e5-small-Quran / README.md
mustapha's picture
Update README.md
ebd2338
|
raw
history blame
1.29 kB
metadata
license: mit
datasets:
  - M-AI-C/quran-en-tafssirs
language:
  - en
import torch.nn.functional as F

from torch import Tensor
from transformers import AutoTokenizer, AutoModel


def average_pool(last_hidden_states: Tensor,
                 attention_mask: Tensor) -> Tensor:
    last_hidden = last_hidden_states.masked_fill(~attention_mask[..., None].bool(), 0.0)
    return last_hidden.sum(dim=1) / attention_mask.sum(dim=1)[..., None]


# Each input text should start with "query: " or "passage: ".
# For tasks other than retrieval, you can simply use the "query: " prefix.
input_texts = ['query: Who is prophet known for patience',
               'query: Who is moses',
               "passage: passage 1",
               "passage: passage 2"]

tokenizer = AutoTokenizer.from_pretrained('intfloat/e5-small')
model = AutoModel.from_pretrained('intfloat/e5-small')

# Tokenize the input texts
batch_dict = tokenizer(input_texts, max_length=512, padding=True, truncation=True, return_tensors='pt')

outputs = model(**batch_dict)
embeddings = average_pool(outputs.last_hidden_state, batch_dict['attention_mask'])

# (Optionally) normalize embeddings
embeddings = F.normalize(embeddings, p=2, dim=1)
scores = (embeddings[:2] @ embeddings[2:].T) * 100
print(scores.tolist())