Modelpublic

cs4248-nlp/paper-s3-a3-bimga-query-only-tinybert-general-4l-312d-taco-hf-20260402-015143

sourceHugging Facemitupdated 6mo agoView on Hugging Face

0likes16downloads

Model Card

cs4248-nlp/paper-s3-a3-bimga-query-only-tinybert-general-4l-312d-taco-hf-20260402-015143

Code-search embedding model trained with the CS4248 two-phase KD pipeline.

Model details

Field	Value
Role	`s3-A3-bimga-query-only`
Phase	Phase 2
Method	`s3-A3-bimga-query-only`
Dataset	`unknown`
Teacher	`unknown`
Student base	`unknown`
Phase 1 epochs	`unknown`
Phase 1 patience	`unknown`
Phase 2 epochs	`unknown`
Phase 2 patience	`unknown`
Batch size	`unknown`
Eval batch size	`unknown`
Learning rate	`unknown`
Seed	`unknown`
Run timestamp	`20260402_015143`

Usage

python

from transformers import AutoModel, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("cs4248-nlp/paper-s3-a3-bimga-query-only-tinybert-general-4l-312d-taco-hf-20260402-015143")
model = AutoModel.from_pretrained("cs4248-nlp/paper-s3-a3-bimga-query-only-tinybert-general-4l-312d-taco-hf-20260402-015143")

Mean-pool the last hidden state to get a fixed-size embedding:

python

import torch

def mean_pool(model_output, attention_mask):
    token_embeddings = model_output.last_hidden_state
    mask = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
    return (token_embeddings * mask).sum(1) / mask.sum(1).clamp(min=1e-9)

inputs = tokenizer("your query here", return_tensors="pt", truncation=True, max_length=160)
with torch.no_grad():
    outputs = model(**inputs)
embedding = mean_pool(outputs, inputs['attention_mask'])