embeddings.py

from transformers import AutoTokenizer, AutoModel
import torch

tokenizer = AutoTokenizer.from_pretrained("sentence-transformers/all-MiniLM-L6-v2")
model = AutoModel.from_pretrained("sentence-transformers/all-MiniLM-L6-v2")

def generate_embeddings(texts):
    # Tokenize the input texts
    encoded_input = tokenizer(texts, padding=True, truncation=True, return_tensors='pt')
    
    # Compute token embeddings
    with torch.no_grad():
        model_output = model(**encoded_input)
    
    # Mean Pooling - Take attention mask into account for correct averaging
    attention_mask = encoded_input['attention_mask']
    input_mask_expanded = attention_mask.unsqueeze(-1).expand(model_output.last_hidden_state.size()).float()
    sum_embeddings = torch.sum(model_output.last_hidden_state * input_mask_expanded, 1)
    sum_mask = torch.clamp(input_mask_expanded.sum(1), min=1e-9)
    embeddings = (sum_embeddings / sum_mask).squeeze()
    
    return embeddings.tolist()