commit files to HF hub

added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "[MASK]": 128000
+}

config.json

@@ -0,0 +1,54 @@
+{
+  "_name_or_path": "MoritzLaurer/DeBERTa-v3-base-mnli-fever-anli",
+  "architectures": [
+    "DebertaV2ForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "custom_pipelines": {
+    "entailment-classification": {
+      "impl": "pipeline.MyPipeline",
+      "pt": [
+        "AutoModelForSequenceClassification"
+      ],
+      "tf": []
+    }
+  },
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "entailment",
+    "1": "neutral",
+    "2": "contradiction"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "contradiction": 2,
+    "entailment": 0,
+    "neutral": 1
+  },
+  "layer_norm_eps": 1e-07,
+  "max_position_embeddings": 512,
+  "max_relative_positions": -1,
+  "model_type": "deberta-v2",
+  "norm_rel_ebd": "layer_norm",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "pooler_dropout": 0,
+  "pooler_hidden_act": "gelu",
+  "pooler_hidden_size": 768,
+  "pos_att_type": [
+    "p2c",
+    "c2p"
+  ],
+  "position_biased_input": false,
+  "position_buckets": 256,
+  "relative_attention": true,
+  "share_att_key": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.30.2",
+  "type_vocab_size": 0,
+  "vocab_size": 128100
+}

pipeline.py

@@ -0,0 +1,88 @@
+from huggingface_hub import Repository
+from typing import List, Union
+from transformers import pipeline
+from transformers.pipelines import PIPELINE_REGISTRY
+from transformers import AutoModelForSequenceClassification, Pipeline
+import torch
+
+
+# from loguru import logger
+
+
+class MyPipeline(Pipeline):
+    def _sanitize_parameters(self, **kwargs):
+        preprocess_kwargs = {}
+        if "hypothesis" in kwargs:
+            preprocess_kwargs["hypothesis"] = kwargs["hypothesis"]
+        return preprocess_kwargs, {}, {}
+    
+    def __call__(
+        self,
+        sequences: Union[str, List[str]],
+        *args,
+        **kwargs,
+    ):
+        if len(args) == 0:
+            pass
+        elif len(args) == 1 and "hypothesis" not in kwargs:
+            kwargs["hypothesis"] = args[0]
+        else:
+            raise ValueError(f"Unable to understand extra arguments {args}")
+
+        return super().__call__(sequences, **kwargs)
+
+
+    def preprocess(self, premise, hypothesis=None):
+        encode_inputs = self.tokenizer(
+                            premise, 
+                            hypothesis,
+                            #  max_length=self.toke,
+                            #  return_token_type_ids=True, 
+                            truncation=True,
+                            return_tensors="pt"
+                        )
+        
+        return {"input_ids": encode_inputs['input_ids']}
+
+    def _forward(self, input_ids):
+        outputs = self.model(input_ids['input_ids'])
+        return outputs
+
+    def postprocess(self, model_outputs):
+        prediction = torch.softmax(model_outputs["logits"][0], -1).tolist()
+        print(prediction)
+        label_names = ["entailment", "neutral", "contradiction"]
+        prediction = {name: round(float(pred) * 100, 1)
+              for pred, name in zip(prediction, label_names)}
+        return prediction
+
+
+# PIPELINE_REGISTRY.register_pipeline(
+#     "test",
+#     pipeline_class=MyPipeline,
+#     pt_model=AutoModelForSequenceClassification,
+#     # default={"pt": ("MoritzLaurer/DeBERTa-v3-base-mnli-fever-anli", "retina")},
+#     # type="text", 
+# )
+
+
+# classifier = pipeline("test", 
+#                       model="MoritzLaurer/DeBERTa-v3-base-mnli-fever-anli",
+#                     #   tokenizer="MoritzLaurer/DeBERTa-v3-base-mnli-fever-anli"
+#             )
+
+# output = classifier(
+#     "Angela Merkel is a politician in Germany and leader of the CDU",
+#     hypothesis="this is a test"
+# )
+# # logger.info(output)
+
+
+# # repo = Repository("entailment-classifier",
+# #                   clone_from="Tverous/entailment-classifier")
+# classifier.save_pretrained("entailment-classifier")
+# # repo.push_to_hub()
+
+
+# logger.info("Finished")
+

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be8178bff2c2a45c2c893a6584cd2b09ce98493f762c591175df839b9b38e6a4
+size 737769017

special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "bos_token": "[CLS]",
+  "cls_token": "[CLS]",
+  "eos_token": "[SEP]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

spm.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
+size 2464616

tokenizer_config.json

@@ -0,0 +1,16 @@
+{
+  "bos_token": "[CLS]",
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_lower_case": false,
+  "eos_token": "[SEP]",
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "sp_model_kwargs": {},
+  "split_by_punct": false,
+  "tokenizer_class": "DebertaV2Tokenizer",
+  "unk_token": "[UNK]",
+  "vocab_type": "spm"
+}