djsull/aha_class

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: klue/roberta-small
+base_model: monologg/kobigbird-bert-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 [klue/roberta-small](https://huggingface.co/klue/roberta-small) on an unknown dataset.
+This model is a fine-tuned version of [monologg/kobigbird-bert-base](https://huggingface.co/monologg/kobigbird-bert-base) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0762
-- F1: 0.9705
-- Roc Auc: 0.9766
-- Accuracy: 0.9565
+- Loss: 0.0761
+- F1: 0.9748
+- Roc Auc: 0.9808
+- Accuracy: 0.9652
 
 ## Model description
 
@@ -51,20 +51,19 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss | F1     | Roc Auc | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:------:|:-------:|:--------:|
-| No log        | 1.0   | 42   | 0.1926          | 0.9218 | 0.9440  | 0.8609   |
-| No log        | 2.0   | 84   | 0.1140          | 0.9580 | 0.9679  | 0.9304   |
-| No log        | 3.0   | 126  | 0.1016          | 0.9536 | 0.9637  | 0.9217   |
-| No log        | 4.0   | 168  | 0.0800          | 0.9705 | 0.9766  | 0.9478   |
-| No log        | 5.0   | 210  | 0.0800          | 0.9661 | 0.9724  | 0.9478   |
-| No log        | 6.0   | 252  | 0.0817          | 0.9620 | 0.9701  | 0.9391   |
-| No log        | 7.0   | 294  | 0.0788          | 0.9661 | 0.9724  | 0.9478   |
-| No log        | 8.0   | 336  | 0.0803          | 0.9705 | 0.9766  | 0.9565   |
-| No log        | 9.0   | 378  | 0.0762          | 0.9705 | 0.9766  | 0.9565   |
-| No log        | 10.0  | 420  | 0.0837          | 0.9664 | 0.9743  | 0.9478   |
-| No log        | 11.0  | 462  | 0.0800          | 0.9664 | 0.9743  | 0.9478   |
-| 0.0807        | 12.0  | 504  | 0.0815          | 0.9664 | 0.9743  | 0.9478   |
-| 0.0807        | 13.0  | 546  | 0.0808          | 0.9664 | 0.9743  | 0.9478   |
-| 0.0807        | 14.0  | 588  | 0.0816          | 0.9664 | 0.9743  | 0.9478   |
+| 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   |
 
 
 ### Framework versions

config.json

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

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:061602a125454932c4542568e14cafa14e39b66c28f184f4c4ea7ee66aac0e8b
-size 272385276
+oid sha256:750d1e0f60305a59a21d40058f15835f7fbdfbaace1a1a013f425e0c30d8932f
+size 457410764

runs/Sep09_07-51-21_129-146-79-76/events.out.tfevents.1725868281.129-146-79-76.2385348.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:251f27747d3a4bcc7fb1a888e74f61dea9a6930d44dcb5a08963b4dae1d0553c
+size 11299

runs/Sep09_07-51-21_129-146-79-76/events.out.tfevents.1725868902.129-146-79-76.2385348.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43a9de58b35f765e97647b547b2cb6047c9df95149f07a582d74531c691241c2
+size 508

special_tokens_map.json

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

tokenizer_config.json

@@ -1,7 +1,7 @@
 {
   "added_tokens_decoder": {
     "0": {
-      "content": "[CLS]",
+      "content": "[PAD]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -9,7 +9,7 @@
       "special": true
     },
     "1": {
-      "content": "[PAD]",
+      "content": "[UNK]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -17,7 +17,7 @@
       "special": true
     },
     "2": {
-      "content": "[SEP]",
+      "content": "[CLS]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -25,7 +25,7 @@
       "special": true
     },
     "3": {
-      "content": "[UNK]",
+      "content": "[SEP]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -39,16 +39,32 @@
       "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": "[CLS]",
+  "bos_token": "<s>",
   "clean_up_tokenization_spaces": true,
   "cls_token": "[CLS]",
   "do_basic_tokenize": true,
   "do_lower_case": false,
-  "eos_token": "[SEP]",
+  "eos_token": "</s>",
   "mask_token": "[MASK]",
-  "model_max_length": 512,
+  "model_max_length": 4096,
   "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 sha256:47683d0e0fb266c7731152f6cdce02035e4240b3858e6f4f96d6919566584e67
+oid sha256:a29829286049265abb7992ca4a9e022dfa82ff8331490ab6958a2530947c7610
 size 5112