Hugging Face's logo

Ascetu
/
yungen

Text Generation
Transformers
Safetensors
florence2
custom_code
Model card Files Community
yungen
File size: 3,276 Bytes 
8ee1ea3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
78
79
80
81
82
83
84
85
86
87
88
89
90
 
import gradio as gr
from transformers import AutoProcessor, AutoModelForCausalLM
import spaces


from PIL import Image 


import subprocess
subprocess.run('pip install flash-attn --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)

model = AutoModelForCausalLM.from_pretrained('Ascetu/yungen', trust_remote_code=True).to("cuda").eval()

processor = AutoProcessor.from_pretrained('Ascetu/yungen', trust_remote_code=True)


TITLE = "# [Florence-2-DocVQA Demo](https://huggingface.co/HuggingFaceM4/Florence-2-DocVQA)"
DESCRIPTION = "The demo for Florence-2 fine-tuned on DocVQA dataset. You can find the notebook [here](https://colab.research.google.com/drive/1hKDrJ5AH_o7I95PtZ9__VlCTNAo1Gjpf?usp=sharing). Read more about Florence-2 fine-tuning [here](finetune-florence2)."


colormap = ['blue','orange','green','purple','brown','pink','gray','olive','cyan','red',
            'lime','indigo','violet','aqua','magenta','coral','gold','tan','skyblue']

@spaces.GPU
def run_example(task_prompt, image, text_input=None):
    if text_input is None:
        prompt = task_prompt
    else:
        prompt = task_prompt + text_input
    inputs = processor(text=prompt, images=image, return_tensors="pt").to("cuda")
    generated_ids = model.generate(
        input_ids=inputs["input_ids"],
        pixel_values=inputs["pixel_values"],
        max_new_tokens=1024,
        early_stopping=False,
        do_sample=False,
        num_beams=3,
    )
    generated_text = processor.batch_decode(generated_ids, skip_special_tokens=False)[0]
    parsed_answer = processor.post_process_generation(
        generated_text,
        task=task_prompt,
        image_size=(image.width, image.height)
    )
    return parsed_answer

def process_image(image, text_input=None):
    image = Image.fromarray(image)  # Convert NumPy array to PIL Image
    task_prompt = '<DocVQA>'
    results = run_example(task_prompt, image, text_input)[task_prompt].replace("<pad>", "")
    return results


css = """
  #output {
    height: 500px; 
    overflow: auto; 
    border: 1px solid #ccc; 
  }
"""

with gr.Blocks(css=css) as demo:
    gr.Markdown(TITLE)
    gr.Markdown(DESCRIPTION)
    with gr.Tab(label="Florence-2 Image Captioning"):
        with gr.Row():
            with gr.Column():
                input_img = gr.Image(label="Input Picture")
                text_input = gr.Textbox(label="Text Input (optional)")
                submit_btn = gr.Button(value="Submit")
            with gr.Column():
                output_text = gr.Textbox(label="Output Text")

        gr.Examples(
            examples=[
                ["idefics2_architecture.png", 'How many tokens per image does it use?'],
                ["idefics2_architecture.png", "What type of encoder does the model use?"],
                ["idefics2_architecture.png", 'Up to which size can the images be?'],
                ["image.jpg", "What's the share of Industry Switchers Gained?"]
            ],
            inputs=[input_img, text_input],
            outputs=[output_text],
            fn=process_image,
            cache_examples=True,
            label='Try the examples below'
        )

        submit_btn.click(process_image, [input_img, text_input], [output_text])

demo.launch(debug=True)