first model

README.md

@@ -0,0 +1,10 @@
+## TextAttack Model Card
+
+This `cmarkea/distilcamembert-base` model was fine-tuned using TextAttackand the `allocine` dataset loaded using the `datasets` library. The model was fine-tuned
+for 1 epochs with a batch size of 16,
+    a maximum sequence length of 512, and an initial learning rate of 5e-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 0.9692, as measured by the
+eval set accuracy, found after 1 epoch.
+
+For more information, check out [TextAttack on Github](https://github.com/QData/TextAttack).

config.json

@@ -0,0 +1,28 @@
+{
+  "_name_or_path": "cmarkea/distilcamembert-base",
+  "architectures": [
+    "CamembertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 2,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "camembert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 6,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float32",
+  "transformers_version": "4.26.1",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 32005
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:556004d4af12dfeee38a8f4aadc5ab961bcf74b235a00cc1f7fe08cf09d2fd15
+size 272425205

sentencepiece.bpe.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:988bc5a00281c6d210a5d34bd143d0363741a432fefe741bf71e61b1869d4314
+size 810912

special_tokens_map.json

@@ -0,0 +1,19 @@
+{
+  "additional_special_tokens": [
+    "<s>NOTUSED",
+    "</s>NOTUSED"
+  ],
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": {
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "unk_token": "<unk>"
+}

tokenizer_config.json

@@ -0,0 +1,25 @@
+{
+  "additional_special_tokens": [
+    "<s>NOTUSED",
+    "</s>NOTUSED"
+  ],
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": {
+    "__type": "AddedToken",
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "model_max_length": 512,
+  "name_or_path": "cmarkea/distilcamembert-base",
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "sp_model_kwargs": {},
+  "special_tokens_map_file": null,
+  "tokenizer_class": "CamembertTokenizer",
+  "unk_token": "<unk>"
+}

train_log.txt

@@ -0,0 +1,17 @@
+Writing logs to ./outputs/2023-02-13-00-40-42-038240/train_log.txt.
+Wrote original training args to ./outputs/2023-02-13-00-40-42-038240/training_args.json.
+***** Running training *****
+  Num examples = 160000
+  Num epochs = 1
+  Num clean epochs = 1
+  Instantaneous batch size per device = 16
+  Total train batch size (w. parallel, distributed & accumulation) = 16
+  Gradient accumulation steps = 1
+  Total optimization steps = 10000
+==========================================================
+Epoch 1
+Running clean epoch 1/1
+Train accuracy: 95.02%
+Eval accuracy: 96.92%
+Best score found. Saved model to ./outputs/2023-02-13-00-40-42-038240/best_model/
+Wrote README to ./outputs/2023-02-13-00-40-42-038240/README.md.

training_args.json

@@ -0,0 +1,38 @@
+{
+    "model_name_or_path": "cmarkea/distilcamembert-base",
+    "attack": null,
+    "dataset": "allocine",
+    "task_type": "classification",
+    "model_max_length": null,
+    "model_num_labels": null,
+    "dataset_train_split": "train",
+    "dataset_eval_split": "validation",
+    "filter_train_by_labels": null,
+    "filter_eval_by_labels": null,
+    "num_epochs": 1,
+    "num_clean_epochs": 1,
+    "attack_epoch_interval": 1,
+    "early_stopping_epochs": null,
+    "learning_rate": 5e-05,
+    "num_warmup_steps": 500,
+    "weight_decay": 0.01,
+    "per_device_train_batch_size": 16,
+    "per_device_eval_batch_size": 32,
+    "gradient_accumulation_steps": 1,
+    "random_seed": 786,
+    "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/2023-02-13-00-40-42-038240",
+    "checkpoint_interval_steps": null,
+    "checkpoint_interval_epochs": null,
+    "save_last": true,
+    "log_to_tb": true,
+    "tb_log_dir": null,
+    "log_to_wandb": false,
+    "wandb_project": "textattack",
+    "logging_interval_step": 1
+}