cs4248-nlp/ft-teacher-all-mpnet-base-v2-taco-20260326-110507

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

Model details

Field	Value
Role	ft-teacher
Phase	Phase 1
Method	ft-teacher
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	20260326_110507

Usage

from transformers import AutoModel, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("cs4248-nlp/ft-teacher-all-mpnet-base-v2-taco-20260326-110507")
model = AutoModel.from_pretrained("cs4248-nlp/ft-teacher-all-mpnet-base-v2-taco-20260326-110507")

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

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'])