Upload folder using huggingface_hub
Browse files- all_runs/checkpoint-2912/added_tokens.json +3 -0
- all_runs/checkpoint-2912/config.json +38 -0
- all_runs/checkpoint-2912/model.safetensors +3 -0
- all_runs/checkpoint-2912/optimizer.pt +3 -0
- all_runs/checkpoint-2912/rng_state.pth +3 -0
- all_runs/checkpoint-2912/scheduler.pt +3 -0
- all_runs/checkpoint-2912/special_tokens_map.json +13 -0
- all_runs/checkpoint-2912/tokenizer.json +0 -0
- all_runs/checkpoint-2912/tokenizer_config.json +63 -0
- all_runs/checkpoint-2912/trainer_state.json +342 -0
- all_runs/checkpoint-2912/training_args.bin +3 -0
- all_runs/checkpoint-2912/vocab.txt +0 -0
all_runs/checkpoint-2912/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<pad>": 28996
|
| 3 |
+
}
|
all_runs/checkpoint-2912/config.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "distilbert/distilbert-base-cased",
|
| 3 |
+
"activation": "gelu",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"DistilBertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"dim": 768,
|
| 9 |
+
"dropout": 0.1,
|
| 10 |
+
"hidden_dim": 3072,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "positive",
|
| 13 |
+
"1": "negative",
|
| 14 |
+
"2": "neutral",
|
| 15 |
+
"3": "rel_by_party"
|
| 16 |
+
},
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"negative": 1,
|
| 20 |
+
"neutral": 2,
|
| 21 |
+
"positive": 0,
|
| 22 |
+
"rel_by_party": 3
|
| 23 |
+
},
|
| 24 |
+
"max_position_embeddings": 512,
|
| 25 |
+
"model_type": "distilbert",
|
| 26 |
+
"n_heads": 12,
|
| 27 |
+
"n_layers": 6,
|
| 28 |
+
"output_past": true,
|
| 29 |
+
"pad_token_id": 28996,
|
| 30 |
+
"problem_type": "single_label_classification",
|
| 31 |
+
"qa_dropout": 0.1,
|
| 32 |
+
"seq_classif_dropout": 0.2,
|
| 33 |
+
"sinusoidal_pos_embds": false,
|
| 34 |
+
"tie_weights_": true,
|
| 35 |
+
"torch_dtype": "float32",
|
| 36 |
+
"transformers_version": "4.40.2",
|
| 37 |
+
"vocab_size": 28997
|
| 38 |
+
}
|
all_runs/checkpoint-2912/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:045813fb304375524765defb2314e02c317ce921eb8df69bc4608be4f93e63d6
|
| 3 |
+
size 263153912
|
all_runs/checkpoint-2912/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:df454e256ed3598c4a80ec00ade3f8dbba1325c70c0169a865ef98d6c80f5e84
|
| 3 |
+
size 526369978
|
all_runs/checkpoint-2912/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b64bf4b4c6b40fab1a3ada5363f5aca1e5a78ae939a935b290ef2b4446abed7e
|
| 3 |
+
size 14244
|
all_runs/checkpoint-2912/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:afe6d0516788adb822b5c806c01f0510f4ff3b63f4d1dd7bff40c1c7f726b8cc
|
| 3 |
+
size 1064
|
all_runs/checkpoint-2912/special_tokens_map.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": {
|
| 5 |
+
"content": "<pad>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": false,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false
|
| 10 |
+
},
|
| 11 |
+
"sep_token": "[SEP]",
|
| 12 |
+
"unk_token": "[UNK]"
|
| 13 |
+
}
|
all_runs/checkpoint-2912/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
all_runs/checkpoint-2912/tokenizer_config.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
},
|
| 43 |
+
"28996": {
|
| 44 |
+
"content": "<pad>",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": false,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": false,
|
| 49 |
+
"special": true
|
| 50 |
+
}
|
| 51 |
+
},
|
| 52 |
+
"clean_up_tokenization_spaces": true,
|
| 53 |
+
"cls_token": "[CLS]",
|
| 54 |
+
"do_lower_case": false,
|
| 55 |
+
"mask_token": "[MASK]",
|
| 56 |
+
"model_max_length": 512,
|
| 57 |
+
"pad_token": "<pad>",
|
| 58 |
+
"sep_token": "[SEP]",
|
| 59 |
+
"strip_accents": null,
|
| 60 |
+
"tokenize_chinese_chars": true,
|
| 61 |
+
"tokenizer_class": "DistilBertTokenizer",
|
| 62 |
+
"unk_token": "[UNK]"
|
| 63 |
+
}
|
all_runs/checkpoint-2912/trainer_state.json
ADDED
|
@@ -0,0 +1,342 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6301119928151375,
|
| 3 |
+
"best_model_checkpoint": "case-analysis-distilbert-base-cased/checkpoint-896",
|
| 4 |
+
"epoch": 13.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2912,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.7661469933184856,
|
| 14 |
+
"eval_f1_macro": 0.5270352134866867,
|
| 15 |
+
"eval_f1_micro": 0.7661469933184856,
|
| 16 |
+
"eval_f1_weighted": 0.7333313004442764,
|
| 17 |
+
"eval_loss": 0.7000867128372192,
|
| 18 |
+
"eval_macro_fpr": 0.1330263157894737,
|
| 19 |
+
"eval_macro_sensitivity": 0.5137440725676019,
|
| 20 |
+
"eval_macro_specificity": 0.8818942592408516,
|
| 21 |
+
"eval_precision": 0.7310834070149471,
|
| 22 |
+
"eval_precision_macro": 0.5790641299401408,
|
| 23 |
+
"eval_recall": 0.7661469933184856,
|
| 24 |
+
"eval_recall_macro": 0.5137440725676019,
|
| 25 |
+
"eval_runtime": 5.4998,
|
| 26 |
+
"eval_samples_per_second": 81.639,
|
| 27 |
+
"eval_steps_per_second": 10.364,
|
| 28 |
+
"eval_weighted_fpr": 0.09234828496042216,
|
| 29 |
+
"eval_weighted_sensitivity": 0.7661469933184856,
|
| 30 |
+
"eval_weighted_specificity": 0.761430043644921,
|
| 31 |
+
"step": 224
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 2.0,
|
| 35 |
+
"eval_accuracy": 0.7750556792873051,
|
| 36 |
+
"eval_f1_macro": 0.5486718521990608,
|
| 37 |
+
"eval_f1_micro": 0.775055679287305,
|
| 38 |
+
"eval_f1_weighted": 0.7492575717498114,
|
| 39 |
+
"eval_loss": 0.7387592792510986,
|
| 40 |
+
"eval_macro_fpr": 0.12078200823200373,
|
| 41 |
+
"eval_macro_sensitivity": 0.5464266787796199,
|
| 42 |
+
"eval_macro_specificity": 0.8914743027952345,
|
| 43 |
+
"eval_precision": 0.7315284458297402,
|
| 44 |
+
"eval_precision_macro": 0.5585311316366537,
|
| 45 |
+
"eval_recall": 0.7750556792873051,
|
| 46 |
+
"eval_recall_macro": 0.5464266787796199,
|
| 47 |
+
"eval_runtime": 23.5227,
|
| 48 |
+
"eval_samples_per_second": 19.088,
|
| 49 |
+
"eval_steps_per_second": 2.423,
|
| 50 |
+
"eval_weighted_fpr": 0.08820960698689956,
|
| 51 |
+
"eval_weighted_sensitivity": 0.7750556792873051,
|
| 52 |
+
"eval_weighted_specificity": 0.7908415318936333,
|
| 53 |
+
"step": 448
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 2.232142857142857,
|
| 57 |
+
"grad_norm": 22.91168785095215,
|
| 58 |
+
"learning_rate": 4.629464285714286e-05,
|
| 59 |
+
"loss": 0.7066,
|
| 60 |
+
"step": 500
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 3.0,
|
| 64 |
+
"eval_accuracy": 0.8017817371937639,
|
| 65 |
+
"eval_f1_macro": 0.576677451756589,
|
| 66 |
+
"eval_f1_micro": 0.8017817371937639,
|
| 67 |
+
"eval_f1_weighted": 0.7759761601356532,
|
| 68 |
+
"eval_loss": 0.7228724956512451,
|
| 69 |
+
"eval_macro_fpr": 0.10757194512338594,
|
| 70 |
+
"eval_macro_sensitivity": 0.5707859310800487,
|
| 71 |
+
"eval_macro_specificity": 0.9026901603449184,
|
| 72 |
+
"eval_precision": 0.7604834726234316,
|
| 73 |
+
"eval_precision_macro": 0.5932281353135314,
|
| 74 |
+
"eval_recall": 0.8017817371937639,
|
| 75 |
+
"eval_recall_macro": 0.5707859310800487,
|
| 76 |
+
"eval_runtime": 17.5142,
|
| 77 |
+
"eval_samples_per_second": 25.636,
|
| 78 |
+
"eval_steps_per_second": 3.254,
|
| 79 |
+
"eval_weighted_fpr": 0.07613344739093242,
|
| 80 |
+
"eval_weighted_sensitivity": 0.8017817371937639,
|
| 81 |
+
"eval_weighted_specificity": 0.8089789041859096,
|
| 82 |
+
"step": 672
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 4.0,
|
| 86 |
+
"eval_accuracy": 0.8062360801781737,
|
| 87 |
+
"eval_f1_macro": 0.6301119928151375,
|
| 88 |
+
"eval_f1_micro": 0.8062360801781738,
|
| 89 |
+
"eval_f1_weighted": 0.7933630533945198,
|
| 90 |
+
"eval_loss": 0.8331468105316162,
|
| 91 |
+
"eval_macro_fpr": 0.10175133159747105,
|
| 92 |
+
"eval_macro_sensitivity": 0.6114882070764424,
|
| 93 |
+
"eval_macro_specificity": 0.9069973738423197,
|
| 94 |
+
"eval_precision": 0.7896315051078991,
|
| 95 |
+
"eval_precision_macro": 0.6674639062867935,
|
| 96 |
+
"eval_recall": 0.8062360801781737,
|
| 97 |
+
"eval_recall_macro": 0.6114882070764424,
|
| 98 |
+
"eval_runtime": 17.9901,
|
| 99 |
+
"eval_samples_per_second": 24.958,
|
| 100 |
+
"eval_steps_per_second": 3.168,
|
| 101 |
+
"eval_weighted_fpr": 0.0741687979539642,
|
| 102 |
+
"eval_weighted_sensitivity": 0.8062360801781737,
|
| 103 |
+
"eval_weighted_specificity": 0.8217534151911048,
|
| 104 |
+
"step": 896
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 4.464285714285714,
|
| 108 |
+
"grad_norm": 0.23617026209831238,
|
| 109 |
+
"learning_rate": 4.258184523809524e-05,
|
| 110 |
+
"loss": 0.3654,
|
| 111 |
+
"step": 1000
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 5.0,
|
| 115 |
+
"eval_accuracy": 0.7683741648106904,
|
| 116 |
+
"eval_f1_macro": 0.5896043825978781,
|
| 117 |
+
"eval_f1_micro": 0.7683741648106904,
|
| 118 |
+
"eval_f1_weighted": 0.7610828501277613,
|
| 119 |
+
"eval_loss": 1.2299734354019165,
|
| 120 |
+
"eval_macro_fpr": 0.10662840460883276,
|
| 121 |
+
"eval_macro_sensitivity": 0.6130556645262528,
|
| 122 |
+
"eval_macro_specificity": 0.9056447032260023,
|
| 123 |
+
"eval_precision": 0.7698582566254396,
|
| 124 |
+
"eval_precision_macro": 0.6085394130060126,
|
| 125 |
+
"eval_recall": 0.7683741648106904,
|
| 126 |
+
"eval_recall_macro": 0.6130556645262528,
|
| 127 |
+
"eval_runtime": 19.5517,
|
| 128 |
+
"eval_samples_per_second": 22.965,
|
| 129 |
+
"eval_steps_per_second": 2.915,
|
| 130 |
+
"eval_weighted_fpr": 0.0913081650570676,
|
| 131 |
+
"eval_weighted_sensitivity": 0.7683741648106904,
|
| 132 |
+
"eval_weighted_specificity": 0.8542046480933186,
|
| 133 |
+
"step": 1120
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 6.0,
|
| 137 |
+
"eval_accuracy": 0.8129175946547884,
|
| 138 |
+
"eval_f1_macro": 0.6299577384875399,
|
| 139 |
+
"eval_f1_micro": 0.8129175946547884,
|
| 140 |
+
"eval_f1_weighted": 0.7972266629935387,
|
| 141 |
+
"eval_loss": 1.0698440074920654,
|
| 142 |
+
"eval_macro_fpr": 0.09572659519735909,
|
| 143 |
+
"eval_macro_sensitivity": 0.6152769461592991,
|
| 144 |
+
"eval_macro_specificity": 0.9133810019905357,
|
| 145 |
+
"eval_precision": 0.7940368363876591,
|
| 146 |
+
"eval_precision_macro": 0.6863588955891834,
|
| 147 |
+
"eval_recall": 0.8129175946547884,
|
| 148 |
+
"eval_recall_macro": 0.6152769461592991,
|
| 149 |
+
"eval_runtime": 21.7346,
|
| 150 |
+
"eval_samples_per_second": 20.658,
|
| 151 |
+
"eval_steps_per_second": 2.623,
|
| 152 |
+
"eval_weighted_fpr": 0.07124681933842239,
|
| 153 |
+
"eval_weighted_sensitivity": 0.8129175946547884,
|
| 154 |
+
"eval_weighted_specificity": 0.8406064133073544,
|
| 155 |
+
"step": 1344
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 6.696428571428571,
|
| 159 |
+
"grad_norm": 0.6412457823753357,
|
| 160 |
+
"learning_rate": 3.8869047619047625e-05,
|
| 161 |
+
"loss": 0.2047,
|
| 162 |
+
"step": 1500
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 7.0,
|
| 166 |
+
"eval_accuracy": 0.7884187082405345,
|
| 167 |
+
"eval_f1_macro": 0.6141215790585401,
|
| 168 |
+
"eval_f1_micro": 0.7884187082405345,
|
| 169 |
+
"eval_f1_weighted": 0.7892035783447391,
|
| 170 |
+
"eval_loss": 1.3299829959869385,
|
| 171 |
+
"eval_macro_fpr": 0.10442766929340637,
|
| 172 |
+
"eval_macro_sensitivity": 0.5958936400112871,
|
| 173 |
+
"eval_macro_specificity": 0.9076274830150074,
|
| 174 |
+
"eval_precision": 0.7959953084611945,
|
| 175 |
+
"eval_precision_macro": 0.6412251711254097,
|
| 176 |
+
"eval_recall": 0.7884187082405345,
|
| 177 |
+
"eval_recall_macro": 0.5958936400112871,
|
| 178 |
+
"eval_runtime": 19.4013,
|
| 179 |
+
"eval_samples_per_second": 23.143,
|
| 180 |
+
"eval_steps_per_second": 2.938,
|
| 181 |
+
"eval_weighted_fpr": 0.08210890233362143,
|
| 182 |
+
"eval_weighted_sensitivity": 0.7884187082405345,
|
| 183 |
+
"eval_weighted_specificity": 0.842091223819495,
|
| 184 |
+
"step": 1568
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 8.0,
|
| 188 |
+
"eval_accuracy": 0.8106904231625836,
|
| 189 |
+
"eval_f1_macro": 0.6163443598585857,
|
| 190 |
+
"eval_f1_micro": 0.8106904231625834,
|
| 191 |
+
"eval_f1_weighted": 0.7947158089605988,
|
| 192 |
+
"eval_loss": 1.3869845867156982,
|
| 193 |
+
"eval_macro_fpr": 0.09827643474073441,
|
| 194 |
+
"eval_macro_sensitivity": 0.6062894445247387,
|
| 195 |
+
"eval_macro_specificity": 0.9106173671247473,
|
| 196 |
+
"eval_precision": 0.7860704056527595,
|
| 197 |
+
"eval_precision_macro": 0.6466819479006202,
|
| 198 |
+
"eval_recall": 0.8106904231625836,
|
| 199 |
+
"eval_recall_macro": 0.6062894445247387,
|
| 200 |
+
"eval_runtime": 24.9242,
|
| 201 |
+
"eval_samples_per_second": 18.015,
|
| 202 |
+
"eval_steps_per_second": 2.287,
|
| 203 |
+
"eval_weighted_fpr": 0.07221750212404418,
|
| 204 |
+
"eval_weighted_sensitivity": 0.8106904231625836,
|
| 205 |
+
"eval_weighted_specificity": 0.8317790453364052,
|
| 206 |
+
"step": 1792
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 8.928571428571429,
|
| 210 |
+
"grad_norm": 0.009310582652688026,
|
| 211 |
+
"learning_rate": 3.5156250000000004e-05,
|
| 212 |
+
"loss": 0.0795,
|
| 213 |
+
"step": 2000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 9.0,
|
| 217 |
+
"eval_accuracy": 0.7951002227171492,
|
| 218 |
+
"eval_f1_macro": 0.6036078586352096,
|
| 219 |
+
"eval_f1_micro": 0.7951002227171492,
|
| 220 |
+
"eval_f1_weighted": 0.7803481654390761,
|
| 221 |
+
"eval_loss": 1.5031198263168335,
|
| 222 |
+
"eval_macro_fpr": 0.10395014224349126,
|
| 223 |
+
"eval_macro_sensitivity": 0.596858890976538,
|
| 224 |
+
"eval_macro_specificity": 0.9067643380561508,
|
| 225 |
+
"eval_precision": 0.7719081858839341,
|
| 226 |
+
"eval_precision_macro": 0.6275149499476224,
|
| 227 |
+
"eval_recall": 0.7951002227171492,
|
| 228 |
+
"eval_recall_macro": 0.596858890976538,
|
| 229 |
+
"eval_runtime": 19.5528,
|
| 230 |
+
"eval_samples_per_second": 22.964,
|
| 231 |
+
"eval_steps_per_second": 2.915,
|
| 232 |
+
"eval_weighted_fpr": 0.07910576096302666,
|
| 233 |
+
"eval_weighted_sensitivity": 0.7951002227171492,
|
| 234 |
+
"eval_weighted_specificity": 0.8319571295074538,
|
| 235 |
+
"step": 2016
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 10.0,
|
| 239 |
+
"eval_accuracy": 0.7728285077951003,
|
| 240 |
+
"eval_f1_macro": 0.6196001194983027,
|
| 241 |
+
"eval_f1_micro": 0.7728285077951002,
|
| 242 |
+
"eval_f1_weighted": 0.7758603070332255,
|
| 243 |
+
"eval_loss": 1.6303768157958984,
|
| 244 |
+
"eval_macro_fpr": 0.10604441860615665,
|
| 245 |
+
"eval_macro_sensitivity": 0.6233043585984762,
|
| 246 |
+
"eval_macro_specificity": 0.9072269655447398,
|
| 247 |
+
"eval_precision": 0.7796278998362599,
|
| 248 |
+
"eval_precision_macro": 0.6171087696246884,
|
| 249 |
+
"eval_recall": 0.7728285077951003,
|
| 250 |
+
"eval_recall_macro": 0.6233043585984762,
|
| 251 |
+
"eval_runtime": 17.8727,
|
| 252 |
+
"eval_samples_per_second": 25.122,
|
| 253 |
+
"eval_steps_per_second": 3.189,
|
| 254 |
+
"eval_weighted_fpr": 0.08923884514435695,
|
| 255 |
+
"eval_weighted_sensitivity": 0.7728285077951003,
|
| 256 |
+
"eval_weighted_specificity": 0.8560793543838588,
|
| 257 |
+
"step": 2240
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 11.0,
|
| 261 |
+
"eval_accuracy": 0.8040089086859689,
|
| 262 |
+
"eval_f1_macro": 0.60971035256129,
|
| 263 |
+
"eval_f1_micro": 0.8040089086859689,
|
| 264 |
+
"eval_f1_weighted": 0.7884103774950618,
|
| 265 |
+
"eval_loss": 1.6553305387496948,
|
| 266 |
+
"eval_macro_fpr": 0.10029712404712406,
|
| 267 |
+
"eval_macro_sensitivity": 0.6047305900247076,
|
| 268 |
+
"eval_macro_specificity": 0.9093226487010472,
|
| 269 |
+
"eval_precision": 0.7801791198576906,
|
| 270 |
+
"eval_precision_macro": 0.6404694360494954,
|
| 271 |
+
"eval_recall": 0.8040089086859689,
|
| 272 |
+
"eval_recall_macro": 0.6047305900247076,
|
| 273 |
+
"eval_runtime": 18.7706,
|
| 274 |
+
"eval_samples_per_second": 23.92,
|
| 275 |
+
"eval_steps_per_second": 3.037,
|
| 276 |
+
"eval_weighted_fpr": 0.07514944491887275,
|
| 277 |
+
"eval_weighted_sensitivity": 0.8040089086859689,
|
| 278 |
+
"eval_weighted_specificity": 0.83328168611822,
|
| 279 |
+
"step": 2464
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 11.160714285714286,
|
| 283 |
+
"grad_norm": 0.003156836610287428,
|
| 284 |
+
"learning_rate": 3.143601190476191e-05,
|
| 285 |
+
"loss": 0.0309,
|
| 286 |
+
"step": 2500
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 12.0,
|
| 290 |
+
"eval_accuracy": 0.799554565701559,
|
| 291 |
+
"eval_f1_macro": 0.6072980420599631,
|
| 292 |
+
"eval_f1_micro": 0.799554565701559,
|
| 293 |
+
"eval_f1_weighted": 0.7860582530348104,
|
| 294 |
+
"eval_loss": 1.6668155193328857,
|
| 295 |
+
"eval_macro_fpr": 0.09988889279594411,
|
| 296 |
+
"eval_macro_sensitivity": 0.6084453437394614,
|
| 297 |
+
"eval_macro_specificity": 0.9106688993495651,
|
| 298 |
+
"eval_precision": 0.777569146196374,
|
| 299 |
+
"eval_precision_macro": 0.6247320881812615,
|
| 300 |
+
"eval_recall": 0.799554565701559,
|
| 301 |
+
"eval_recall_macro": 0.6084453437394614,
|
| 302 |
+
"eval_runtime": 17.8921,
|
| 303 |
+
"eval_samples_per_second": 25.095,
|
| 304 |
+
"eval_steps_per_second": 3.186,
|
| 305 |
+
"eval_weighted_fpr": 0.07712082262210797,
|
| 306 |
+
"eval_weighted_sensitivity": 0.799554565701559,
|
| 307 |
+
"eval_weighted_specificity": 0.8431210316967014,
|
| 308 |
+
"step": 2688
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 13.0,
|
| 312 |
+
"eval_accuracy": 0.8040089086859689,
|
| 313 |
+
"eval_f1_macro": 0.5912164538573286,
|
| 314 |
+
"eval_f1_micro": 0.8040089086859689,
|
| 315 |
+
"eval_f1_weighted": 0.7845547674905504,
|
| 316 |
+
"eval_loss": 1.7547632455825806,
|
| 317 |
+
"eval_macro_fpr": 0.10303078503610634,
|
| 318 |
+
"eval_macro_sensitivity": 0.5846762537939009,
|
| 319 |
+
"eval_macro_specificity": 0.9063991203807322,
|
| 320 |
+
"eval_precision": 0.7724328584664579,
|
| 321 |
+
"eval_precision_macro": 0.6059236646315017,
|
| 322 |
+
"eval_recall": 0.8040089086859689,
|
| 323 |
+
"eval_recall_macro": 0.5846762537939009,
|
| 324 |
+
"eval_runtime": 18.8418,
|
| 325 |
+
"eval_samples_per_second": 23.83,
|
| 326 |
+
"eval_steps_per_second": 3.025,
|
| 327 |
+
"eval_weighted_fpr": 0.07514944491887275,
|
| 328 |
+
"eval_weighted_sensitivity": 0.8040089086859689,
|
| 329 |
+
"eval_weighted_specificity": 0.8215875728369607,
|
| 330 |
+
"step": 2912
|
| 331 |
+
}
|
| 332 |
+
],
|
| 333 |
+
"logging_steps": 500,
|
| 334 |
+
"max_steps": 6720,
|
| 335 |
+
"num_input_tokens_seen": 0,
|
| 336 |
+
"num_train_epochs": 30,
|
| 337 |
+
"save_steps": 500,
|
| 338 |
+
"total_flos": 3086070586540032.0,
|
| 339 |
+
"train_batch_size": 8,
|
| 340 |
+
"trial_name": null,
|
| 341 |
+
"trial_params": null
|
| 342 |
+
}
|
all_runs/checkpoint-2912/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3168244efe16c0aeafe631cea29ca8c9c2bbb665a169c0d40c2e4d22eaca40d9
|
| 3 |
+
size 5048
|
all_runs/checkpoint-2912/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|