(self, embed_size=128, window_size=5, workers=3, iter=5, **kwargs)
| 35 | num_walks=num_walks, walk_length=walk_length, workers=workers, verbose=1) |
| 36 | |
| 37 | def train(self, embed_size=128, window_size=5, workers=3, iter=5, **kwargs): |
| 38 | |
| 39 | kwargs["sentences"] = self.sentences |
| 40 | kwargs["min_count"] = kwargs.get("min_count", 0) |
| 41 | kwargs["vector_size"] = embed_size |
| 42 | kwargs["sg"] = 1 # skip gram |
| 43 | kwargs["hs"] = 1 # deepwalk use Hierarchical Softmax |
| 44 | kwargs["workers"] = workers |
| 45 | kwargs["window"] = window_size |
| 46 | kwargs["epochs"] = iter |
| 47 | |
| 48 | print("Learning embedding vectors...") |
| 49 | model = Word2Vec(**kwargs) |
| 50 | print("Learning embedding vectors done!") |
| 51 | |
| 52 | self.w2v_model = model |
| 53 | return model |
| 54 | |
| 55 | def get_embeddings(self, ): |
| 56 | if self.w2v_model is None: |
no outgoing calls