File size: 1,221 Bytes
67a2e99 |
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 |
import gradio as gr
import numpy as np
import cv2
import os,glob
import json
with gr.Blocks() as demo:
video_upload = gr.UploadButton(label="Upload the Video", file_types=["video"])
slider = gr.Slider(maximum=200,interactive=True,steps=1)
frames = []
def get_frame(video):
frames.clear()
cap = cv2.VideoCapture(video.name)
i = 0
for i in range(201):
ret, frame = cap.read()
if ret == False:
break
frames.append(frame)
i += 1
cap.release()
cv2.destroyAllWindows()
video_upload.upload(fn=get_frame, inputs=[video_upload])
def return_frame(index):
img = frames[index]
return img
slider.change(return_frame,slider,gr.Image(shape=(1280, 720),type="numpy"))
question = gr.Textbox(label="Question")
model_type = gr.CheckboxGroup(["SurgGPT","LCGN"],label="Model Choice")
answer = gr.Textbox(label="Answer")
predict = gr.Button(value="Predict")
def predict_ans(index,question,model_choice):
return "hi"
predict.click(fn=predict_ans,inputs=[slider,question,model_type],outputs=[answer])
if __name__ == "__main__":
demo.launch()
|