Hesam Ossanloo
commited on
Commit
·
02fa43f
1
Parent(s):
1e49e46
add NO model
Browse files- model/checkpoint-52560/config.json +43 -0
- model/checkpoint-52560/optimizer.pt +3 -0
- model/checkpoint-52560/pytorch_model.bin +3 -0
- model/checkpoint-52560/rng_state.pth +3 -0
- model/checkpoint-52560/scheduler.pt +3 -0
- model/checkpoint-52560/special_tokens_map.json +7 -0
- model/checkpoint-52560/tokenizer.json +0 -0
- model/checkpoint-52560/tokenizer_config.json +15 -0
- model/checkpoint-52560/trainer_state.json +682 -0
- model/checkpoint-52560/training_args.bin +3 -0
- model/checkpoint-52560/vocab.txt +0 -0
model/checkpoint-52560/config.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "emilyalsentzer/Bio_ClinicalBERT",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForTokenClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "O",
|
| 13 |
+
"1": "B-problem",
|
| 14 |
+
"2": "I-problem",
|
| 15 |
+
"3": "B-test",
|
| 16 |
+
"4": "I-test",
|
| 17 |
+
"5": "B-treatment",
|
| 18 |
+
"6": "I-treatment"
|
| 19 |
+
},
|
| 20 |
+
"initializer_range": 0.02,
|
| 21 |
+
"intermediate_size": 3072,
|
| 22 |
+
"label2id": {
|
| 23 |
+
"B-problem": 1,
|
| 24 |
+
"B-test": 3,
|
| 25 |
+
"B-treatment": 5,
|
| 26 |
+
"I-problem": 2,
|
| 27 |
+
"I-test": 4,
|
| 28 |
+
"I-treatment": 6,
|
| 29 |
+
"O": 0
|
| 30 |
+
},
|
| 31 |
+
"layer_norm_eps": 1e-12,
|
| 32 |
+
"max_position_embeddings": 512,
|
| 33 |
+
"model_type": "bert",
|
| 34 |
+
"num_attention_heads": 12,
|
| 35 |
+
"num_hidden_layers": 12,
|
| 36 |
+
"pad_token_id": 0,
|
| 37 |
+
"position_embedding_type": "absolute",
|
| 38 |
+
"torch_dtype": "float32",
|
| 39 |
+
"transformers_version": "4.31.0",
|
| 40 |
+
"type_vocab_size": 2,
|
| 41 |
+
"use_cache": true,
|
| 42 |
+
"vocab_size": 28996
|
| 43 |
+
}
|
model/checkpoint-52560/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:124f9ccf0dc7f6bb3a321ebb853c000a7ab5dda74009a5ae04b4a3845c56e538
|
| 3 |
+
size 861917189
|
model/checkpoint-52560/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:06a63d8aa6fd550bdabbd93b0fecf83a9d812c458179a037e3cb1e24c7e83eb7
|
| 3 |
+
size 430967977
|
model/checkpoint-52560/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7632877891bae50972b1d3c73300d1b837ae7d6c74df887ed4123c5f0b188516
|
| 3 |
+
size 14575
|
model/checkpoint-52560/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9f67393655e6c6060fa3988a7592744ce5fc39e0bf4ea462c179e144b19be72f
|
| 3 |
+
size 627
|
model/checkpoint-52560/special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
model/checkpoint-52560/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model/checkpoint-52560/tokenizer_config.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"clean_up_tokenization_spaces": true,
|
| 3 |
+
"cls_token": "[CLS]",
|
| 4 |
+
"do_basic_tokenize": true,
|
| 5 |
+
"do_lower_case": true,
|
| 6 |
+
"mask_token": "[MASK]",
|
| 7 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 8 |
+
"never_split": null,
|
| 9 |
+
"pad_token": "[PAD]",
|
| 10 |
+
"sep_token": "[SEP]",
|
| 11 |
+
"strip_accents": null,
|
| 12 |
+
"tokenize_chinese_chars": true,
|
| 13 |
+
"tokenizer_class": "BertTokenizer",
|
| 14 |
+
"unk_token": "[UNK]"
|
| 15 |
+
}
|
model/checkpoint-52560/trainer_state.json
ADDED
|
@@ -0,0 +1,682 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 52560,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 1.9809741248097414e-05,
|
| 13 |
+
"loss": 0.6459,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 1.9619482496194826e-05,
|
| 19 |
+
"loss": 0.4121,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.09,
|
| 24 |
+
"learning_rate": 1.9429223744292238e-05,
|
| 25 |
+
"loss": 0.3439,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.11,
|
| 30 |
+
"learning_rate": 1.923896499238965e-05,
|
| 31 |
+
"loss": 0.3113,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.14,
|
| 36 |
+
"learning_rate": 1.9048706240487066e-05,
|
| 37 |
+
"loss": 0.2776,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.17,
|
| 42 |
+
"learning_rate": 1.8858447488584478e-05,
|
| 43 |
+
"loss": 0.2665,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.2,
|
| 48 |
+
"learning_rate": 1.866818873668189e-05,
|
| 49 |
+
"loss": 0.2538,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.23,
|
| 54 |
+
"learning_rate": 1.8477929984779303e-05,
|
| 55 |
+
"loss": 0.2583,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.26,
|
| 60 |
+
"learning_rate": 1.8287671232876715e-05,
|
| 61 |
+
"loss": 0.2515,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.29,
|
| 66 |
+
"learning_rate": 1.8097412480974127e-05,
|
| 67 |
+
"loss": 0.218,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.31,
|
| 72 |
+
"learning_rate": 1.790715372907154e-05,
|
| 73 |
+
"loss": 0.2041,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.34,
|
| 78 |
+
"learning_rate": 1.771689497716895e-05,
|
| 79 |
+
"loss": 0.1916,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.37,
|
| 84 |
+
"learning_rate": 1.7526636225266364e-05,
|
| 85 |
+
"loss": 0.1879,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.4,
|
| 90 |
+
"learning_rate": 1.7336377473363776e-05,
|
| 91 |
+
"loss": 0.2072,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.43,
|
| 96 |
+
"learning_rate": 1.7146118721461188e-05,
|
| 97 |
+
"loss": 0.2066,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.46,
|
| 102 |
+
"learning_rate": 1.69558599695586e-05,
|
| 103 |
+
"loss": 0.2069,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.49,
|
| 108 |
+
"learning_rate": 1.6765601217656012e-05,
|
| 109 |
+
"loss": 0.1913,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.51,
|
| 114 |
+
"learning_rate": 1.6575342465753425e-05,
|
| 115 |
+
"loss": 0.1942,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.54,
|
| 120 |
+
"learning_rate": 1.6385083713850837e-05,
|
| 121 |
+
"loss": 0.171,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.57,
|
| 126 |
+
"learning_rate": 1.619482496194825e-05,
|
| 127 |
+
"loss": 0.1609,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.6,
|
| 132 |
+
"learning_rate": 1.6004566210045664e-05,
|
| 133 |
+
"loss": 0.1736,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.63,
|
| 138 |
+
"learning_rate": 1.5814307458143077e-05,
|
| 139 |
+
"loss": 0.1696,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.66,
|
| 144 |
+
"learning_rate": 1.562404870624049e-05,
|
| 145 |
+
"loss": 0.1559,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.68,
|
| 150 |
+
"learning_rate": 1.54337899543379e-05,
|
| 151 |
+
"loss": 0.1577,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.71,
|
| 156 |
+
"learning_rate": 1.5243531202435313e-05,
|
| 157 |
+
"loss": 0.151,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.74,
|
| 162 |
+
"learning_rate": 1.5053272450532725e-05,
|
| 163 |
+
"loss": 0.147,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.77,
|
| 168 |
+
"learning_rate": 1.4863013698630138e-05,
|
| 169 |
+
"loss": 0.145,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.8,
|
| 174 |
+
"learning_rate": 1.467275494672755e-05,
|
| 175 |
+
"loss": 0.163,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.83,
|
| 180 |
+
"learning_rate": 1.4482496194824962e-05,
|
| 181 |
+
"loss": 0.162,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.86,
|
| 186 |
+
"learning_rate": 1.4292237442922376e-05,
|
| 187 |
+
"loss": 0.1484,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.88,
|
| 192 |
+
"learning_rate": 1.4101978691019788e-05,
|
| 193 |
+
"loss": 0.1385,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.91,
|
| 198 |
+
"learning_rate": 1.39117199391172e-05,
|
| 199 |
+
"loss": 0.1436,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.94,
|
| 204 |
+
"learning_rate": 1.3721461187214612e-05,
|
| 205 |
+
"loss": 0.1292,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.97,
|
| 210 |
+
"learning_rate": 1.3531202435312025e-05,
|
| 211 |
+
"loss": 0.1482,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.0,
|
| 216 |
+
"learning_rate": 1.3340943683409437e-05,
|
| 217 |
+
"loss": 0.1274,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.0,
|
| 222 |
+
"eval_accuracy": 0.9567564969397891,
|
| 223 |
+
"eval_f1": 0.8423064061529643,
|
| 224 |
+
"eval_loss": 0.14508096873760223,
|
| 225 |
+
"eval_precision": 0.8260095568256348,
|
| 226 |
+
"eval_recall": 0.8592592592592593,
|
| 227 |
+
"eval_runtime": 35.0181,
|
| 228 |
+
"eval_samples_per_second": 250.156,
|
| 229 |
+
"eval_steps_per_second": 62.539,
|
| 230 |
+
"step": 17520
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.03,
|
| 234 |
+
"learning_rate": 1.3150684931506849e-05,
|
| 235 |
+
"loss": 0.1016,
|
| 236 |
+
"step": 18000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.06,
|
| 240 |
+
"learning_rate": 1.2960426179604261e-05,
|
| 241 |
+
"loss": 0.0952,
|
| 242 |
+
"step": 18500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.08,
|
| 246 |
+
"learning_rate": 1.2770167427701677e-05,
|
| 247 |
+
"loss": 0.0896,
|
| 248 |
+
"step": 19000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.11,
|
| 252 |
+
"learning_rate": 1.2579908675799089e-05,
|
| 253 |
+
"loss": 0.1053,
|
| 254 |
+
"step": 19500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.14,
|
| 258 |
+
"learning_rate": 1.2389649923896501e-05,
|
| 259 |
+
"loss": 0.0977,
|
| 260 |
+
"step": 20000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.17,
|
| 264 |
+
"learning_rate": 1.2199391171993913e-05,
|
| 265 |
+
"loss": 0.0992,
|
| 266 |
+
"step": 20500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.2,
|
| 270 |
+
"learning_rate": 1.2009132420091326e-05,
|
| 271 |
+
"loss": 0.0977,
|
| 272 |
+
"step": 21000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.23,
|
| 276 |
+
"learning_rate": 1.1818873668188738e-05,
|
| 277 |
+
"loss": 0.0809,
|
| 278 |
+
"step": 21500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.26,
|
| 282 |
+
"learning_rate": 1.162861491628615e-05,
|
| 283 |
+
"loss": 0.09,
|
| 284 |
+
"step": 22000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.28,
|
| 288 |
+
"learning_rate": 1.1438356164383562e-05,
|
| 289 |
+
"loss": 0.0993,
|
| 290 |
+
"step": 22500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.31,
|
| 294 |
+
"learning_rate": 1.1248097412480976e-05,
|
| 295 |
+
"loss": 0.0991,
|
| 296 |
+
"step": 23000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.34,
|
| 300 |
+
"learning_rate": 1.1057838660578388e-05,
|
| 301 |
+
"loss": 0.0964,
|
| 302 |
+
"step": 23500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.37,
|
| 306 |
+
"learning_rate": 1.08675799086758e-05,
|
| 307 |
+
"loss": 0.0836,
|
| 308 |
+
"step": 24000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.4,
|
| 312 |
+
"learning_rate": 1.0677321156773213e-05,
|
| 313 |
+
"loss": 0.0969,
|
| 314 |
+
"step": 24500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.43,
|
| 318 |
+
"learning_rate": 1.0487062404870625e-05,
|
| 319 |
+
"loss": 0.0819,
|
| 320 |
+
"step": 25000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.46,
|
| 324 |
+
"learning_rate": 1.0296803652968037e-05,
|
| 325 |
+
"loss": 0.1056,
|
| 326 |
+
"step": 25500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.48,
|
| 330 |
+
"learning_rate": 1.010654490106545e-05,
|
| 331 |
+
"loss": 0.0796,
|
| 332 |
+
"step": 26000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.51,
|
| 336 |
+
"learning_rate": 9.916286149162863e-06,
|
| 337 |
+
"loss": 0.077,
|
| 338 |
+
"step": 26500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.54,
|
| 342 |
+
"learning_rate": 9.726027397260275e-06,
|
| 343 |
+
"loss": 0.0737,
|
| 344 |
+
"step": 27000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.57,
|
| 348 |
+
"learning_rate": 9.535768645357688e-06,
|
| 349 |
+
"loss": 0.0968,
|
| 350 |
+
"step": 27500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.6,
|
| 354 |
+
"learning_rate": 9.3455098934551e-06,
|
| 355 |
+
"loss": 0.0774,
|
| 356 |
+
"step": 28000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.63,
|
| 360 |
+
"learning_rate": 9.155251141552512e-06,
|
| 361 |
+
"loss": 0.0734,
|
| 362 |
+
"step": 28500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.66,
|
| 366 |
+
"learning_rate": 8.964992389649924e-06,
|
| 367 |
+
"loss": 0.0792,
|
| 368 |
+
"step": 29000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.68,
|
| 372 |
+
"learning_rate": 8.774733637747336e-06,
|
| 373 |
+
"loss": 0.0771,
|
| 374 |
+
"step": 29500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.71,
|
| 378 |
+
"learning_rate": 8.584474885844748e-06,
|
| 379 |
+
"loss": 0.076,
|
| 380 |
+
"step": 30000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.74,
|
| 384 |
+
"learning_rate": 8.394216133942162e-06,
|
| 385 |
+
"loss": 0.0794,
|
| 386 |
+
"step": 30500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 1.77,
|
| 390 |
+
"learning_rate": 8.203957382039575e-06,
|
| 391 |
+
"loss": 0.0842,
|
| 392 |
+
"step": 31000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 1.8,
|
| 396 |
+
"learning_rate": 8.013698630136987e-06,
|
| 397 |
+
"loss": 0.0831,
|
| 398 |
+
"step": 31500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 1.83,
|
| 402 |
+
"learning_rate": 7.823439878234399e-06,
|
| 403 |
+
"loss": 0.0804,
|
| 404 |
+
"step": 32000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.86,
|
| 408 |
+
"learning_rate": 7.633181126331813e-06,
|
| 409 |
+
"loss": 0.081,
|
| 410 |
+
"step": 32500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 1.88,
|
| 414 |
+
"learning_rate": 7.442922374429224e-06,
|
| 415 |
+
"loss": 0.0731,
|
| 416 |
+
"step": 33000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.91,
|
| 420 |
+
"learning_rate": 7.252663622526636e-06,
|
| 421 |
+
"loss": 0.0615,
|
| 422 |
+
"step": 33500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 1.94,
|
| 426 |
+
"learning_rate": 7.0624048706240486e-06,
|
| 427 |
+
"loss": 0.0823,
|
| 428 |
+
"step": 34000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 1.97,
|
| 432 |
+
"learning_rate": 6.8721461187214625e-06,
|
| 433 |
+
"loss": 0.0613,
|
| 434 |
+
"step": 34500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.0,
|
| 438 |
+
"learning_rate": 6.681887366818875e-06,
|
| 439 |
+
"loss": 0.0777,
|
| 440 |
+
"step": 35000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.0,
|
| 444 |
+
"eval_accuracy": 0.9699693444846775,
|
| 445 |
+
"eval_f1": 0.8876556316957822,
|
| 446 |
+
"eval_loss": 0.11980047076940536,
|
| 447 |
+
"eval_precision": 0.8790862167845537,
|
| 448 |
+
"eval_recall": 0.8963937621832359,
|
| 449 |
+
"eval_runtime": 35.3315,
|
| 450 |
+
"eval_samples_per_second": 247.938,
|
| 451 |
+
"eval_steps_per_second": 61.984,
|
| 452 |
+
"step": 35040
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.03,
|
| 456 |
+
"learning_rate": 6.491628614916287e-06,
|
| 457 |
+
"loss": 0.0397,
|
| 458 |
+
"step": 35500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.05,
|
| 462 |
+
"learning_rate": 6.301369863013699e-06,
|
| 463 |
+
"loss": 0.0581,
|
| 464 |
+
"step": 36000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.08,
|
| 468 |
+
"learning_rate": 6.111111111111112e-06,
|
| 469 |
+
"loss": 0.0481,
|
| 470 |
+
"step": 36500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.11,
|
| 474 |
+
"learning_rate": 5.920852359208524e-06,
|
| 475 |
+
"loss": 0.042,
|
| 476 |
+
"step": 37000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.14,
|
| 480 |
+
"learning_rate": 5.7305936073059365e-06,
|
| 481 |
+
"loss": 0.0472,
|
| 482 |
+
"step": 37500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.17,
|
| 486 |
+
"learning_rate": 5.540334855403349e-06,
|
| 487 |
+
"loss": 0.046,
|
| 488 |
+
"step": 38000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.2,
|
| 492 |
+
"learning_rate": 5.350076103500762e-06,
|
| 493 |
+
"loss": 0.042,
|
| 494 |
+
"step": 38500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.23,
|
| 498 |
+
"learning_rate": 5.159817351598174e-06,
|
| 499 |
+
"loss": 0.0428,
|
| 500 |
+
"step": 39000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.25,
|
| 504 |
+
"learning_rate": 4.969558599695586e-06,
|
| 505 |
+
"loss": 0.0458,
|
| 506 |
+
"step": 39500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.28,
|
| 510 |
+
"learning_rate": 4.779299847792998e-06,
|
| 511 |
+
"loss": 0.0371,
|
| 512 |
+
"step": 40000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.31,
|
| 516 |
+
"learning_rate": 4.589041095890411e-06,
|
| 517 |
+
"loss": 0.0582,
|
| 518 |
+
"step": 40500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.34,
|
| 522 |
+
"learning_rate": 4.3987823439878235e-06,
|
| 523 |
+
"loss": 0.04,
|
| 524 |
+
"step": 41000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.37,
|
| 528 |
+
"learning_rate": 4.2085235920852366e-06,
|
| 529 |
+
"loss": 0.0559,
|
| 530 |
+
"step": 41500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.4,
|
| 534 |
+
"learning_rate": 4.018264840182649e-06,
|
| 535 |
+
"loss": 0.0442,
|
| 536 |
+
"step": 42000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.43,
|
| 540 |
+
"learning_rate": 3.828006088280061e-06,
|
| 541 |
+
"loss": 0.0482,
|
| 542 |
+
"step": 42500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.45,
|
| 546 |
+
"learning_rate": 3.6377473363774736e-06,
|
| 547 |
+
"loss": 0.0434,
|
| 548 |
+
"step": 43000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.48,
|
| 552 |
+
"learning_rate": 3.447488584474886e-06,
|
| 553 |
+
"loss": 0.0457,
|
| 554 |
+
"step": 43500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.51,
|
| 558 |
+
"learning_rate": 3.2572298325722984e-06,
|
| 559 |
+
"loss": 0.0404,
|
| 560 |
+
"step": 44000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.54,
|
| 564 |
+
"learning_rate": 3.066971080669711e-06,
|
| 565 |
+
"loss": 0.0521,
|
| 566 |
+
"step": 44500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.57,
|
| 570 |
+
"learning_rate": 2.876712328767123e-06,
|
| 571 |
+
"loss": 0.0605,
|
| 572 |
+
"step": 45000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 2.6,
|
| 576 |
+
"learning_rate": 2.6864535768645362e-06,
|
| 577 |
+
"loss": 0.05,
|
| 578 |
+
"step": 45500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 2.63,
|
| 582 |
+
"learning_rate": 2.4961948249619484e-06,
|
| 583 |
+
"loss": 0.0471,
|
| 584 |
+
"step": 46000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 2.65,
|
| 588 |
+
"learning_rate": 2.3059360730593606e-06,
|
| 589 |
+
"loss": 0.0403,
|
| 590 |
+
"step": 46500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.68,
|
| 594 |
+
"learning_rate": 2.1156773211567732e-06,
|
| 595 |
+
"loss": 0.0533,
|
| 596 |
+
"step": 47000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.71,
|
| 600 |
+
"learning_rate": 1.925418569254186e-06,
|
| 601 |
+
"loss": 0.04,
|
| 602 |
+
"step": 47500
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 2.74,
|
| 606 |
+
"learning_rate": 1.7351598173515982e-06,
|
| 607 |
+
"loss": 0.0435,
|
| 608 |
+
"step": 48000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 2.77,
|
| 612 |
+
"learning_rate": 1.5449010654490107e-06,
|
| 613 |
+
"loss": 0.0417,
|
| 614 |
+
"step": 48500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 2.8,
|
| 618 |
+
"learning_rate": 1.3546423135464233e-06,
|
| 619 |
+
"loss": 0.036,
|
| 620 |
+
"step": 49000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 2.83,
|
| 624 |
+
"learning_rate": 1.1643835616438357e-06,
|
| 625 |
+
"loss": 0.0364,
|
| 626 |
+
"step": 49500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 2.85,
|
| 630 |
+
"learning_rate": 9.74124809741248e-07,
|
| 631 |
+
"loss": 0.0339,
|
| 632 |
+
"step": 50000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 2.88,
|
| 636 |
+
"learning_rate": 7.838660578386606e-07,
|
| 637 |
+
"loss": 0.0291,
|
| 638 |
+
"step": 50500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 2.91,
|
| 642 |
+
"learning_rate": 5.936073059360731e-07,
|
| 643 |
+
"loss": 0.0351,
|
| 644 |
+
"step": 51000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 2.94,
|
| 648 |
+
"learning_rate": 4.0334855403348556e-07,
|
| 649 |
+
"loss": 0.0393,
|
| 650 |
+
"step": 51500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 2.97,
|
| 654 |
+
"learning_rate": 2.1308980213089802e-07,
|
| 655 |
+
"loss": 0.0342,
|
| 656 |
+
"step": 52000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.0,
|
| 660 |
+
"learning_rate": 2.2831050228310502e-08,
|
| 661 |
+
"loss": 0.0438,
|
| 662 |
+
"step": 52500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.0,
|
| 666 |
+
"eval_accuracy": 0.9766719005351363,
|
| 667 |
+
"eval_f1": 0.9076281864883202,
|
| 668 |
+
"eval_loss": 0.10344000160694122,
|
| 669 |
+
"eval_precision": 0.9026412184306921,
|
| 670 |
+
"eval_recall": 0.9126705653021443,
|
| 671 |
+
"eval_runtime": 35.4439,
|
| 672 |
+
"eval_samples_per_second": 247.151,
|
| 673 |
+
"eval_steps_per_second": 61.788,
|
| 674 |
+
"step": 52560
|
| 675 |
+
}
|
| 676 |
+
],
|
| 677 |
+
"max_steps": 52560,
|
| 678 |
+
"num_train_epochs": 3,
|
| 679 |
+
"total_flos": 4848998991932040.0,
|
| 680 |
+
"trial_name": null,
|
| 681 |
+
"trial_params": null
|
| 682 |
+
}
|
model/checkpoint-52560/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7c82c4e51275ddff46bf99412a77befffdc5e79558b7fbe63cfd77812286ac34
|
| 3 |
+
size 4027
|
model/checkpoint-52560/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|