app.py

'''
contains the infernce code for fastsam.py
'''
import os 
import sys 
import cv2 
import torch
import numpy as np 
import gradio as gr 
from PIL import Image 
import matplotlib.pyplot as plt 
from ultralytics import FastSAM 
from ultralytics.models.fastsam import FastSAMPrompt 

def inference(input_image,show_all=True,text_prompt=''):

    #model = FastSAM('FastSAM.pt')
    model = FastSAM(params['./FastSAM.pt'])
    device = torch.device("cuda") if torch.cuda.is_available() else "cpu"
    text_prompt = params['text_prompt']
    input_image = Image.fromarray(input_image).convert("RGB")
    w,h = input_image.size
    scale = params['input_size']/max(w,h) 

    new_w,new_h = int(w*scale),int(h*scale) 
    input_image = input_image.resize((new_w,new_h))
    

    ## inference results
    # Run inference on an image
    results = model(source, device='cpu', retina_masks=True, imgsz=1024, conf=0.4, iou=0.9)

    # Prepare a Prompt Process object
    prompt_process = FastSAMPrompt(source, results, device='cpu')

    if text_prompt:
        ann = prompt_process.text_prompt(text=text_prompt)
    if show_all:
        gr.Warning("Annotations are enabled, Overiding the text prompt!")
        ann = prompt_process.everything_prompt()

    result = prompt_process.plot_to_result(annotations=ann,bboxes=None,points=None,point_label=None,withContours=True,better_quality=False,)

    return result 


params = {
            'input_size':1024,
            'point_prompt':[[0,0]],
            'text_prompt':'jeep',
            'output_dir':'./testimg.png'
}

title = "Usage of FastSAM"
description = "Implementation of pre-trained fast-sam model for spaces."

demo = gr.Interface(inference,inputs=[gr.Image(sources=['clipboard','upload']),
                                      gr.Checkbox(True,label='Show All Annotations'),
                                      gr.Textbox('',label='Provide the text prompt: '),],
                    outputs = [gr.Image(label='Output')],
                    title = title,
                    description = description)

demo.launch(debug = True)