Spaces:
Runtime error
Runtime error
File size: 4,030 Bytes
1d42b83 c431f44 1d42b83 c431f44 1d42b83 c431f44 1d42b83 c431f44 1d42b83 c431f44 1d42b83 |
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 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 |
import string
import gradio as gr
import requests
import torch
from transformers import BlipForQuestionAnswering, BlipProcessor
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
processor = BlipProcessor.from_pretrained("Salesforce/blip-vqa-capfilt-large")
model_vqa = BlipForQuestionAnswering.from_pretrained("Salesforce/blip-vqa-capfilt-large").to(device)
from transformers import BlipProcessor, Blip2ForConditionalGeneration
cap_processor = BlipProcessor.from_pretrained("Salesforce/blip2-flan-t5-xl")
cap_model = Blip2ForConditionalGeneration.from_pretrained("Salesforce/blip2-flan-t5-xl")
def caption(input_image):
inputs = processor(input_image, return_tensors="pt")
inputs["num_beams"] = 4
inputs['num_return_sequences'] =4
out = model.generate(**inputs)
return "\n".join(processor.decode(out[0], skip_special_tokens=True))
def gpt3(input_text):
response = openai.Completion.create(
engine="text-davinci-003",
prompt=input_text,
max_tokens=10,
n=1,
stop=None,
temperature=0.7,
)
answer = response.choices[0].text.strip()
return answer
def inference_chat(input_image,input_text):
inputs = processor(images=input_image, text=input_text,return_tensors="pt")
inputs["max_length"] = 10
inputs["num_beams"] = 5
inputs['num_return_sequences'] =4
out = model_vqa.generate(**inputs)
return "\n".join(processor.batch_decode(out, skip_special_tokens=True))
with gr.Blocks(
css="""
.message.svelte-w6rprc.svelte-w6rprc.svelte-w6rprc {font-size: 20px; margin-top: 20px}
#component-21 > div.wrap.svelte-w6rprc {height: 600px;}
"""
) as iface:
state = gr.State([])
#caption_output = None
#gr.Markdown(title)
#gr.Markdown(description)
#gr.Markdown(article)
with gr.Row():
with gr.Column(scale=1):
image_input = gr.Image(type="pil")
with gr.Row():
with gr.Column(scale=1):
chat_input = gr.Textbox(lines=1, label="VQA Input(问题输入)")
with gr.Row():
clear_button = gr.Button(value="Clear", interactive=True)
submit_button = gr.Button(
value="Submit", interactive=True, variant="primary"
)
cap_submit_button = gr.Button(
value="Submit", interactive=True, variant="primary"
)
with gr.Column():
caption_output = gr.Textbox(lines=0, label="VQA Output(模型答案输出)")
caption_output_v1 = gr.Textbox(lines=0, label="Caption Output(模型caption输出)")
image_input.change(
lambda: ("", "", []),
[],
[ caption_output, state],
queue=False,
)
chat_input.submit(
inference_chat,
[
image_input,
chat_input,
],
[ caption_output],
)
clear_button.click(
lambda: ("", [], []),
[],
[chat_input, state],
queue=False,
)
submit_button.click(
inference_chat,
[
image_input,
chat_input,
],
[caption_output],
)
cap_submit_button.click(
caption,
[
image_input,
],
[caption_output_v1],
)
# examples = gr.Examples(
# examples=examples,
# inputs=[image_input, chat_input],
# )
iface.queue(concurrency_count=1, api_open=False, max_size=10)
iface.launch(enable_queue=True) |