Upload model

hf_modeling_grounding.py

@@ -7,6 +7,7 @@ import torch.nn as nn
 import torch.nn.functional as F
 from torchaudio import transforms
 from transformers import PreTrainedModel, PretrainedConfig
+from transformers.utils.hub import cached_file
 
 
 def sum_with_lens(features, lens):
@@ -256,37 +257,6 @@ class DotProduct(nn.Module):
         return score
 
 
-class Vocabulary(object):
-    """Simple vocabulary wrapper."""
-
-    def __init__(self):
-        self.word2idx = {}
-        self.idx2word = {}
-        self.idx = 0
-
-    def add_word(self, word):
-        if not word in self.word2idx:
-            self.word2idx[word] = self.idx
-            self.idx2word[self.idx] = word
-            self.idx += 1
-
-    def __call__(self, word):
-        if not word in self.word2idx:
-            return self.word2idx["<unk>"]
-        return self.word2idx[word]
-
-    def __len__(self):
-        return len(self.word2idx)
-
-    def state_dict(self):
-        return self.word2idx
-
-    def load_state_dict(self, state_dict):
-        self.word2idx = state_dict
-        self.idx2word = {idx: word for word, idx in self.word2idx.items()}
-        self.idx = len(self.word2idx)
-
-
 class BiEncoder(nn.Module):
 
     def __init__(self,
@@ -425,6 +395,6 @@ class Cnn8RnnW2vMeanGroundingModel(PreTrainedModel):
                         **kwargs):
         model = super().from_pretrained(pretrained_model_name_or_path,
                                         *model_args, **kwargs)
-        model.vocab_mapping = json.load(
-            open(os.path.join(pretrained_model_name_or_path, "vocab.json")))
+        vocab_path = cached_file(pretrained_model_name_or_path, "vocab.json")
+        model.vocab_mapping = json.load(open(vocab_path))
         return model