model upload

README.md

@@ -0,0 +1,10 @@
+## TextAttack Model Card
+
+            This `distilbert` model was fine-tuned using TextAttack. The model was fine-tuned
+            for 3 epochs with a batch size of 8,
+             a maximum sequence length of 512, and an initial learning rate of 3e-05.
+            Since this was a classification task, the model was trained with a cross-entropy loss function.
+            The best score the model achieved on this task was 0.9004, as measured by the
+            eval set accuracy, found after 3 epochs.
+
+            For more information, check out [TextAttack on Github](https://github.com/QData/TextAttack).

config.json

@@ -0,0 +1,43 @@
+{
+  "_name_or_path": "distilbert/distilbert-base-multilingual-cased",
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "id2label": {
+    "0": "Mainland China Politics",
+    "1": "HongKong Macau Politics",
+    "2": "International News",
+    "3": "Financial News",
+    "4": "Culture",
+    "5": "Entertainment",
+    "6": "Sports"
+  },
+  "initializer_range": 0.02,
+  "label2id": {
+    "Culture": 4,
+    "Entertainment": 5,
+    "Financial News": 3,
+    "HongKong Macau Politics": 1,
+    "International News": 2,
+    "Mainland China Politics": 0,
+    "Sports": 6
+  },
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "output_past": true,
+  "pad_token_id": 0,
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.38.1",
+  "vocab_size": 119547
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:259c5aa25f7404d1f6c55ca81386d7406d70fc4c669bcc56c570442457f40ceb
+size 541332756

special_tokens_map.json

@@ -0,0 +1,7 @@
+{
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

tokenizer_config.json

@@ -0,0 +1,55 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_lower_case": false,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "DistilBertTokenizer",
+  "unk_token": "[UNK]"
+}

train_log.txt

@@ -0,0 +1,29 @@
+Writing logs to ./outputs/2024-03-01-10-24-35-896622/train_log.txt.
+Wrote original training args to ./outputs/2024-03-01-10-24-35-896622/training_args.json.
+***** Running training *****
+  Num examples = 50000
+  Num epochs = 3
+  Num clean epochs = 3
+  Instantaneous batch size per device = 8
+  Total train batch size (w. parallel, distributed & accumulation) = 8
+  Gradient accumulation steps = 1
+  Total optimization steps = 18750
+==========================================================
+Epoch 1
+Running clean epoch 1/3
+Train accuracy: 83.54%
+Eval accuracy: 88.30%
+Best score found. Saved model to ./outputs/2024-03-01-10-24-35-896622/best_model/
+==========================================================
+Epoch 2
+Running clean epoch 2/3
+Train accuracy: 91.31%
+Eval accuracy: 89.43%
+Best score found. Saved model to ./outputs/2024-03-01-10-24-35-896622/best_model/
+==========================================================
+Epoch 3
+Running clean epoch 3/3
+Train accuracy: 95.09%
+Eval accuracy: 90.04%
+Best score found. Saved model to ./outputs/2024-03-01-10-24-35-896622/best_model/
+Wrote README to ./outputs/2024-03-01-10-24-35-896622/README.md.

training_args.json

@@ -0,0 +1 @@
+{"num_epochs": 3, "num_clean_epochs": 1, "attack_epoch_interval": 1, "early_stopping_epochs": null, "learning_rate": 3e-05, "num_warmup_steps": 500, "weight_decay": 0.01, "per_device_train_batch_size": 8, "per_device_eval_batch_size": 32, "gradient_accumulation_steps": 1, "random_seed": 718, "parallel": false, "load_best_model_at_end": false, "alpha": 1.0, "num_train_adv_examples": -1, "query_budget_train": null, "attack_num_workers_per_device": 1, "output_dir": "./outputs/2024-03-01-10-24-35-896622", "checkpoint_interval_steps": null, "checkpoint_interval_epochs": null, "save_last": true, "log_to_tb": false, "tb_log_dir": null, "log_to_wandb": false, "wandb_project": "textattack", "logging_interval_step": 1}