Initial commit

config.json

@@ -0,0 +1,42 @@
+{
+  "_name_or_path": "sgugger/finetuned-bert-mrpc",
+  "architectures": [
+    "BertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "custom_pipelines": {
+    "pair-classification": {
+      "impl": "pair_classification.PairClassificationPipeline",
+      "pt": "AutoModelForSequenceClassification",
+      "tf": "TFAutoModelForSequenceClassification"
+    }
+  },
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "not equivalent",
+    "1": "equivalent"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "equivalent": 1,
+    "not equivalent": 0
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "torch_dtype": "float32",
+  "transformers_version": "4.21.0.dev0",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 28996
+}

pair_classification.py

@@ -0,0 +1,25 @@
+from transformers import Pipeline
+import torch
+
+class PairClassificationPipeline(Pipeline):
+    def _sanitize_parameters(self, **kwargs):
+        preprocess_kwargs = {}
+        if "text_pair" in kwargs:
+            preprocess_kwargs["text_pair"] = kwargs["text_pair"]
+        return preprocess_kwargs, {}, {}
+
+    def preprocess(self, text, text_pair=None):
+        return self.tokenizer(text, text_pair=text_pair, return_tensors="pt")
+
+    def _forward(self, model_inputs):
+        return self.model(**model_inputs)
+
+    def postprocess(self, model_outputs):
+        logits = model_outputs.logits
+        probabilities = torch.nn.functional.softmax(logits, dim=-1)
+        
+        best_class = probabilities.argmax().item()
+        label = self.model.config.id2label[best_class]
+        score = probabilities.squeeze()[best_class].item()
+        logits = logits.squeeze().tolist()
+        return {"label": label, "score": score, "logits": logits}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fb4ff71c41b993ad8531e00ca325ef3c7938156427dccae355771c026a75d4ad
+size 433315437

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 +1,14 @@
+{
+  "cls_token": "[CLS]",
+  "do_lower_case": false,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "name_or_path": "sgugger/finetuned-bert-mrpc",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "special_tokens_map_file": null,
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}