |
+
{
|
| 2579 |
+
"epoch": 4.95,
|
| 2580 |
+
"learning_rate": 2.1342494989719166e-07,
|
| 2581 |
+
"loss": 0.0,
|
| 2582 |
+
"step": 42300
|
| 2583 |
+
},
|
| 2584 |
+
{
|
| 2585 |
+
"epoch": 4.97,
|
| 2586 |
+
"learning_rate": 1.6137008406860833e-07,
|
| 2587 |
+
"loss": 0.0,
|
| 2588 |
+
"step": 42400
|
| 2589 |
+
},
|
| 2590 |
+
{
|
| 2591 |
+
"epoch": 4.98,
|
| 2592 |
+
"learning_rate": 1.0931521824002499e-07,
|
| 2593 |
+
"loss": 0.0,
|
| 2594 |
+
"step": 42500
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 4.99,
|
| 2598 |
+
"learning_rate": 5.7260352411441666e-08,
|
| 2599 |
+
"loss": 0.0,
|
| 2600 |
+
"step": 42600
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 5.0,
|
| 2604 |
+
"eval_f1": 0.9649930701479859,
|
| 2605 |
+
"eval_loss": 0.22864048182964325,
|
| 2606 |
+
"eval_runtime": 47.3601,
|
| 2607 |
+
"eval_samples_per_second": 722.57,
|
| 2608 |
+
"eval_steps_per_second": 90.329,
|
| 2609 |
+
"step": 42690
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"epoch": 5.0,
|
| 2613 |
+
"step": 42690,
|
| 2614 |
+
"total_flos": 8.986242362554368e+16,
|
| 2615 |
+
"train_loss": 0.023754316836012614,
|
| 2616 |
+
"train_runtime": 9371.3538,
|
| 2617 |
+
"train_samples_per_second": 145.781,
|
| 2618 |
+
"train_steps_per_second": 4.555
|
| 2619 |
+
}
|
| 2620 |
+
],
|
| 2621 |
+
"max_steps": 42690,
|
| 2622 |
+
"num_train_epochs": 5,
|
| 2623 |
+
"total_flos": 8.986242362554368e+16,
|
| 2624 |
+
"trial_name": null,
|
| 2625 |
+
"trial_params": null
|
| 2626 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:18ba0e007e1225cae19bb9520a0ff430f110d0e17d262a656de7dfef78d4f05c
|
| 3 |
+
size 3759
|