Upload v2 of URL classifier model (hybrid BERT + tabular)

Training Metrics:
- Eval Loss: 0.07472482323646545
- Eval F1 Macro: 0.9319480242737234
- Eval Accuracy: 0.9817232375979112

config.json

@@ -1,6 +1,6 @@
 {
   "architectures": [
-    "BertForSequenceClassification"
+    "BertModel"
   ],
   "attention_probs_dropout_prob": 0.1,
   "classifier_dropout": null,
@@ -16,7 +16,7 @@
   "model_type": "bert",
   "num_attention_heads": 12,
   "num_hidden_layers": 12,
-  "pad_token_id": 0,
+  "pad_token_id": 1,
   "position_embedding_type": "absolute",
   "transformers_version": "4.57.0",
   "type_vocab_size": 2,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5d9a3e637d2b82a295a3ea5555e3a0040b5645d59093f47df6bdeac1f6486c68
-size 442499064
+oid sha256:91f225c4b06118f3c2848477bd64c4b76a2928ce8e330b42ba114faa706dcc0e
+size 442503856

special_tokens_map.json

@@ -1,4 +1,10 @@
 {
+  "additional_special_tokens": [
+    "[DOMAIN]",
+    "[PATH]",
+    "[IP]",
+    "[IPv6]"
+  ],
   "cls_token": {
     "content": "[CLS]",
     "lstrip": false,

tabular_config.json

@@ -0,0 +1,4 @@
+{
+  "num_labels": 2,
+  "num_tabular_features": 6
+}

tokenizer_config.json

@@ -1,7 +1,7 @@
 {
   "added_tokens_decoder": {
     "0": {
-      "content": "[PAD]",
+      "content": "[UNK]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -9,7 +9,7 @@
       "special": true
     },
     "1": {
-      "content": "[UNK]",
+      "content": "[PAD]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -39,28 +39,53 @@
       "rstrip": false,
       "single_word": false,
       "special": true
+    },
+    "5": {
+      "content": "[DOMAIN]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "6": {
+      "content": "[PATH]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "7": {
+      "content": "[IP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "8": {
+      "content": "[IPv6]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
     }
   },
+  "additional_special_tokens": [
+    "[DOMAIN]",
+    "[PATH]",
+    "[IP]",
+    "[IPv6]"
+  ],
   "clean_up_tokenization_spaces": true,
   "cls_token": "[CLS]",
-  "do_basic_tokenize": true,
-  "do_lower_case": true,
   "extra_special_tokens": {},
   "mask_token": "[MASK]",
-  "max_len": 64,
-  "max_length": 64,
-  "model_max_length": 64,
-  "never_split": null,
-  "pad_to_multiple_of": null,
+  "model_max_length": 1024,
   "pad_token": "[PAD]",
-  "pad_token_type_id": 0,
-  "padding_side": "right",
   "sep_token": "[SEP]",
-  "stride": 0,
-  "strip_accents": null,
-  "tokenize_chinese_chars": true,
-  "tokenizer_class": "BertTokenizer",
-  "truncation_side": "right",
-  "truncation_strategy": "longest_first",
+  "tokenizer_class": "PreTrainedTokenizerFast",
   "unk_token": "[UNK]"
 }

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4bf205e4a14c0b396ee497711eeaf558cb79f3225536ee3e5fea5d95bccf5321
+size 5777