File size: 1,625 Bytes
9261660 9618b1a 3f81199 9ac52f4 9618b1a 9261660 5a73fa0 9618b1a 5a73fa0 9618b1a 5a73fa0 9618b1a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 |
import gradio as gr
from deepseek_vl.models import VLChatProcessor, MultiModalityCausalLM
from deepseek_vl.utils.io import load_pil_images
import torch
model_path = "deepseek-ai/deepseek-vl-1.3b-chat"
vl_chat_processor = VLChatProcessor.from_pretrained(model_path)
tokenizer = vl_chat_processor.tokenizer
vl_gpt = MultiModalityCausalLM.from_pretrained(model_path, trust_remote_code=True).to("cpu")
def qa(image, question):
conversation = [
{"role": "User", "content": "<image_placeholder>" + question, "images": [image]},
{"role": "Assistant", "content": ""}
]
pil_images = load_pil_images(conversation)
prepare_inputs = vl_chat_processor(
conversations=conversation,
images=pil_images,
force_batchify=True
).to("cpu")
inputs_embeds = vl_gpt.prepare_inputs_embeds(**prepare_inputs)
outputs = vl_gpt.language_model.generate(
inputs_embeds=inputs_embeds,
attention_mask=prepare_inputs.attention_mask,
pad_token_id=tokenizer.eos_token_id,
bos_token_id=tokenizer.bos_token_id,
eos_token_id=tokenizer.eos_token_id,
max_new_tokens=512,
do_sample=False,
use_cache=True
)
answer = tokenizer.decode(outputs[0].cpu().tolist(), skip_special_tokens=True)
return answer
demo = gr.Interface(
fn=qa,
inputs=[gr.Image(type="pil", label="Upload Image"), gr.Textbox(label="Enter your question")],
outputs="text",
title="DeepSeek-VL Multimodal QA Demo",
description="Upload an image and enter a question. Experience DeepSeek-VL's vision-language capabilities."
)
demo.launch() |