from transformers import AutoTokenizer, AutoModel
import torch

tokenizer = AutoTokenizer.from_pretrained("yenstdi/ntu-bge-small-zh-simcse-job-talent-matching")
model = AutoModel.from_pretrained("yenstdi/ntu-bge-small-zh-simcse-job-talent-matching")

def encode(texts):
    inputs = tokenizer(texts, padding=True, truncation=True, max_length=512, return_tensors="pt")
    with torch.no_grad():
        outputs = model(**inputs)
    mask = inputs["attention_mask"].unsqueeze(-1)
    embeddings = (outputs.last_hidden_state * mask).sum(1) / mask.sum(1).clamp(min=1e-6)
    return torch.nn.functional.normalize(embeddings, dim=-1)

job_emb = encode(["Software Engineer - Python, ML experience required"])
talent_emb = encode(["5 years Python developer with ML projects"])
score = (job_emb @ talent_emb.T).item()
print(f"Cosine similarity: {score:.4f}")

Trained by

National Taiwan University (NTU)