djsull/aha_class

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: monologg/kobigbird-bert-base
+base_model: klue/roberta-base
 tags:
 - generated_from_trainer
 metrics:
@@ -15,12 +15,12 @@ should probably proofread and complete it, then remove this comment. -->
 
 # aha_class
 
-This model is a fine-tuned version of [monologg/kobigbird-bert-base](https://huggingface.co/monologg/kobigbird-bert-base) on an unknown dataset.
+This model is a fine-tuned version of [klue/roberta-base](https://huggingface.co/klue/roberta-base) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0761
-- F1: 0.9748
-- Roc Auc: 0.9808
-- Accuracy: 0.9652
+- Loss: 0.0885
+- F1: 0.9580
+- Roc Auc: 0.9679
+- Accuracy: 0.9391
 
 ## Model description
 
@@ -51,19 +51,17 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss | F1     | Roc Auc | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:------:|:-------:|:--------:|
-| No log        | 1.0   | 42   | 0.3253          | 0.8814 | 0.9082  | 0.8435   |
-| No log        | 2.0   | 84   | 0.1631          | 0.9661 | 0.9724  | 0.9478   |
-| No log        | 3.0   | 126  | 0.1213          | 0.9661 | 0.9724  | 0.9478   |
-| No log        | 4.0   | 168  | 0.0921          | 0.9705 | 0.9766  | 0.9565   |
-| No log        | 5.0   | 210  | 0.0869          | 0.9705 | 0.9766  | 0.9565   |
-| No log        | 6.0   | 252  | 0.0868          | 0.9748 | 0.9808  | 0.9652   |
-| No log        | 7.0   | 294  | 0.0825          | 0.9746 | 0.9788  | 0.9652   |
-| No log        | 8.0   | 336  | 0.0761          | 0.9748 | 0.9808  | 0.9652   |
-| No log        | 9.0   | 378  | 0.0820          | 0.9707 | 0.9785  | 0.9565   |
-| No log        | 10.0  | 420  | 0.0774          | 0.9748 | 0.9808  | 0.9652   |
-| No log        | 11.0  | 462  | 0.0772          | 0.9707 | 0.9785  | 0.9565   |
-| 0.1111        | 12.0  | 504  | 0.0767          | 0.9748 | 0.9808  | 0.9652   |
-| 0.1111        | 13.0  | 546  | 0.0785          | 0.9748 | 0.9808  | 0.9652   |
+| No log        | 1.0   | 42   | 0.1723          | 0.95   | 0.9635  | 0.9217   |
+| No log        | 2.0   | 84   | 0.1160          | 0.9576 | 0.9659  | 0.9391   |
+| No log        | 3.0   | 126  | 0.1064          | 0.9492 | 0.9595  | 0.9304   |
+| No log        | 4.0   | 168  | 0.0974          | 0.9540 | 0.9657  | 0.9304   |
+| No log        | 5.0   | 210  | 0.0968          | 0.9580 | 0.9679  | 0.9304   |
+| No log        | 6.0   | 252  | 0.0885          | 0.9580 | 0.9679  | 0.9391   |
+| No log        | 7.0   | 294  | 0.1005          | 0.9580 | 0.9679  | 0.9391   |
+| No log        | 8.0   | 336  | 0.0921          | 0.9664 | 0.9743  | 0.9478   |
+| No log        | 9.0   | 378  | 0.1055          | 0.9580 | 0.9679  | 0.9391   |
+| No log        | 10.0  | 420  | 0.0988          | 0.9664 | 0.9743  | 0.9478   |
+| No log        | 11.0  | 462  | 0.0993          | 0.9664 | 0.9743  | 0.9478   |
 
 
 ### Framework versions

config.json

@@ -1,16 +1,14 @@
 {
-  "_name_or_path": "monologg/kobigbird-bert-base",
+  "_name_or_path": "klue/roberta-base",
   "architectures": [
-    "BigBirdForSequenceClassification"
+    "RobertaForSequenceClassification"
   ],
   "attention_probs_dropout_prob": 0.1,
-  "attention_type": "block_sparse",
-  "block_size": 64,
-  "bos_token_id": 5,
+  "bos_token_id": 0,
   "classifier_dropout": null,
-  "eos_token_id": 6,
+  "eos_token_id": 2,
   "gradient_checkpointing": false,
-  "hidden_act": "gelu_new",
+  "hidden_act": "gelu",
   "hidden_dropout_prob": 0.1,
   "hidden_size": 768,
   "id2label": {
@@ -25,22 +23,18 @@
     "Service": 1,
     "Token": 2
   },
-  "layer_norm_eps": 1e-12,
-  "max_position_embeddings": 4096,
-  "model_type": "big_bird",
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "roberta",
   "num_attention_heads": 12,
   "num_hidden_layers": 12,
-  "num_random_blocks": 3,
-  "pad_token_id": 0,
+  "pad_token_id": 1,
   "position_embedding_type": "absolute",
   "problem_type": "multi_label_classification",
-  "rescale_embeddings": false,
-  "sep_token_id": 3,
   "tokenizer_class": "BertTokenizer",
   "torch_dtype": "float32",
   "transformers_version": "4.42.3",
-  "type_vocab_size": 2,
-  "use_bias": true,
+  "type_vocab_size": 1,
   "use_cache": true,
-  "vocab_size": 32500
+  "vocab_size": 32000
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:750d1e0f60305a59a21d40058f15835f7fbdfbaace1a1a013f425e0c30d8932f
-size 457410764
+oid sha256:7e7501a4ea7a7116c2056c188862a55419a7ea53bec1a59c52e3f7630326e82b
+size 442505820

runs/Sep10_07-10-38_129-146-79-76/events.out.tfevents.1725952239.129-146-79-76.2854019.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93779abc7d31a808705a709c1cb832667d9602accadafd0b6136e0ac65b7788c
+size 10049

runs/Sep10_07-10-38_129-146-79-76/events.out.tfevents.1725953200.129-146-79-76.2854019.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e16372ff507c3f6165fb24714420daf179a06bf97a4f14f812e0fe4ea2ceec8f
+size 508

special_tokens_map.json

@@ -1,6 +1,6 @@
 {
   "bos_token": {
-    "content": "<s>",
+    "content": "[CLS]",
     "lstrip": false,
     "normalized": false,
     "rstrip": false,
@@ -14,7 +14,7 @@
     "single_word": false
   },
   "eos_token": {
-    "content": "</s>",
+    "content": "[SEP]",
     "lstrip": false,
     "normalized": false,
     "rstrip": false,

tokenizer_config.json

@@ -1,7 +1,7 @@
 {
   "added_tokens_decoder": {
     "0": {
-      "content": "[PAD]",
+      "content": "[CLS]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -9,7 +9,7 @@
       "special": true
     },
     "1": {
-      "content": "[UNK]",
+      "content": "[PAD]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -17,7 +17,7 @@
       "special": true
     },
     "2": {
-      "content": "[CLS]",
+      "content": "[SEP]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -25,7 +25,7 @@
       "special": true
     },
     "3": {
-      "content": "[SEP]",
+      "content": "[UNK]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -39,32 +39,16 @@
       "rstrip": false,
       "single_word": false,
       "special": true
-    },
-    "5": {
-      "content": "<s>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "6": {
-      "content": "</s>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
     }
   },
-  "bos_token": "<s>",
+  "bos_token": "[CLS]",
   "clean_up_tokenization_spaces": true,
   "cls_token": "[CLS]",
   "do_basic_tokenize": true,
   "do_lower_case": false,
-  "eos_token": "</s>",
+  "eos_token": "[SEP]",
   "mask_token": "[MASK]",
-  "model_max_length": 4096,
+  "model_max_length": 512,
   "never_split": null,
   "pad_token": "[PAD]",
   "sep_token": "[SEP]",

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a29829286049265abb7992ca4a9e022dfa82ff8331490ab6958a2530947c7610
+oid sha256:c4c17129074b1635e7ea1818ff25d4eb9b99fe7e40b930fe5b6f259c28ed2a8d
 size 5112