agentic-rag / project /db /dense_embeddings.py
devank2000's picture
fix2
3366ea9 verified
Raw History Blame Contribute Delete
1.31 kB
import os
import spaces
from langchain_core.embeddings import Embeddings
from sentence_transformers import SentenceTransformer
class DenseEmbeddings(Embeddings):
def __init__(self, model_name: str):
is_zero_gpu = os.environ.get("SPACES_ZERO_GPU", "").lower() in {"1", "true"}
model = SentenceTransformer(
model_name,
device="cuda" if is_zero_gpu else "cpu",
)
dimension = model.get_embedding_dimension()
if not isinstance(dimension, int) or dimension <= 0:
raise ValueError(f"Cannot determine embedding dimension for '{model_name}'.")
self.vector_size = dimension
# Capture the model in the worker, rather than sending CUDA weights as arguments.
@spaces.GPU(duration=30)
def encode(texts: list[str]) -> list[list[float]]:
return model.encode(
texts,
show_progress_bar=False,
convert_to_numpy=True,
).tolist()
self._encode = encode
def embed_documents(self, texts: list[str]) -> list[list[float]]:
return self._encode([text.replace("\n", " ") for text in texts])
def embed_query(self, text: str) -> list[float]:
return self.embed_documents([text])[0]