File size: 2,246 Bytes
ec5bfd8 |
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 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 |
import gradio as gr
import torch
from transformers import Qwen2VLForConditionalGeneration, AutoProcessor
from PIL import Image
from byaldi import RAGMultiModalModel
from qwen_vl_utils import process_vision_info
# Model and processor names
RAG_MODEL = "vidore/colpali"
QWN_MODEL = "Qwen/Qwen2-VL-7B-Instruct"
def load_models():
RAG = RAGMultiModalModel.from_pretrained(RAG_MODEL)
model = Qwen2VLForConditionalGeneration.from_pretrained(
QWN_MODEL,
torch_dtype=torch.bfloat16,
attn_implementation="flash_attention_2",
device_map="auto",
trust_remote_code=True
).eval()
processor = AutoProcessor.from_pretrained(QWN_MODEL, trust_remote_code=True)
return RAG, model, processor
RAG, model, processor = load_models()
def document_rag(image, text_query):
messages = [
{
"role": "user",
"content": [
{
"type": "image",
"image": image,
},
{"type": "text", "text": text_query},
],
}
]
text = processor.apply_chat_template(
messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
text=[text],
images=image_inputs,
videos=video_inputs,
padding=True,
return_tensors="pt",
)
inputs = inputs.to(model.device)
generated_ids = model.generate(**inputs, max_new_tokens=50)
generated_ids_trimmed = [
out_ids[len(in_ids):] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
return output_text[0]
# Define the Gradio interface
iface = gr.Interface(
fn=document_rag,
inputs=[
gr.Image(type="pil", label="Upload an image"),
gr.Textbox(label="Enter your text query")
],
outputs=gr.Textbox(label="Result"),
title="Document Processor",
description="Upload an image and enter a text query to process the document.",
)
# Launch the app
if __name__ == "__main__":
iface.launch() |