Upload folder using huggingface_hub
Browse files- best_metadata.json +10 -0
- best_model.safetensors +3 -0
- inference_config.json +11 -0
- label_map.json +89 -0
- model_config.json +38 -0
- tokenizer/tokenizer.json +0 -0
- tokenizer/tokenizer_config.json +23 -0
best_metadata.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"run_dir": "/Users/baseconnect/hawks_utils/reducer-data-agent/pipeline-output/training_runs/20260306_001508",
|
| 3 |
+
"promoted_at": "2026-03-06T08:03:00",
|
| 4 |
+
"primary_metric": "macro_f1",
|
| 5 |
+
"accuracy": 0.8684807256235828,
|
| 6 |
+
"macro_f1": 0.833427366079523,
|
| 7 |
+
"weighted_f1": 0.8698534999896407,
|
| 8 |
+
"best_epoch": 15,
|
| 9 |
+
"test_file": "/Users/baseconnect/hawks_utils/reducer-data-agent/test-data/Hawks_ver6.0_alternative.csv"
|
| 10 |
+
}
|
best_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:256d79ef28da90dc5f300dc5b569c3e1d855b21c4b67e17e8ff5c94c868a1138
|
| 3 |
+
size 639688828
|
inference_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"max_length": 512,
|
| 3 |
+
"title_empty_token": "[TITLE_EMPTY]",
|
| 4 |
+
"title_repeat": true,
|
| 5 |
+
"url_token": "[URL]",
|
| 6 |
+
"pooling": "cls",
|
| 7 |
+
"pick_map": {
|
| 8 |
+
"Decline": 0,
|
| 9 |
+
"Pick": 1
|
| 10 |
+
}
|
| 11 |
+
}
|
label_map.json
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"CSR活動": 0,
|
| 3 |
+
"DX": 1,
|
| 4 |
+
"M&A": 2,
|
| 5 |
+
"イベント": 3,
|
| 6 |
+
"サイバー攻撃": 4,
|
| 7 |
+
"システム障害": 5,
|
| 8 |
+
"ダイバーシティ": 6,
|
| 9 |
+
"ハラスメント問題": 7,
|
| 10 |
+
"リニューアルオープン": 8,
|
| 11 |
+
"不法就労": 9,
|
| 12 |
+
"不適切広告": 10,
|
| 13 |
+
"不適切発言": 11,
|
| 14 |
+
"事故": 12,
|
| 15 |
+
"事業効率化": 13,
|
| 16 |
+
"事業縮小": 14,
|
| 17 |
+
"人事異動": 15,
|
| 18 |
+
"人材育成": 16,
|
| 19 |
+
"代表者死亡": 17,
|
| 20 |
+
"企業間トラブル": 18,
|
| 21 |
+
"企業間取引解消": 19,
|
| 22 |
+
"価格変更": 20,
|
| 23 |
+
"債務超過": 21,
|
| 24 |
+
"債権回収不能": 22,
|
| 25 |
+
"債権譲渡": 23,
|
| 26 |
+
"公害": 24,
|
| 27 |
+
"再建型倒産": 25,
|
| 28 |
+
"労働争議": 26,
|
| 29 |
+
"労働災害": 27,
|
| 30 |
+
"労働者の自殺・失踪": 28,
|
| 31 |
+
"反社会的勢力関与": 29,
|
| 32 |
+
"受賞": 30,
|
| 33 |
+
"品質管理不正": 31,
|
| 34 |
+
"商品表示の不備": 32,
|
| 35 |
+
"広報活動": 33,
|
| 36 |
+
"役員変更": 34,
|
| 37 |
+
"従業員の不祥事": 35,
|
| 38 |
+
"従業員数減少": 36,
|
| 39 |
+
"情報漏洩": 37,
|
| 40 |
+
"批判的投稿": 38,
|
| 41 |
+
"提携": 39,
|
| 42 |
+
"新商品の展開": 40,
|
| 43 |
+
"新工場竣工": 41,
|
| 44 |
+
"新店舗開店": 42,
|
| 45 |
+
"新施設開設": 43,
|
| 46 |
+
"新規事業参入": 44,
|
| 47 |
+
"新規採用": 45,
|
| 48 |
+
"新設備導入": 46,
|
| 49 |
+
"本社住所移転": 47,
|
| 50 |
+
"株主変更": 48,
|
| 51 |
+
"株主総会開催": 49,
|
| 52 |
+
"株式保有率変更": 50,
|
| 53 |
+
"株式公開": 51,
|
| 54 |
+
"業績悪化": 52,
|
| 55 |
+
"業績改善": 53,
|
| 56 |
+
"横領": 54,
|
| 57 |
+
"海外進出": 55,
|
| 58 |
+
"清算型倒産": 56,
|
| 59 |
+
"特許取得": 57,
|
| 60 |
+
"生産停止": 58,
|
| 61 |
+
"生産再開": 59,
|
| 62 |
+
"生産拡大": 60,
|
| 63 |
+
"異物混入": 61,
|
| 64 |
+
"知的財産権の侵害": 62,
|
| 65 |
+
"研究開発": 63,
|
| 66 |
+
"社名変更": 64,
|
| 67 |
+
"税務違反": 65,
|
| 68 |
+
"粉飾決算": 66,
|
| 69 |
+
"組織体制変更": 67,
|
| 70 |
+
"経営計画": 68,
|
| 71 |
+
"職場環境改善": 69,
|
| 72 |
+
"行政処分": 70,
|
| 73 |
+
"補助金獲得": 71,
|
| 74 |
+
"設備投資": 72,
|
| 75 |
+
"訴訟": 73,
|
| 76 |
+
"詐欺": 74,
|
| 77 |
+
"認定取得": 75,
|
| 78 |
+
"談合": 76,
|
| 79 |
+
"販路拡大": 77,
|
| 80 |
+
"賃上げ": 78,
|
| 81 |
+
"賃金未払い": 79,
|
| 82 |
+
"資本金の変更": 80,
|
| 83 |
+
"資金調達": 81,
|
| 84 |
+
"贈収賄": 82,
|
| 85 |
+
"通信障害": 83,
|
| 86 |
+
"逮捕": 84,
|
| 87 |
+
"違法残業": 85,
|
| 88 |
+
"顧客トラブル": 86
|
| 89 |
+
}
|
model_config.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"DebertaV2ForMaskedLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"bos_token_id": 96871,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
+
"eos_token_id": 96868,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 3072,
|
| 14 |
+
"layer_norm_eps": 1e-07,
|
| 15 |
+
"legacy": true,
|
| 16 |
+
"max_position_embeddings": 512,
|
| 17 |
+
"max_relative_positions": -1,
|
| 18 |
+
"model_type": "deberta-v2",
|
| 19 |
+
"norm_rel_ebd": "layer_norm",
|
| 20 |
+
"num_attention_heads": 12,
|
| 21 |
+
"num_hidden_layers": 12,
|
| 22 |
+
"pad_token_id": 96869,
|
| 23 |
+
"pooler_dropout": 0,
|
| 24 |
+
"pooler_hidden_act": "gelu",
|
| 25 |
+
"pooler_hidden_size": 768,
|
| 26 |
+
"pos_att_type": [
|
| 27 |
+
"p2c",
|
| 28 |
+
"c2p"
|
| 29 |
+
],
|
| 30 |
+
"position_biased_input": false,
|
| 31 |
+
"position_buckets": 256,
|
| 32 |
+
"relative_attention": true,
|
| 33 |
+
"share_att_key": true,
|
| 34 |
+
"tie_word_embeddings": true,
|
| 35 |
+
"transformers_version": "5.1.0",
|
| 36 |
+
"type_vocab_size": 0,
|
| 37 |
+
"vocab_size": 96873
|
| 38 |
+
}
|
tokenizer/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer/tokenizer_config.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": true,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "[CLS]",
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"cls_token": "[CLS]",
|
| 7 |
+
"do_lower_case": false,
|
| 8 |
+
"eos_token": "[SEP]",
|
| 9 |
+
"extra_special_tokens": [
|
| 10 |
+
"[TITLE_EMPTY]"
|
| 11 |
+
],
|
| 12 |
+
"is_local": true,
|
| 13 |
+
"keep_accents": true,
|
| 14 |
+
"mask_token": "[MASK]",
|
| 15 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 16 |
+
"pad_token": "[PAD]",
|
| 17 |
+
"sep_token": "[SEP]",
|
| 18 |
+
"sp_model_kwargs": {},
|
| 19 |
+
"split_by_punct": false,
|
| 20 |
+
"tokenizer_class": "DebertaV2Tokenizer",
|
| 21 |
+
"unk_id": 1,
|
| 22 |
+
"unk_token": "[UNK]"
|
| 23 |
+
}
|