Upload 11 files
Browse files- config.json +64 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +55 -0
- trainer_state.json +1211 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "distilbert-base-uncased",
|
| 3 |
+
"activation": "gelu",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"HierarchicalMultiLabelClassifier"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"dim": 768,
|
| 9 |
+
"dropout": 0.1,
|
| 10 |
+
"hidden_dim": 3072,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "EVENT",
|
| 13 |
+
"1": "EVENT_Illness",
|
| 14 |
+
"2": "EVENT_Accident",
|
| 15 |
+
"3": "EVENT_Other",
|
| 16 |
+
"4": "CAUSE",
|
| 17 |
+
"5": "CAUSE_Just_Happens",
|
| 18 |
+
"6": "CAUSE_Material_Physical",
|
| 19 |
+
"7": "CAUSE_Spirits_Gods",
|
| 20 |
+
"8": "CAUSE_Witchcraft_Sorcery",
|
| 21 |
+
"9": "CAUSE_Rule_Violation_Taboo",
|
| 22 |
+
"10": "CAUSE_Other",
|
| 23 |
+
"11": "ACTION",
|
| 24 |
+
"12": "ACTION_Physical_Material",
|
| 25 |
+
"13": "ACTION_Technical_Specialist",
|
| 26 |
+
"14": "ACTION_Divination",
|
| 27 |
+
"15": "ACTION_Shaman_Medium_Healer",
|
| 28 |
+
"16": "ACTION_Priest_High_Religion",
|
| 29 |
+
"17": "ACTION_Other"
|
| 30 |
+
},
|
| 31 |
+
"initializer_range": 0.02,
|
| 32 |
+
"label2id": {
|
| 33 |
+
"ACTION": 11,
|
| 34 |
+
"ACTION_Divination": 14,
|
| 35 |
+
"ACTION_Other": 17,
|
| 36 |
+
"ACTION_Physical_Material": 12,
|
| 37 |
+
"ACTION_Priest_High_Religion": 16,
|
| 38 |
+
"ACTION_Shaman_Medium_Healer": 15,
|
| 39 |
+
"ACTION_Technical_Specialist": 13,
|
| 40 |
+
"CAUSE": 4,
|
| 41 |
+
"CAUSE_Just_Happens": 5,
|
| 42 |
+
"CAUSE_Material_Physical": 6,
|
| 43 |
+
"CAUSE_Other": 10,
|
| 44 |
+
"CAUSE_Rule_Violation_Taboo": 9,
|
| 45 |
+
"CAUSE_Spirits_Gods": 7,
|
| 46 |
+
"CAUSE_Witchcraft_Sorcery": 8,
|
| 47 |
+
"EVENT": 0,
|
| 48 |
+
"EVENT_Accident": 2,
|
| 49 |
+
"EVENT_Illness": 1,
|
| 50 |
+
"EVENT_Other": 3
|
| 51 |
+
},
|
| 52 |
+
"max_position_embeddings": 512,
|
| 53 |
+
"model_type": "distilbert",
|
| 54 |
+
"n_heads": 12,
|
| 55 |
+
"n_layers": 6,
|
| 56 |
+
"pad_token_id": 0,
|
| 57 |
+
"qa_dropout": 0.1,
|
| 58 |
+
"seq_classif_dropout": 0.2,
|
| 59 |
+
"sinusoidal_pos_embds": false,
|
| 60 |
+
"tie_weights_": true,
|
| 61 |
+
"torch_dtype": "float32",
|
| 62 |
+
"transformers_version": "4.42.4",
|
| 63 |
+
"vocab_size": 30522
|
| 64 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2418e84da1103ed5fa5ec08c1b4f3cf110d222e7f74f1bbf1fba59d6021a664
|
| 3 |
+
size 265938696
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bd5d633e79e518d0459d8bee314a49fd9b2566d41b2e4b3e4482fee0a7d5973e
|
| 3 |
+
size 531945530
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2b5a8f2106c34eeb573aad7cb7f7a0eb352cc85dc08adef1137eba3bdb370c4a
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:10b3b185fea1205fbfcb56f550d303bc0b6cb7d325de973a2d3dc5d3e34c96cf
|
| 3 |
+
size 1064
|
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 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_lower_case": true,
|
| 47 |
+
"mask_token": "[MASK]",
|
| 48 |
+
"model_max_length": 512,
|
| 49 |
+
"pad_token": "[PAD]",
|
| 50 |
+
"sep_token": "[SEP]",
|
| 51 |
+
"strip_accents": null,
|
| 52 |
+
"tokenize_chinese_chars": true,
|
| 53 |
+
"tokenizer_class": "DistilBertTokenizer",
|
| 54 |
+
"unk_token": "[UNK]"
|
| 55 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1211 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6518611919550361,
|
| 3 |
+
"best_model_checkpoint": "drive/MyDrive/NLP_HRAF//Models/HRAF_MultiLabel_Hierarchical/Model_1_BaseTest/Hierarchy_test_fold_1/checkpoint-10790",
|
| 4 |
+
"epoch": 13.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 10790,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.12048192771084337,
|
| 13 |
+
"grad_norm": 4.097501277923584,
|
| 14 |
+
"learning_rate": 1.9839357429718877e-05,
|
| 15 |
+
"loss": 2.1939,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.24096385542168675,
|
| 20 |
+
"grad_norm": 3.734555721282959,
|
| 21 |
+
"learning_rate": 1.967871485943775e-05,
|
| 22 |
+
"loss": 1.7255,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.3614457831325301,
|
| 27 |
+
"grad_norm": 3.1774344444274902,
|
| 28 |
+
"learning_rate": 1.951807228915663e-05,
|
| 29 |
+
"loss": 1.5907,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.4819277108433735,
|
| 34 |
+
"grad_norm": 5.335158824920654,
|
| 35 |
+
"learning_rate": 1.9357429718875505e-05,
|
| 36 |
+
"loss": 1.5148,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.6024096385542169,
|
| 41 |
+
"grad_norm": 10.060626983642578,
|
| 42 |
+
"learning_rate": 1.9196787148594377e-05,
|
| 43 |
+
"loss": 1.4449,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.7228915662650602,
|
| 48 |
+
"grad_norm": 6.7067084312438965,
|
| 49 |
+
"learning_rate": 1.9036144578313255e-05,
|
| 50 |
+
"loss": 1.3943,
|
| 51 |
+
"step": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.8433734939759037,
|
| 55 |
+
"grad_norm": 5.713496685028076,
|
| 56 |
+
"learning_rate": 1.887550200803213e-05,
|
| 57 |
+
"loss": 1.3414,
|
| 58 |
+
"step": 700
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.963855421686747,
|
| 62 |
+
"grad_norm": 5.983292102813721,
|
| 63 |
+
"learning_rate": 1.8714859437751005e-05,
|
| 64 |
+
"loss": 1.3267,
|
| 65 |
+
"step": 800
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 1.0,
|
| 69 |
+
"eval_action_metrics": {
|
| 70 |
+
"accuracy": 0.568414707655214,
|
| 71 |
+
"f1": 0.2052586938083121,
|
| 72 |
+
"loss": 0.28379881381988525,
|
| 73 |
+
"roc_auc": 0.557318161535029
|
| 74 |
+
},
|
| 75 |
+
"eval_cause_metrics": {
|
| 76 |
+
"accuracy": 0.5563592525617842,
|
| 77 |
+
"f1": 0.12061206120612061,
|
| 78 |
+
"loss": 0.4551348090171814,
|
| 79 |
+
"roc_auc": 0.5310209707194632
|
| 80 |
+
},
|
| 81 |
+
"eval_event_metrics": {
|
| 82 |
+
"accuracy": 0.6865581675708258,
|
| 83 |
+
"f1": 0.6954503249767873,
|
| 84 |
+
"loss": 0.4480176568031311,
|
| 85 |
+
"roc_auc": 0.7856260330258503
|
| 86 |
+
},
|
| 87 |
+
"eval_f1": 0.4580407659608961,
|
| 88 |
+
"eval_loss": 1.2922335863113403,
|
| 89 |
+
"eval_main_metrics": {
|
| 90 |
+
"accuracy": 0.5792646172393008,
|
| 91 |
+
"f1": 0.8108419838523645,
|
| 92 |
+
"loss": 1.066821813583374,
|
| 93 |
+
"roc_auc": 0.8017363757036454
|
| 94 |
+
},
|
| 95 |
+
"eval_runtime": 19.3038,
|
| 96 |
+
"eval_samples_per_second": 85.942,
|
| 97 |
+
"eval_steps_per_second": 10.775,
|
| 98 |
+
"step": 830
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.0843373493975903,
|
| 102 |
+
"grad_norm": 2.3373818397521973,
|
| 103 |
+
"learning_rate": 1.855421686746988e-05,
|
| 104 |
+
"loss": 1.2549,
|
| 105 |
+
"step": 900
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.2048192771084336,
|
| 109 |
+
"grad_norm": 4.526793956756592,
|
| 110 |
+
"learning_rate": 1.8393574297188755e-05,
|
| 111 |
+
"loss": 1.184,
|
| 112 |
+
"step": 1000
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.3253012048192772,
|
| 116 |
+
"grad_norm": 4.537077903747559,
|
| 117 |
+
"learning_rate": 1.8232931726907634e-05,
|
| 118 |
+
"loss": 1.16,
|
| 119 |
+
"step": 1100
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.4457831325301205,
|
| 123 |
+
"grad_norm": 9.092710494995117,
|
| 124 |
+
"learning_rate": 1.807228915662651e-05,
|
| 125 |
+
"loss": 1.1414,
|
| 126 |
+
"step": 1200
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 1.5662650602409638,
|
| 130 |
+
"grad_norm": 11.58354377746582,
|
| 131 |
+
"learning_rate": 1.7911646586345384e-05,
|
| 132 |
+
"loss": 1.1569,
|
| 133 |
+
"step": 1300
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 1.6867469879518073,
|
| 137 |
+
"grad_norm": 8.659878730773926,
|
| 138 |
+
"learning_rate": 1.775100401606426e-05,
|
| 139 |
+
"loss": 1.0936,
|
| 140 |
+
"step": 1400
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 1.8072289156626506,
|
| 144 |
+
"grad_norm": 4.855623245239258,
|
| 145 |
+
"learning_rate": 1.7590361445783134e-05,
|
| 146 |
+
"loss": 1.1401,
|
| 147 |
+
"step": 1500
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 1.927710843373494,
|
| 151 |
+
"grad_norm": 5.588471412658691,
|
| 152 |
+
"learning_rate": 1.742971887550201e-05,
|
| 153 |
+
"loss": 1.0588,
|
| 154 |
+
"step": 1600
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 2.0,
|
| 158 |
+
"eval_action_metrics": {
|
| 159 |
+
"accuracy": 0.6118143459915611,
|
| 160 |
+
"f1": 0.37830319888734354,
|
| 161 |
+
"loss": 0.3113616406917572,
|
| 162 |
+
"roc_auc": 0.6275566914121131
|
| 163 |
+
},
|
| 164 |
+
"eval_cause_metrics": {
|
| 165 |
+
"accuracy": 0.5750452079566004,
|
| 166 |
+
"f1": 0.20998278829604128,
|
| 167 |
+
"loss": 0.46545448899269104,
|
| 168 |
+
"roc_auc": 0.5587718371637969
|
| 169 |
+
},
|
| 170 |
+
"eval_event_metrics": {
|
| 171 |
+
"accuracy": 0.6913803496081977,
|
| 172 |
+
"f1": 0.7058823529411764,
|
| 173 |
+
"loss": 0.5396388173103333,
|
| 174 |
+
"roc_auc": 0.7942354115176712
|
| 175 |
+
},
|
| 176 |
+
"eval_f1": 0.5340634299499163,
|
| 177 |
+
"eval_loss": 1.181040644645691,
|
| 178 |
+
"eval_main_metrics": {
|
| 179 |
+
"accuracy": 0.6377335744424352,
|
| 180 |
+
"f1": 0.8420853796751038,
|
| 181 |
+
"loss": 1.2089903354644775,
|
| 182 |
+
"roc_auc": 0.8308065869363559
|
| 183 |
+
},
|
| 184 |
+
"eval_runtime": 19.4619,
|
| 185 |
+
"eval_samples_per_second": 85.243,
|
| 186 |
+
"eval_steps_per_second": 10.688,
|
| 187 |
+
"step": 1660
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 2.0481927710843375,
|
| 191 |
+
"grad_norm": 4.3426384925842285,
|
| 192 |
+
"learning_rate": 1.7269076305220884e-05,
|
| 193 |
+
"loss": 1.0881,
|
| 194 |
+
"step": 1700
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 2.1686746987951806,
|
| 198 |
+
"grad_norm": 15.190299987792969,
|
| 199 |
+
"learning_rate": 1.710843373493976e-05,
|
| 200 |
+
"loss": 0.9603,
|
| 201 |
+
"step": 1800
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 2.289156626506024,
|
| 205 |
+
"grad_norm": 10.794776916503906,
|
| 206 |
+
"learning_rate": 1.6947791164658637e-05,
|
| 207 |
+
"loss": 0.9664,
|
| 208 |
+
"step": 1900
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 2.4096385542168672,
|
| 212 |
+
"grad_norm": 10.562026977539062,
|
| 213 |
+
"learning_rate": 1.6787148594377512e-05,
|
| 214 |
+
"loss": 0.9303,
|
| 215 |
+
"step": 2000
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 2.5301204819277108,
|
| 219 |
+
"grad_norm": 4.81545877456665,
|
| 220 |
+
"learning_rate": 1.6626506024096387e-05,
|
| 221 |
+
"loss": 0.9407,
|
| 222 |
+
"step": 2100
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 2.6506024096385543,
|
| 226 |
+
"grad_norm": 4.499901294708252,
|
| 227 |
+
"learning_rate": 1.6465863453815262e-05,
|
| 228 |
+
"loss": 0.9614,
|
| 229 |
+
"step": 2200
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 2.7710843373493974,
|
| 233 |
+
"grad_norm": 4.810569763183594,
|
| 234 |
+
"learning_rate": 1.6305220883534137e-05,
|
| 235 |
+
"loss": 0.9482,
|
| 236 |
+
"step": 2300
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 2.891566265060241,
|
| 240 |
+
"grad_norm": 8.373404502868652,
|
| 241 |
+
"learning_rate": 1.6144578313253015e-05,
|
| 242 |
+
"loss": 0.8748,
|
| 243 |
+
"step": 2400
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 3.0,
|
| 247 |
+
"eval_action_metrics": {
|
| 248 |
+
"accuracy": 0.621458710066305,
|
| 249 |
+
"f1": 0.4234990316333118,
|
| 250 |
+
"loss": 0.31134167313575745,
|
| 251 |
+
"roc_auc": 0.6527716928319338
|
| 252 |
+
},
|
| 253 |
+
"eval_cause_metrics": {
|
| 254 |
+
"accuracy": 0.5822784810126582,
|
| 255 |
+
"f1": 0.2585924713584288,
|
| 256 |
+
"loss": 0.489061564207077,
|
| 257 |
+
"roc_auc": 0.5754498216809775
|
| 258 |
+
},
|
| 259 |
+
"eval_event_metrics": {
|
| 260 |
+
"accuracy": 0.6937914406268837,
|
| 261 |
+
"f1": 0.7111111111111111,
|
| 262 |
+
"loss": 0.6078017354011536,
|
| 263 |
+
"roc_auc": 0.798155101571195
|
| 264 |
+
},
|
| 265 |
+
"eval_f1": 0.5597986447252539,
|
| 266 |
+
"eval_loss": 1.1940369606018066,
|
| 267 |
+
"eval_main_metrics": {
|
| 268 |
+
"accuracy": 0.6497890295358649,
|
| 269 |
+
"f1": 0.8459919647981634,
|
| 270 |
+
"loss": 1.356093406677246,
|
| 271 |
+
"roc_auc": 0.8375618095347757
|
| 272 |
+
},
|
| 273 |
+
"eval_runtime": 19.4292,
|
| 274 |
+
"eval_samples_per_second": 85.387,
|
| 275 |
+
"eval_steps_per_second": 10.706,
|
| 276 |
+
"step": 2490
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 3.0120481927710845,
|
| 280 |
+
"grad_norm": 7.537188529968262,
|
| 281 |
+
"learning_rate": 1.5983935742971887e-05,
|
| 282 |
+
"loss": 0.9201,
|
| 283 |
+
"step": 2500
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 3.1325301204819276,
|
| 287 |
+
"grad_norm": 6.753769397735596,
|
| 288 |
+
"learning_rate": 1.5823293172690762e-05,
|
| 289 |
+
"loss": 0.7751,
|
| 290 |
+
"step": 2600
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 3.253012048192771,
|
| 294 |
+
"grad_norm": 6.366585731506348,
|
| 295 |
+
"learning_rate": 1.566265060240964e-05,
|
| 296 |
+
"loss": 0.7857,
|
| 297 |
+
"step": 2700
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 3.3734939759036147,
|
| 301 |
+
"grad_norm": 8.141886711120605,
|
| 302 |
+
"learning_rate": 1.5502008032128516e-05,
|
| 303 |
+
"loss": 0.7844,
|
| 304 |
+
"step": 2800
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 3.4939759036144578,
|
| 308 |
+
"grad_norm": 12.003459930419922,
|
| 309 |
+
"learning_rate": 1.534136546184739e-05,
|
| 310 |
+
"loss": 0.7853,
|
| 311 |
+
"step": 2900
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 3.6144578313253013,
|
| 315 |
+
"grad_norm": 4.899393558502197,
|
| 316 |
+
"learning_rate": 1.5180722891566266e-05,
|
| 317 |
+
"loss": 0.7652,
|
| 318 |
+
"step": 3000
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 3.734939759036145,
|
| 322 |
+
"grad_norm": 7.926997661590576,
|
| 323 |
+
"learning_rate": 1.5020080321285142e-05,
|
| 324 |
+
"loss": 0.7908,
|
| 325 |
+
"step": 3100
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 3.855421686746988,
|
| 329 |
+
"grad_norm": 3.9363718032836914,
|
| 330 |
+
"learning_rate": 1.4859437751004017e-05,
|
| 331 |
+
"loss": 0.8142,
|
| 332 |
+
"step": 3200
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 3.9759036144578315,
|
| 336 |
+
"grad_norm": 4.785311222076416,
|
| 337 |
+
"learning_rate": 1.4698795180722894e-05,
|
| 338 |
+
"loss": 0.75,
|
| 339 |
+
"step": 3300
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 4.0,
|
| 343 |
+
"eval_action_metrics": {
|
| 344 |
+
"accuracy": 0.6262808921036769,
|
| 345 |
+
"f1": 0.47516457211250746,
|
| 346 |
+
"loss": 0.3067108690738678,
|
| 347 |
+
"roc_auc": 0.6846639081578841
|
| 348 |
+
},
|
| 349 |
+
"eval_cause_metrics": {
|
| 350 |
+
"accuracy": 0.6100060277275468,
|
| 351 |
+
"f1": 0.38606271777003487,
|
| 352 |
+
"loss": 0.5579449534416199,
|
| 353 |
+
"roc_auc": 0.6297609166955901
|
| 354 |
+
},
|
| 355 |
+
"eval_event_metrics": {
|
| 356 |
+
"accuracy": 0.6955997588908981,
|
| 357 |
+
"f1": 0.7214808437365476,
|
| 358 |
+
"loss": 0.6555235981941223,
|
| 359 |
+
"roc_auc": 0.8122881121542476
|
| 360 |
+
},
|
| 361 |
+
"eval_f1": 0.6080738826170755,
|
| 362 |
+
"eval_loss": 1.2421071529388428,
|
| 363 |
+
"eval_main_metrics": {
|
| 364 |
+
"accuracy": 0.6594333936106088,
|
| 365 |
+
"f1": 0.8495873968492123,
|
| 366 |
+
"loss": 1.4700279235839844,
|
| 367 |
+
"roc_auc": 0.8373546457094286
|
| 368 |
+
},
|
| 369 |
+
"eval_runtime": 19.3706,
|
| 370 |
+
"eval_samples_per_second": 85.645,
|
| 371 |
+
"eval_steps_per_second": 10.738,
|
| 372 |
+
"step": 3320
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 4.096385542168675,
|
| 376 |
+
"grad_norm": 2.819312334060669,
|
| 377 |
+
"learning_rate": 1.4538152610441769e-05,
|
| 378 |
+
"loss": 0.6905,
|
| 379 |
+
"step": 3400
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 4.216867469879518,
|
| 383 |
+
"grad_norm": 15.132315635681152,
|
| 384 |
+
"learning_rate": 1.4377510040160642e-05,
|
| 385 |
+
"loss": 0.66,
|
| 386 |
+
"step": 3500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 4.337349397590361,
|
| 390 |
+
"grad_norm": 13.252199172973633,
|
| 391 |
+
"learning_rate": 1.4216867469879519e-05,
|
| 392 |
+
"loss": 0.6854,
|
| 393 |
+
"step": 3600
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 4.457831325301205,
|
| 397 |
+
"grad_norm": 3.5887160301208496,
|
| 398 |
+
"learning_rate": 1.4056224899598394e-05,
|
| 399 |
+
"loss": 0.6423,
|
| 400 |
+
"step": 3700
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 4.578313253012048,
|
| 404 |
+
"grad_norm": 3.583491563796997,
|
| 405 |
+
"learning_rate": 1.3895582329317269e-05,
|
| 406 |
+
"loss": 0.6372,
|
| 407 |
+
"step": 3800
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 4.698795180722891,
|
| 411 |
+
"grad_norm": 9.358288764953613,
|
| 412 |
+
"learning_rate": 1.3734939759036146e-05,
|
| 413 |
+
"loss": 0.6457,
|
| 414 |
+
"step": 3900
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 4.8192771084337345,
|
| 418 |
+
"grad_norm": 6.61803674697876,
|
| 419 |
+
"learning_rate": 1.357429718875502e-05,
|
| 420 |
+
"loss": 0.6402,
|
| 421 |
+
"step": 4000
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 4.9397590361445785,
|
| 425 |
+
"grad_norm": 12.123319625854492,
|
| 426 |
+
"learning_rate": 1.3413654618473897e-05,
|
| 427 |
+
"loss": 0.6447,
|
| 428 |
+
"step": 4100
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 5.0,
|
| 432 |
+
"eval_action_metrics": {
|
| 433 |
+
"accuracy": 0.6274864376130199,
|
| 434 |
+
"f1": 0.47335025380710655,
|
| 435 |
+
"loss": 0.4026564955711365,
|
| 436 |
+
"roc_auc": 0.6765029816234635
|
| 437 |
+
},
|
| 438 |
+
"eval_cause_metrics": {
|
| 439 |
+
"accuracy": 0.6244725738396625,
|
| 440 |
+
"f1": 0.42625169147496617,
|
| 441 |
+
"loss": 0.5328236222267151,
|
| 442 |
+
"roc_auc": 0.6485007620686012
|
| 443 |
+
},
|
| 444 |
+
"eval_event_metrics": {
|
| 445 |
+
"accuracy": 0.6949969861362266,
|
| 446 |
+
"f1": 0.7206896551724137,
|
| 447 |
+
"loss": 0.7433874011039734,
|
| 448 |
+
"roc_auc": 0.8115834972530086
|
| 449 |
+
},
|
| 450 |
+
"eval_f1": 0.6173571829119144,
|
| 451 |
+
"eval_loss": 1.272179365158081,
|
| 452 |
+
"eval_main_metrics": {
|
| 453 |
+
"accuracy": 0.650994575045208,
|
| 454 |
+
"f1": 0.8491371311931712,
|
| 455 |
+
"loss": 1.7594246864318848,
|
| 456 |
+
"roc_auc": 0.8347010566163695
|
| 457 |
+
},
|
| 458 |
+
"eval_runtime": 19.3728,
|
| 459 |
+
"eval_samples_per_second": 85.635,
|
| 460 |
+
"eval_steps_per_second": 10.737,
|
| 461 |
+
"step": 4150
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 5.0602409638554215,
|
| 465 |
+
"grad_norm": 4.934209823608398,
|
| 466 |
+
"learning_rate": 1.3253012048192772e-05,
|
| 467 |
+
"loss": 0.5871,
|
| 468 |
+
"step": 4200
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 5.180722891566265,
|
| 472 |
+
"grad_norm": 3.2406387329101562,
|
| 473 |
+
"learning_rate": 1.309236947791165e-05,
|
| 474 |
+
"loss": 0.5427,
|
| 475 |
+
"step": 4300
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 5.301204819277109,
|
| 479 |
+
"grad_norm": 5.232382774353027,
|
| 480 |
+
"learning_rate": 1.2931726907630524e-05,
|
| 481 |
+
"loss": 0.5588,
|
| 482 |
+
"step": 4400
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 5.421686746987952,
|
| 486 |
+
"grad_norm": 14.17167854309082,
|
| 487 |
+
"learning_rate": 1.2771084337349398e-05,
|
| 488 |
+
"loss": 0.5816,
|
| 489 |
+
"step": 4500
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 5.542168674698795,
|
| 493 |
+
"grad_norm": 15.979377746582031,
|
| 494 |
+
"learning_rate": 1.2610441767068273e-05,
|
| 495 |
+
"loss": 0.5946,
|
| 496 |
+
"step": 4600
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 5.662650602409639,
|
| 500 |
+
"grad_norm": 7.11401891708374,
|
| 501 |
+
"learning_rate": 1.244979919678715e-05,
|
| 502 |
+
"loss": 0.5395,
|
| 503 |
+
"step": 4700
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 5.783132530120482,
|
| 507 |
+
"grad_norm": 2.767531156539917,
|
| 508 |
+
"learning_rate": 1.2289156626506024e-05,
|
| 509 |
+
"loss": 0.5364,
|
| 510 |
+
"step": 4800
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 5.903614457831325,
|
| 514 |
+
"grad_norm": 4.462075233459473,
|
| 515 |
+
"learning_rate": 1.2128514056224901e-05,
|
| 516 |
+
"loss": 0.5586,
|
| 517 |
+
"step": 4900
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 6.0,
|
| 521 |
+
"eval_action_metrics": {
|
| 522 |
+
"accuracy": 0.6359252561784208,
|
| 523 |
+
"f1": 0.4963144963144963,
|
| 524 |
+
"loss": 0.3993087410926819,
|
| 525 |
+
"roc_auc": 0.6909881952050626
|
| 526 |
+
},
|
| 527 |
+
"eval_cause_metrics": {
|
| 528 |
+
"accuracy": 0.6172393007836046,
|
| 529 |
+
"f1": 0.4432786885245902,
|
| 530 |
+
"loss": 0.5477549433708191,
|
| 531 |
+
"roc_auc": 0.6586722400792753
|
| 532 |
+
},
|
| 533 |
+
"eval_event_metrics": {
|
| 534 |
+
"accuracy": 0.6895720313441832,
|
| 535 |
+
"f1": 0.7254658385093167,
|
| 536 |
+
"loss": 0.775439441204071,
|
| 537 |
+
"roc_auc": 0.8210486138922756
|
| 538 |
+
},
|
| 539 |
+
"eval_f1": 0.628022957090326,
|
| 540 |
+
"eval_loss": 1.3330581188201904,
|
| 541 |
+
"eval_main_metrics": {
|
| 542 |
+
"accuracy": 0.6377335744424352,
|
| 543 |
+
"f1": 0.8470328050129008,
|
| 544 |
+
"loss": 1.7559667825698853,
|
| 545 |
+
"roc_auc": 0.8309945868949555
|
| 546 |
+
},
|
| 547 |
+
"eval_runtime": 19.4268,
|
| 548 |
+
"eval_samples_per_second": 85.398,
|
| 549 |
+
"eval_steps_per_second": 10.707,
|
| 550 |
+
"step": 4980
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 6.024096385542169,
|
| 554 |
+
"grad_norm": 4.994393825531006,
|
| 555 |
+
"learning_rate": 1.1967871485943776e-05,
|
| 556 |
+
"loss": 0.5326,
|
| 557 |
+
"step": 5000
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 6.144578313253012,
|
| 561 |
+
"grad_norm": 2.5982210636138916,
|
| 562 |
+
"learning_rate": 1.1807228915662651e-05,
|
| 563 |
+
"loss": 0.4935,
|
| 564 |
+
"step": 5100
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 6.265060240963855,
|
| 568 |
+
"grad_norm": 9.731000900268555,
|
| 569 |
+
"learning_rate": 1.1646586345381528e-05,
|
| 570 |
+
"loss": 0.4444,
|
| 571 |
+
"step": 5200
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 6.385542168674699,
|
| 575 |
+
"grad_norm": 3.476001501083374,
|
| 576 |
+
"learning_rate": 1.1485943775100403e-05,
|
| 577 |
+
"loss": 0.4772,
|
| 578 |
+
"step": 5300
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 6.506024096385542,
|
| 582 |
+
"grad_norm": 9.325688362121582,
|
| 583 |
+
"learning_rate": 1.132530120481928e-05,
|
| 584 |
+
"loss": 0.4758,
|
| 585 |
+
"step": 5400
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 6.626506024096385,
|
| 589 |
+
"grad_norm": 10.23697566986084,
|
| 590 |
+
"learning_rate": 1.1164658634538153e-05,
|
| 591 |
+
"loss": 0.467,
|
| 592 |
+
"step": 5500
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 6.746987951807229,
|
| 596 |
+
"grad_norm": 6.919719219207764,
|
| 597 |
+
"learning_rate": 1.1004016064257028e-05,
|
| 598 |
+
"loss": 0.4596,
|
| 599 |
+
"step": 5600
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 6.867469879518072,
|
| 603 |
+
"grad_norm": 3.161492347717285,
|
| 604 |
+
"learning_rate": 1.0843373493975904e-05,
|
| 605 |
+
"loss": 0.4901,
|
| 606 |
+
"step": 5700
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 6.9879518072289155,
|
| 610 |
+
"grad_norm": 5.633391857147217,
|
| 611 |
+
"learning_rate": 1.068273092369478e-05,
|
| 612 |
+
"loss": 0.5053,
|
| 613 |
+
"step": 5800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 7.0,
|
| 617 |
+
"eval_action_metrics": {
|
| 618 |
+
"accuracy": 0.6347197106690777,
|
| 619 |
+
"f1": 0.4936014625228519,
|
| 620 |
+
"loss": 0.44190824031829834,
|
| 621 |
+
"roc_auc": 0.6908239016672751
|
| 622 |
+
},
|
| 623 |
+
"eval_cause_metrics": {
|
| 624 |
+
"accuracy": 0.6208559373116335,
|
| 625 |
+
"f1": 0.4570337364735837,
|
| 626 |
+
"loss": 0.5815909504890442,
|
| 627 |
+
"roc_auc": 0.6677868823597467
|
| 628 |
+
},
|
| 629 |
+
"eval_event_metrics": {
|
| 630 |
+
"accuracy": 0.6714888487040386,
|
| 631 |
+
"f1": 0.7207713941341904,
|
| 632 |
+
"loss": 0.7780100703239441,
|
| 633 |
+
"roc_auc": 0.8228283501014495
|
| 634 |
+
},
|
| 635 |
+
"eval_f1": 0.6299279587105135,
|
| 636 |
+
"eval_loss": 1.403409481048584,
|
| 637 |
+
"eval_main_metrics": {
|
| 638 |
+
"accuracy": 0.6443640747438216,
|
| 639 |
+
"f1": 0.8483052417114281,
|
| 640 |
+
"loss": 2.0327484607696533,
|
| 641 |
+
"roc_auc": 0.8334165115229015
|
| 642 |
+
},
|
| 643 |
+
"eval_runtime": 19.4509,
|
| 644 |
+
"eval_samples_per_second": 85.292,
|
| 645 |
+
"eval_steps_per_second": 10.694,
|
| 646 |
+
"step": 5810
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 7.108433734939759,
|
| 650 |
+
"grad_norm": 2.695713996887207,
|
| 651 |
+
"learning_rate": 1.0522088353413654e-05,
|
| 652 |
+
"loss": 0.4098,
|
| 653 |
+
"step": 5900
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 7.228915662650603,
|
| 657 |
+
"grad_norm": 3.300992727279663,
|
| 658 |
+
"learning_rate": 1.0361445783132531e-05,
|
| 659 |
+
"loss": 0.4406,
|
| 660 |
+
"step": 6000
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 7.349397590361446,
|
| 664 |
+
"grad_norm": 9.012105941772461,
|
| 665 |
+
"learning_rate": 1.0200803212851406e-05,
|
| 666 |
+
"loss": 0.3848,
|
| 667 |
+
"step": 6100
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 7.469879518072289,
|
| 671 |
+
"grad_norm": 1.8880501985549927,
|
| 672 |
+
"learning_rate": 1.0040160642570283e-05,
|
| 673 |
+
"loss": 0.4086,
|
| 674 |
+
"step": 6200
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 7.590361445783133,
|
| 678 |
+
"grad_norm": 6.892731666564941,
|
| 679 |
+
"learning_rate": 9.879518072289156e-06,
|
| 680 |
+
"loss": 0.4284,
|
| 681 |
+
"step": 6300
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 7.710843373493976,
|
| 685 |
+
"grad_norm": 7.901873588562012,
|
| 686 |
+
"learning_rate": 9.718875502008033e-06,
|
| 687 |
+
"loss": 0.45,
|
| 688 |
+
"step": 6400
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 7.831325301204819,
|
| 692 |
+
"grad_norm": 4.076874256134033,
|
| 693 |
+
"learning_rate": 9.558232931726908e-06,
|
| 694 |
+
"loss": 0.4319,
|
| 695 |
+
"step": 6500
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 7.951807228915663,
|
| 699 |
+
"grad_norm": 4.073435306549072,
|
| 700 |
+
"learning_rate": 9.397590361445785e-06,
|
| 701 |
+
"loss": 0.42,
|
| 702 |
+
"step": 6600
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 8.0,
|
| 706 |
+
"eval_action_metrics": {
|
| 707 |
+
"accuracy": 0.6395418927064497,
|
| 708 |
+
"f1": 0.5042735042735043,
|
| 709 |
+
"loss": 0.45456641912460327,
|
| 710 |
+
"roc_auc": 0.6954778710802808
|
| 711 |
+
},
|
| 712 |
+
"eval_cause_metrics": {
|
| 713 |
+
"accuracy": 0.6298975286317059,
|
| 714 |
+
"f1": 0.4623179227359088,
|
| 715 |
+
"loss": 0.5961005687713623,
|
| 716 |
+
"roc_auc": 0.6706782159043465
|
| 717 |
+
},
|
| 718 |
+
"eval_event_metrics": {
|
| 719 |
+
"accuracy": 0.6853526220614828,
|
| 720 |
+
"f1": 0.7223140495867769,
|
| 721 |
+
"loss": 0.8032154440879822,
|
| 722 |
+
"roc_auc": 0.8192863575314914
|
| 723 |
+
},
|
| 724 |
+
"eval_f1": 0.6344324161815224,
|
| 725 |
+
"eval_loss": 1.4469618797302246,
|
| 726 |
+
"eval_main_metrics": {
|
| 727 |
+
"accuracy": 0.646172393007836,
|
| 728 |
+
"f1": 0.8488241881298993,
|
| 729 |
+
"loss": 2.039750337600708,
|
| 730 |
+
"roc_auc": 0.835544105032868
|
| 731 |
+
},
|
| 732 |
+
"eval_runtime": 19.422,
|
| 733 |
+
"eval_samples_per_second": 85.419,
|
| 734 |
+
"eval_steps_per_second": 10.71,
|
| 735 |
+
"step": 6640
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 8.072289156626505,
|
| 739 |
+
"grad_norm": 2.453636407852173,
|
| 740 |
+
"learning_rate": 9.23694779116466e-06,
|
| 741 |
+
"loss": 0.3554,
|
| 742 |
+
"step": 6700
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 8.19277108433735,
|
| 746 |
+
"grad_norm": 3.2704885005950928,
|
| 747 |
+
"learning_rate": 9.076305220883535e-06,
|
| 748 |
+
"loss": 0.3916,
|
| 749 |
+
"step": 6800
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 8.313253012048193,
|
| 753 |
+
"grad_norm": 5.025237083435059,
|
| 754 |
+
"learning_rate": 8.91566265060241e-06,
|
| 755 |
+
"loss": 0.3619,
|
| 756 |
+
"step": 6900
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 8.433734939759036,
|
| 760 |
+
"grad_norm": 7.339418888092041,
|
| 761 |
+
"learning_rate": 8.755020080321286e-06,
|
| 762 |
+
"loss": 0.345,
|
| 763 |
+
"step": 7000
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 8.55421686746988,
|
| 767 |
+
"grad_norm": 11.077627182006836,
|
| 768 |
+
"learning_rate": 8.594377510040161e-06,
|
| 769 |
+
"loss": 0.3704,
|
| 770 |
+
"step": 7100
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 8.674698795180722,
|
| 774 |
+
"grad_norm": 4.094336986541748,
|
| 775 |
+
"learning_rate": 8.433734939759038e-06,
|
| 776 |
+
"loss": 0.4081,
|
| 777 |
+
"step": 7200
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 8.795180722891565,
|
| 781 |
+
"grad_norm": 2.9691665172576904,
|
| 782 |
+
"learning_rate": 8.273092369477911e-06,
|
| 783 |
+
"loss": 0.3692,
|
| 784 |
+
"step": 7300
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 8.91566265060241,
|
| 788 |
+
"grad_norm": 2.072866916656494,
|
| 789 |
+
"learning_rate": 8.112449799196788e-06,
|
| 790 |
+
"loss": 0.3552,
|
| 791 |
+
"step": 7400
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 9.0,
|
| 795 |
+
"eval_action_metrics": {
|
| 796 |
+
"accuracy": 0.6305003013863774,
|
| 797 |
+
"f1": 0.509478672985782,
|
| 798 |
+
"loss": 0.4554847180843353,
|
| 799 |
+
"roc_auc": 0.7022230335483347
|
| 800 |
+
},
|
| 801 |
+
"eval_cause_metrics": {
|
| 802 |
+
"accuracy": 0.6106088004822182,
|
| 803 |
+
"f1": 0.43440594059405946,
|
| 804 |
+
"loss": 0.6525366902351379,
|
| 805 |
+
"roc_auc": 0.6608236377080599
|
| 806 |
+
},
|
| 807 |
+
"eval_event_metrics": {
|
| 808 |
+
"accuracy": 0.6799276672694394,
|
| 809 |
+
"f1": 0.724025974025974,
|
| 810 |
+
"loss": 0.9302711486816406,
|
| 811 |
+
"roc_auc": 0.8233804035411074
|
| 812 |
+
},
|
| 813 |
+
"eval_f1": 0.6287379744042451,
|
| 814 |
+
"eval_loss": 1.5192142724990845,
|
| 815 |
+
"eval_main_metrics": {
|
| 816 |
+
"accuracy": 0.6449668474984931,
|
| 817 |
+
"f1": 0.8470413100111649,
|
| 818 |
+
"loss": 2.102555513381958,
|
| 819 |
+
"roc_auc": 0.8330059034412807
|
| 820 |
+
},
|
| 821 |
+
"eval_runtime": 19.4322,
|
| 822 |
+
"eval_samples_per_second": 85.374,
|
| 823 |
+
"eval_steps_per_second": 10.704,
|
| 824 |
+
"step": 7470
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 9.036144578313253,
|
| 828 |
+
"grad_norm": 6.187836170196533,
|
| 829 |
+
"learning_rate": 7.951807228915663e-06,
|
| 830 |
+
"loss": 0.3559,
|
| 831 |
+
"step": 7500
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 9.156626506024097,
|
| 835 |
+
"grad_norm": 9.382346153259277,
|
| 836 |
+
"learning_rate": 7.79116465863454e-06,
|
| 837 |
+
"loss": 0.3518,
|
| 838 |
+
"step": 7600
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 9.27710843373494,
|
| 842 |
+
"grad_norm": 3.6472980976104736,
|
| 843 |
+
"learning_rate": 7.630522088353415e-06,
|
| 844 |
+
"loss": 0.3199,
|
| 845 |
+
"step": 7700
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 9.397590361445783,
|
| 849 |
+
"grad_norm": 6.128402233123779,
|
| 850 |
+
"learning_rate": 7.469879518072289e-06,
|
| 851 |
+
"loss": 0.3212,
|
| 852 |
+
"step": 7800
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 9.518072289156626,
|
| 856 |
+
"grad_norm": 2.361755132675171,
|
| 857 |
+
"learning_rate": 7.309236947791165e-06,
|
| 858 |
+
"loss": 0.3245,
|
| 859 |
+
"step": 7900
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 9.638554216867469,
|
| 863 |
+
"grad_norm": 4.379687786102295,
|
| 864 |
+
"learning_rate": 7.148594377510041e-06,
|
| 865 |
+
"loss": 0.3294,
|
| 866 |
+
"step": 8000
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 9.759036144578314,
|
| 870 |
+
"grad_norm": 6.3661370277404785,
|
| 871 |
+
"learning_rate": 6.987951807228917e-06,
|
| 872 |
+
"loss": 0.3325,
|
| 873 |
+
"step": 8100
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 9.879518072289157,
|
| 877 |
+
"grad_norm": 2.1560158729553223,
|
| 878 |
+
"learning_rate": 6.8273092369477925e-06,
|
| 879 |
+
"loss": 0.3192,
|
| 880 |
+
"step": 8200
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 10.0,
|
| 884 |
+
"grad_norm": 4.878142356872559,
|
| 885 |
+
"learning_rate": 6.666666666666667e-06,
|
| 886 |
+
"loss": 0.3227,
|
| 887 |
+
"step": 8300
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 10.0,
|
| 891 |
+
"eval_action_metrics": {
|
| 892 |
+
"accuracy": 0.6317058468957203,
|
| 893 |
+
"f1": 0.5257076834199884,
|
| 894 |
+
"loss": 0.3950282335281372,
|
| 895 |
+
"roc_auc": 0.7138452801103403
|
| 896 |
+
},
|
| 897 |
+
"eval_cause_metrics": {
|
| 898 |
+
"accuracy": 0.6190476190476191,
|
| 899 |
+
"f1": 0.48321342925659466,
|
| 900 |
+
"loss": 0.6035521626472473,
|
| 901 |
+
"roc_auc": 0.6868496637340858
|
| 902 |
+
},
|
| 903 |
+
"eval_event_metrics": {
|
| 904 |
+
"accuracy": 0.6949969861362266,
|
| 905 |
+
"f1": 0.73347022587269,
|
| 906 |
+
"loss": 1.0751080513000488,
|
| 907 |
+
"roc_auc": 0.8277649690562431
|
| 908 |
+
},
|
| 909 |
+
"eval_f1": 0.6476335229111712,
|
| 910 |
+
"eval_loss": 1.5520377159118652,
|
| 911 |
+
"eval_main_metrics": {
|
| 912 |
+
"accuracy": 0.6353224834237492,
|
| 913 |
+
"f1": 0.8481427530954114,
|
| 914 |
+
"loss": 2.005683422088623,
|
| 915 |
+
"roc_auc": 0.8296800426683087
|
| 916 |
+
},
|
| 917 |
+
"eval_runtime": 19.4155,
|
| 918 |
+
"eval_samples_per_second": 85.447,
|
| 919 |
+
"eval_steps_per_second": 10.713,
|
| 920 |
+
"step": 8300
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 10.120481927710843,
|
| 924 |
+
"grad_norm": 4.528349876403809,
|
| 925 |
+
"learning_rate": 6.5060240963855425e-06,
|
| 926 |
+
"loss": 0.3223,
|
| 927 |
+
"step": 8400
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 10.240963855421686,
|
| 931 |
+
"grad_norm": 6.405284881591797,
|
| 932 |
+
"learning_rate": 6.345381526104418e-06,
|
| 933 |
+
"loss": 0.2805,
|
| 934 |
+
"step": 8500
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 10.36144578313253,
|
| 938 |
+
"grad_norm": 4.246714115142822,
|
| 939 |
+
"learning_rate": 6.184738955823294e-06,
|
| 940 |
+
"loss": 0.2912,
|
| 941 |
+
"step": 8600
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 10.481927710843374,
|
| 945 |
+
"grad_norm": 2.8840444087982178,
|
| 946 |
+
"learning_rate": 6.02409638554217e-06,
|
| 947 |
+
"loss": 0.303,
|
| 948 |
+
"step": 8700
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 10.602409638554217,
|
| 952 |
+
"grad_norm": 6.155231952667236,
|
| 953 |
+
"learning_rate": 5.863453815261044e-06,
|
| 954 |
+
"loss": 0.289,
|
| 955 |
+
"step": 8800
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 10.72289156626506,
|
| 959 |
+
"grad_norm": 5.590630531311035,
|
| 960 |
+
"learning_rate": 5.70281124497992e-06,
|
| 961 |
+
"loss": 0.2866,
|
| 962 |
+
"step": 8900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 10.843373493975903,
|
| 966 |
+
"grad_norm": 2.1477112770080566,
|
| 967 |
+
"learning_rate": 5.542168674698796e-06,
|
| 968 |
+
"loss": 0.2908,
|
| 969 |
+
"step": 9000
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 10.963855421686747,
|
| 973 |
+
"grad_norm": 3.363445520401001,
|
| 974 |
+
"learning_rate": 5.381526104417672e-06,
|
| 975 |
+
"loss": 0.3135,
|
| 976 |
+
"step": 9100
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 11.0,
|
| 980 |
+
"eval_action_metrics": {
|
| 981 |
+
"accuracy": 0.6365280289330922,
|
| 982 |
+
"f1": 0.5067319461444308,
|
| 983 |
+
"loss": 0.4702241122722626,
|
| 984 |
+
"roc_auc": 0.6962618149365137
|
| 985 |
+
},
|
| 986 |
+
"eval_cause_metrics": {
|
| 987 |
+
"accuracy": 0.621458710066305,
|
| 988 |
+
"f1": 0.4755501222493888,
|
| 989 |
+
"loss": 0.6331600546836853,
|
| 990 |
+
"roc_auc": 0.680847681852707
|
| 991 |
+
},
|
| 992 |
+
"eval_event_metrics": {
|
| 993 |
+
"accuracy": 0.6943942133815552,
|
| 994 |
+
"f1": 0.73487616727568,
|
| 995 |
+
"loss": 0.8929408192634583,
|
| 996 |
+
"roc_auc": 0.8306706066392446
|
| 997 |
+
},
|
| 998 |
+
"eval_f1": 0.6414051572871415,
|
| 999 |
+
"eval_loss": 1.5805182456970215,
|
| 1000 |
+
"eval_main_metrics": {
|
| 1001 |
+
"accuracy": 0.6401446654611211,
|
| 1002 |
+
"f1": 0.8484623934790664,
|
| 1003 |
+
"loss": 2.2667479515075684,
|
| 1004 |
+
"roc_auc": 0.8336254542725883
|
| 1005 |
+
},
|
| 1006 |
+
"eval_runtime": 19.3897,
|
| 1007 |
+
"eval_samples_per_second": 85.561,
|
| 1008 |
+
"eval_steps_per_second": 10.727,
|
| 1009 |
+
"step": 9130
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 11.08433734939759,
|
| 1013 |
+
"grad_norm": 3.677710771560669,
|
| 1014 |
+
"learning_rate": 5.220883534136547e-06,
|
| 1015 |
+
"loss": 0.2626,
|
| 1016 |
+
"step": 9200
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 11.204819277108435,
|
| 1020 |
+
"grad_norm": 3.71104097366333,
|
| 1021 |
+
"learning_rate": 5.060240963855422e-06,
|
| 1022 |
+
"loss": 0.2811,
|
| 1023 |
+
"step": 9300
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 11.325301204819278,
|
| 1027 |
+
"grad_norm": 4.658101558685303,
|
| 1028 |
+
"learning_rate": 4.899598393574298e-06,
|
| 1029 |
+
"loss": 0.2556,
|
| 1030 |
+
"step": 9400
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 11.44578313253012,
|
| 1034 |
+
"grad_norm": 2.3544187545776367,
|
| 1035 |
+
"learning_rate": 4.7389558232931736e-06,
|
| 1036 |
+
"loss": 0.2675,
|
| 1037 |
+
"step": 9500
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 11.566265060240964,
|
| 1041 |
+
"grad_norm": 2.9858105182647705,
|
| 1042 |
+
"learning_rate": 4.578313253012049e-06,
|
| 1043 |
+
"loss": 0.2601,
|
| 1044 |
+
"step": 9600
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 11.686746987951807,
|
| 1048 |
+
"grad_norm": 5.517003059387207,
|
| 1049 |
+
"learning_rate": 4.4176706827309244e-06,
|
| 1050 |
+
"loss": 0.2706,
|
| 1051 |
+
"step": 9700
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 11.80722891566265,
|
| 1055 |
+
"grad_norm": 5.479954242706299,
|
| 1056 |
+
"learning_rate": 4.2570281124497995e-06,
|
| 1057 |
+
"loss": 0.2678,
|
| 1058 |
+
"step": 9800
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 11.927710843373493,
|
| 1062 |
+
"grad_norm": 3.5991973876953125,
|
| 1063 |
+
"learning_rate": 4.096385542168675e-06,
|
| 1064 |
+
"loss": 0.2555,
|
| 1065 |
+
"step": 9900
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 12.0,
|
| 1069 |
+
"eval_action_metrics": {
|
| 1070 |
+
"accuracy": 0.6329113924050633,
|
| 1071 |
+
"f1": 0.5242030696576152,
|
| 1072 |
+
"loss": 0.41705745458602905,
|
| 1073 |
+
"roc_auc": 0.7097399699809339
|
| 1074 |
+
},
|
| 1075 |
+
"eval_cause_metrics": {
|
| 1076 |
+
"accuracy": 0.6262808921036769,
|
| 1077 |
+
"f1": 0.5017221584385763,
|
| 1078 |
+
"loss": 0.6140196919441223,
|
| 1079 |
+
"roc_auc": 0.7016332915830403
|
| 1080 |
+
},
|
| 1081 |
+
"eval_event_metrics": {
|
| 1082 |
+
"accuracy": 0.6871609403254972,
|
| 1083 |
+
"f1": 0.7303643724696356,
|
| 1084 |
+
"loss": 1.0149368047714233,
|
| 1085 |
+
"roc_auc": 0.8280933246892902
|
| 1086 |
+
},
|
| 1087 |
+
"eval_f1": 0.6511378774267244,
|
| 1088 |
+
"eval_loss": 1.6208207607269287,
|
| 1089 |
+
"eval_main_metrics": {
|
| 1090 |
+
"accuracy": 0.6395418927064497,
|
| 1091 |
+
"f1": 0.8482619091410705,
|
| 1092 |
+
"loss": 2.2304847240448,
|
| 1093 |
+
"roc_auc": 0.8319157462619947
|
| 1094 |
+
},
|
| 1095 |
+
"eval_runtime": 19.4554,
|
| 1096 |
+
"eval_samples_per_second": 85.272,
|
| 1097 |
+
"eval_steps_per_second": 10.691,
|
| 1098 |
+
"step": 9960
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 12.048192771084338,
|
| 1102 |
+
"grad_norm": 4.944530010223389,
|
| 1103 |
+
"learning_rate": 3.93574297188755e-06,
|
| 1104 |
+
"loss": 0.2796,
|
| 1105 |
+
"step": 10000
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 12.168674698795181,
|
| 1109 |
+
"grad_norm": 4.537522792816162,
|
| 1110 |
+
"learning_rate": 3.7751004016064258e-06,
|
| 1111 |
+
"loss": 0.2397,
|
| 1112 |
+
"step": 10100
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 12.289156626506024,
|
| 1116 |
+
"grad_norm": 5.054214954376221,
|
| 1117 |
+
"learning_rate": 3.6144578313253016e-06,
|
| 1118 |
+
"loss": 0.2311,
|
| 1119 |
+
"step": 10200
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 12.409638554216867,
|
| 1123 |
+
"grad_norm": 1.6431657075881958,
|
| 1124 |
+
"learning_rate": 3.453815261044177e-06,
|
| 1125 |
+
"loss": 0.2515,
|
| 1126 |
+
"step": 10300
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 12.53012048192771,
|
| 1130 |
+
"grad_norm": 2.898925304412842,
|
| 1131 |
+
"learning_rate": 3.2931726907630525e-06,
|
| 1132 |
+
"loss": 0.2424,
|
| 1133 |
+
"step": 10400
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 12.650602409638553,
|
| 1137 |
+
"grad_norm": 3.5978894233703613,
|
| 1138 |
+
"learning_rate": 3.132530120481928e-06,
|
| 1139 |
+
"loss": 0.2573,
|
| 1140 |
+
"step": 10500
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 12.771084337349398,
|
| 1144 |
+
"grad_norm": 3.019956588745117,
|
| 1145 |
+
"learning_rate": 2.9718875502008034e-06,
|
| 1146 |
+
"loss": 0.2469,
|
| 1147 |
+
"step": 10600
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 12.891566265060241,
|
| 1151 |
+
"grad_norm": 4.129029273986816,
|
| 1152 |
+
"learning_rate": 2.811244979919679e-06,
|
| 1153 |
+
"loss": 0.2377,
|
| 1154 |
+
"step": 10700
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 13.0,
|
| 1158 |
+
"eval_action_metrics": {
|
| 1159 |
+
"accuracy": 0.6298975286317059,
|
| 1160 |
+
"f1": 0.5311438278595696,
|
| 1161 |
+
"loss": 0.40619078278541565,
|
| 1162 |
+
"roc_auc": 0.7197446351060808
|
| 1163 |
+
},
|
| 1164 |
+
"eval_cause_metrics": {
|
| 1165 |
+
"accuracy": 0.6220614828209765,
|
| 1166 |
+
"f1": 0.5011520737327189,
|
| 1167 |
+
"loss": 0.6362375617027283,
|
| 1168 |
+
"roc_auc": 0.7008556294988456
|
| 1169 |
+
},
|
| 1170 |
+
"eval_event_metrics": {
|
| 1171 |
+
"accuracy": 0.6841470765521398,
|
| 1172 |
+
"f1": 0.729192042224929,
|
| 1173 |
+
"loss": 1.0542970895767212,
|
| 1174 |
+
"roc_auc": 0.8268163000692555
|
| 1175 |
+
},
|
| 1176 |
+
"eval_f1": 0.6518611919550361,
|
| 1177 |
+
"eval_loss": 1.6575490236282349,
|
| 1178 |
+
"eval_main_metrics": {
|
| 1179 |
+
"accuracy": 0.6305003013863774,
|
| 1180 |
+
"f1": 0.8459568240029272,
|
| 1181 |
+
"loss": 2.175412654876709,
|
| 1182 |
+
"roc_auc": 0.8282710537312776
|
| 1183 |
+
},
|
| 1184 |
+
"eval_runtime": 19.3864,
|
| 1185 |
+
"eval_samples_per_second": 85.575,
|
| 1186 |
+
"eval_steps_per_second": 10.729,
|
| 1187 |
+
"step": 10790
|
| 1188 |
+
}
|
| 1189 |
+
],
|
| 1190 |
+
"logging_steps": 100,
|
| 1191 |
+
"max_steps": 12450,
|
| 1192 |
+
"num_input_tokens_seen": 0,
|
| 1193 |
+
"num_train_epochs": 15,
|
| 1194 |
+
"save_steps": 500,
|
| 1195 |
+
"stateful_callbacks": {
|
| 1196 |
+
"TrainerControl": {
|
| 1197 |
+
"args": {
|
| 1198 |
+
"should_epoch_stop": false,
|
| 1199 |
+
"should_evaluate": false,
|
| 1200 |
+
"should_log": false,
|
| 1201 |
+
"should_save": true,
|
| 1202 |
+
"should_training_stop": false
|
| 1203 |
+
},
|
| 1204 |
+
"attributes": {}
|
| 1205 |
+
}
|
| 1206 |
+
},
|
| 1207 |
+
"total_flos": 7890557846663520.0,
|
| 1208 |
+
"train_batch_size": 8,
|
| 1209 |
+
"trial_name": null,
|
| 1210 |
+
"trial_params": null
|
| 1211 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e281bffd89136ee0dd827c8128960558474cc42ab286246ed0831eb92fb72fdd
|
| 3 |
+
size 5368
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|