app.py

import torch
from peft import PeftModel
from transformers import AutoTokenizer, AutoModel
import gradio as gr

BASE_MODEL = "Qwen/Qwen3-Embedding-0.6B"
LORA_MODEL = "gauravprasadgp/qwen3-embedding_0.6B_lora"

tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL, trust_remote_code=True)
base_model = AutoModel.from_pretrained(BASE_MODEL, trust_remote_code=True, torch_dtype=torch.float16)
model = PeftModel.from_pretrained(base_model, LORA_MODEL, torch_dtype=torch.float16)

model.eval()

def get_embedding(text):
    inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True)
    with torch.no_grad():
        outputs = model(**inputs)
        embedding = outputs.last_hidden_state[:, 0, :].squeeze().cpu().numpy()
    return embedding.tolist()

iface = gr.Interface(
    fn=get_embedding,
    inputs=gr.Textbox(lines=2, placeholder="Enter text to embed..."),
    outputs="json",
    title="Qwen3 Embedding (LoRA Fine-Tuned)",
    description="Enter a sentence to get the embedding vector using the fine-tuned Qwen3-0.6B model.",
    allow_flagging="never"
)

if __name__ == "__main__":
    iface.launch()