ayo/ipex-llm
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

LLM: modify transformersembeddings.embed() in langchain (#10051)

Browse source
This commit is contained in:
Zhicun 2024-02-05 10:42:10 +08:00 committed by GitHub
parent ad050107b3
commit 676d6923f2
2 changed files with 8 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
python/llm/src/bigdl/llm/langchain/embeddings/transformersembeddings.py
View file

@ -135,7 +135,7 @@ class TransformersEmbeddings(BaseModel, Embeddings):
extra = Extra.forbid
def embed(self, text: str):
def embed(self, text: str, **kwargs):
"""Compute doc embeddings using a HuggingFace transformer model.
Args:
@ -144,7 +144,7 @@ class TransformersEmbeddings(BaseModel, Embeddings):
Returns:
List of embeddings, one for each text.
"""
input_ids = self.tokenizer.encode(text, return_tensors="pt") # shape: [1, T]
input_ids = self.tokenizer.encode(text, return_tensors="pt", **kwargs) # shape: [1, T]
embeddings = self.model(input_ids, return_dict=False)[0] # shape: [1, T, N]
embeddings = embeddings.squeeze(0).detach().numpy()
embeddings = np.mean(embeddings, axis=0)

6
python/llm/test/langchain/test_transformers_api.py
View file

@ -130,6 +130,12 @@ class Test_Langchain_Transformers_API(TestCase):
self.assertTrue(res)
"""
def test_embed_kwargs(self):
embeddings = TransformersEmbeddings.from_model_id(model_id=self.llama_model_path)
encode_kwargs = {"truncation": True, "max_length": 512}
en_texts = ["hello","goodbye"]
embeddings.embed(en_texts,**encode_kwargs)
if __name__ == '__main__':
pytest.main([__file__])