Training in progress, step 100

config.json

@@ -1,46 +1,54 @@
 {
-  "_name_or_path": "kykim/bert-kor-base",
+  "_name_or_path": "kykim/funnel-kor-base",
+  "activation_dropout": 0.0,
   "architectures": [
-    "BertForSequenceClassification"
+    "FunnelForSequenceClassification"
   ],
-  "attention_probs_dropout_prob": 0.1,
-  "classifier_dropout": null,
-  "directionality": "bidi",
+  "attention_dropout": 0.1,
+  "attention_type": "relative_shift",
+  "block_repeats": [
+    1,
+    1,
+    1
+  ],
+  "block_sizes": [
+    6,
+    6,
+    6
+  ],
+  "d_head": 64,
+  "d_inner": 3072,
+  "d_model": 768,
   "embedding_size": 768,
-  "gradient_checkpointing": false,
-  "hidden_act": "gelu",
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 768,
+  "hidden_act": "gelu_new",
+  "hidden_dropout": 0.1,
   "id2label": {
     "0": "item",
     "1": "non_item",
     "2": "garbage",
     "3": "review"
   },
-  "initializer_range": 0.02,
-  "intermediate_size": 3072,
+  "initializer_range": 0.1,
+  "initializer_std": 0.02,
   "label2id": {
     "garbage": 2,
     "item": 0,
     "non_item": 1,
     "review": 3
   },
-  "layer_norm_eps": 1e-12,
+  "layer_norm_eps": 1e-09,
   "max_position_embeddings": 512,
-  "model_type": "bert",
-  "num_attention_heads": 12,
-  "num_hidden_layers": 12,
-  "pad_token_id": 0,
-  "pooler_fc_size": 768,
-  "pooler_num_attention_heads": 12,
-  "pooler_num_fc_layers": 3,
-  "pooler_size_per_head": 128,
-  "pooler_type": "first_token_transform",
-  "position_embedding_type": "absolute",
+  "model_type": "funnel",
+  "n_head": 12,
+  "num_decoder_layers": 2,
+  "pool_q_only": true,
+  "pooling_type": "mean",
   "problem_type": "single_label_classification",
+  "rel_attn_type": "factorized",
+  "separate_cls": true,
   "torch_dtype": "float32",
   "transformers_version": "4.32.0",
-  "type_vocab_size": 2,
-  "use_cache": true,
+  "truncate_seq": true,
+  "type_vocab_size": 3,
   "vocab_size": 42000
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bb3494620bcb22a346ddd8998dc474e2040ba8aa074aca54907ab52367d7efdf
-size 473270510
+oid sha256:87bfce63c841cd79bb028372104a643497711b535441e65ff80bea01d6e4ac56
+size 684546438

special_tokens_map.json

@@ -1,7 +1,9 @@
 {
-  "cls_token": "[CLS]",
-  "mask_token": "[MASK]",
-  "pad_token": "[PAD]",
-  "sep_token": "[SEP]",
-  "unk_token": "[UNK]"
+  "bos_token": "<s>",
+  "cls_token": "<cls>",
+  "eos_token": "</s>",
+  "mask_token": "<mask>",
+  "pad_token": "<pad>",
+  "sep_token": "<sep>",
+  "unk_token": "<unk>"
 }

tokenizer.json

@@ -10,7 +10,7 @@
   "added_tokens": [
     {
       "id": 0,
-      "content": "[PAD]",
+      "content": "<pad>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -19,7 +19,7 @@
     },
     {
       "id": 1,
-      "content": "[UNK]",
+      "content": "<unk>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -28,7 +28,7 @@
     },
     {
       "id": 2,
-      "content": "[CLS]",
+      "content": "<cls>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -37,7 +37,7 @@
     },
     {
       "id": 3,
-      "content": "[SEP]",
+      "content": "<sep>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -46,7 +46,25 @@
     },
     {
       "id": 4,
-      "content": "[MASK]",
+      "content": "<mask>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 5,
+      "content": "<s>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 6,
+      "content": "</s>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -69,8 +87,8 @@
     "single": [
       {
         "SpecialToken": {
-          "id": "[CLS]",
-          "type_id": 0
+          "id": "<cls>",
+          "type_id": 2
         }
       },
       {
@@ -81,7 +99,7 @@
       },
       {
         "SpecialToken": {
-          "id": "[SEP]",
+          "id": "<sep>",
           "type_id": 0
         }
       }
@@ -89,8 +107,8 @@
     "pair": [
       {
         "SpecialToken": {
-          "id": "[CLS]",
-          "type_id": 0
+          "id": "<cls>",
+          "type_id": 2
         }
       },
       {
@@ -101,7 +119,7 @@
       },
       {
         "SpecialToken": {
-          "id": "[SEP]",
+          "id": "<sep>",
           "type_id": 0
         }
       },
@@ -113,28 +131,28 @@
       },
       {
         "SpecialToken": {
-          "id": "[SEP]",
+          "id": "<sep>",
           "type_id": 1
         }
       }
     ],
     "special_tokens": {
-      "[CLS]": {
-        "id": "[CLS]",
+      "<cls>": {
+        "id": "<cls>",
         "ids": [
           2
         ],
         "tokens": [
-          "[CLS]"
+          "<cls>"
         ]
       },
-      "[SEP]": {
-        "id": "[SEP]",
+      "<sep>": {
+        "id": "<sep>",
         "ids": [
           3
         ],
         "tokens": [
-          "[SEP]"
+          "<sep>"
         ]
       }
     }
@@ -146,19 +164,19 @@
   },
   "model": {
     "type": "WordPiece",
-    "unk_token": "[UNK]",
+    "unk_token": "<unk>",
     "continuing_subword_prefix": "##",
     "max_input_chars_per_word": 100,
     "vocab": {
-      "[PAD]": 0,
-      "[UNK]": 1,
-      "[CLS]": 2,
-      "[SEP]": 3,
-      "[MASK]": 4,
-      "[unused0]": 5,
-      "[unused1]": 6,
-      "[unused2]": 7,
-      "[unused3]": 8,
+      "<pad>": 0,
+      "<unk>": 1,
+      "<cls>": 2,
+      "<sep>": 3,
+      "<mask>": 4,
+      "<s>": 5,
+      "</s>": 6,
+      "<eod>": 7,
+      "<eop>": 8,
       "[unused4]": 9,
       "[unused5]": 10,
       "[unused6]": 11,

tokenizer_config.json

@@ -1,15 +1,19 @@
 {
+  "bos_token": "<s>",
+  "clean_text": true,
   "clean_up_tokenization_spaces": true,
-  "cls_token": "[CLS]",
+  "cls_token": "<cls>",
   "do_basic_tokenize": true,
   "do_lower_case": true,
-  "mask_token": "[MASK]",
+  "eos_token": "</s>",
+  "mask_token": "<mask>",
   "model_max_length": 512,
   "never_split": null,
-  "pad_token": "[PAD]",
-  "sep_token": "[SEP]",
+  "pad_token": "<pad>",
+  "sep_token": "<sep>",
   "strip_accents": false,
   "tokenize_chinese_chars": true,
-  "tokenizer_class": "BertTokenizer",
-  "unk_token": "[UNK]"
+  "tokenizer_class": "FunnelTokenizer",
+  "unk_token": "<unk>",
+  "wordpieces_prefix": "##"
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a1f7d315177df05497b5918048ea9caed7d58672564736c0b364143181096f72
+oid sha256:5d9a5a04723c7b140f278b33665efcfdaa8934f2abd07e0d4349157e35f33bb1
 size 4472

vocab.txt

@@ -1,12 +1,12 @@
-[PAD]
-[UNK]
-[CLS]
-[SEP]
-[MASK]
-[unused0]
-[unused1]
-[unused2]
-[unused3]
+<pad>
+<unk>
+<cls>
+<sep>
+<mask>
+<s>
+</s>
+<eod>
+<eop>
 [unused4]
 [unused5]
 [unused6]