Upload 2 files

bg.py

@@ -0,0 +1,149 @@
+import cv2
+import os
+from PIL import Image
+import numpy as np
+import torch
+from torch.autograd import Variable
+from torchvision import transforms
+import torch.nn.functional as F
+import matplotlib.pyplot as plt
+import warnings
+import random
+import tempfile
+
+
+warnings.filterwarnings("ignore")
+
+os.system("git clone https://github.com/xuebinqin/DIS")
+os.system("mv DIS/IS-Net/* .")
+
+# project imports
+from data_loader_cache import normalize, im_reader, im_preprocess 
+from models import *
+
+#Helpers
+device = 'cuda' if torch.cuda.is_available() else 'cpu'
+
+
+class GOSNormalize(object):
+    '''
+    Normalize the Image using torch.transforms
+    '''
+    def __init__(self, mean=[0.485,0.456,0.406], std=[0.229,0.224,0.225]):
+        self.mean = mean
+        self.std = std
+
+    def __call__(self,image):
+        image = normalize(image,self.mean,self.std)
+        return image
+
+
+transform =  transforms.Compose([GOSNormalize([0.5,0.5,0.5],[1.0,1.0,1.0])])
+
+def load_image(im_path, hypar):
+    im = im_reader(im_path)
+    im, im_shp = im_preprocess(im, hypar["cache_size"])
+    im = torch.divide(im,255.0)
+    shape = torch.from_numpy(np.array(im_shp))
+    return transform(im).unsqueeze(0), shape.unsqueeze(0) # make a batch of image, shape
+
+
+def build_model(hypar,device):
+    net = hypar["model"]#GOSNETINC(3,1)
+
+    # convert to half precision
+    if(hypar["model_digit"]=="half"):
+        net.half()
+        for layer in net.modules():
+            if isinstance(layer, nn.BatchNorm2d):
+                layer.float()
+
+    net.to(device)
+
+    if(hypar["restore_model"]!=""):
+        net.load_state_dict(torch.load(hypar["model_path"]+"/"+hypar["restore_model"], map_location=device))
+        net.to(device)
+    net.eval()  
+    return net
+
+    
+def predict(net,  inputs_val, shapes_val, hypar, device):
+    '''
+    Given an Image, predict the mask
+    '''
+    net.eval()
+
+    if(hypar["model_digit"]=="full"):
+        inputs_val = inputs_val.type(torch.FloatTensor)
+    else:
+        inputs_val = inputs_val.type(torch.HalfTensor)
+
+  
+    inputs_val_v = Variable(inputs_val, requires_grad=False).to(device) # wrap inputs in Variable
+   
+    ds_val = net(inputs_val_v)[0] # list of 6 results
+
+    pred_val = ds_val[0][0,:,:,:] # B x 1 x H x W    # we want the first one which is the most accurate prediction
+
+    ## recover the prediction spatial size to the orignal image size
+    pred_val = torch.squeeze(F.upsample(torch.unsqueeze(pred_val,0),(shapes_val[0][0],shapes_val[0][1]),mode='bilinear'))
+
+    ma = torch.max(pred_val)
+    mi = torch.min(pred_val)
+    pred_val = (pred_val-mi)/(ma-mi) # max = 1
+
+    if device == 'cuda': torch.cuda.empty_cache()
+    return (pred_val.detach().cpu().numpy()*255).astype(np.uint8) # it is the mask we need
+    
+# Set Parameters
+hypar = {} # paramters for inferencing
+
+
+hypar["model_path"] ="./saved_models" ## load trained weights from this path
+hypar["restore_model"] = "isnet.pth" ## name of the to-be-loaded weights
+hypar["interm_sup"] = False ## indicate if activate intermediate feature supervision
+
+##  choose floating point accuracy --
+hypar["model_digit"] = "full" ## indicates "half" or "full" accuracy of float number
+hypar["seed"] = 0
+
+hypar["cache_size"] = [1024, 1024] ## cached input spatial resolution, can be configured into different size
+
+## data augmentation parameters ---
+hypar["input_size"] = [1024, 1024] ## mdoel input spatial size, usually use the same value hypar["cache_size"], which means we don't further resize the images
+hypar["crop_size"] = [1024, 1024] ## random crop size from the input, it is usually set as smaller than hypar["cache_size"], e.g., [920,920] for data augmentation
+
+hypar["model"] = ISNetDIS()
+
+ # Build Model
+net = build_model(hypar, device)
+
+
+
+def inference(image: Image):
+    # Save the image to a temporary file
+    with tempfile.NamedTemporaryFile(suffix='.jpg') as temp:
+        image.save(temp.name)
+        image_path = temp.name
+    
+        image_tensor, orig_size = load_image(image_path, hypar) 
+        mask = predict(net, image_tensor, orig_size, hypar, device)
+
+        pil_mask = Image.fromarray(mask).convert('L')
+        im_rgb = image.convert("RGBA")
+
+        im_rgba = im_rgb.copy()
+        im_rgba.putalpha(pil_mask)
+        output_path = "output.png"
+        im_rgba.save(output_path, format="PNG", mode="RGBA")
+
+        return im_rgba
+        
+def paste_transparent_image(output_path, transparent_path):
+    transparent_image = transparent_path.resize(output_path.size)
+    
+    result_image = Image.alpha_composite(output_path.convert('RGBA'), transparent_image)
+
+    return result_image
+
+

handler.py

@@ -0,0 +1,140 @@
+from typing import  Dict, List, Any
+import base64
+from PIL import Image
+from io import BytesIO
+from diffusers import StableDiffusionControlNetPipeline, ControlNetModel
+import torch
+from bg import inference
+from bg import paste_transparent_image
+
+
+
+
+import numpy as np
+import cv2
+import controlnet_hinter
+
+# set device
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+if device.type != 'cuda':
+    raise ValueError("need to run on GPU")
+# set mixed precision dtype
+dtype = torch.bfloat16 if torch.cuda.get_device_capability()[0] == 8 else torch.float16
+
+# controlnet mapping for controlnet id and control hinter
+CONTROLNET_MAPPING = {
+    "canny_edge": {
+        "model_id": "lllyasviel/sd-controlnet-canny",
+        "hinter": controlnet_hinter.hint_canny
+    },
+    "pose": {
+        "model_id": "lllyasviel/sd-controlnet-openpose",
+        "hinter": controlnet_hinter.hint_openpose
+    },
+    "depth": {
+        "model_id": "lllyasviel/sd-controlnet-depth",
+        "hinter": controlnet_hinter.hint_depth
+    },
+    "scribble": {
+        "model_id": "lllyasviel/sd-controlnet-scribble",
+        "hinter": controlnet_hinter.hint_scribble,
+    },
+    "segmentation": {
+        "model_id": "lllyasviel/sd-controlnet-seg",
+        "hinter": controlnet_hinter.hint_segmentation,
+    },
+    "normal": {
+        "model_id": "lllyasviel/sd-controlnet-normal",
+        "hinter": controlnet_hinter.hint_normal,
+    },
+    "hed": {
+        "model_id": "lllyasviel/sd-controlnet-hed",
+        "hinter": controlnet_hinter.hint_hed,
+    },
+    "hough": {
+        "model_id": "lllyasviel/sd-controlnet-mlsd",
+        "hinter": controlnet_hinter.hint_hough,
+    }
+}
+
+
+class EndpointHandler():
+    def __init__(self, path=""):
+        # define default controlnet id and load controlnet
+        self.control_type = "normal"
+        self.controlnet = ControlNetModel.from_pretrained(CONTROLNET_MAPPING[self.control_type]["model_id"],torch_dtype=dtype).to(device)
+        
+        # Load StableDiffusionControlNetPipeline 
+        self.stable_diffusion_id = "runwayml/stable-diffusion-v1-5"
+        self.pipe = StableDiffusionControlNetPipeline.from_pretrained(self.stable_diffusion_id, 
+                                                                      controlnet=self.controlnet, 
+                                                                      torch_dtype=dtype,
+                                                                      safety_checker=None).to(device)
+        # Define Generator with seed
+        self.generator = torch.Generator(device="cpu").manual_seed(3)
+
+    def __call__(self, data: Any) -> List[List[Dict[str, float]]]:
+        """
+        :param data: A dictionary contains `inputs` and optional `image` field.
+        :return: A dictionary with `image` field contains image in base64.
+        """
+        prompt = data.pop("inputs", None)
+        image = data.pop("image", None)
+        controlnet_type = data.pop("controlnet_type", None)
+        
+        # Check if neither prompt nor image is provided
+        if prompt is None and image is None:
+            return {"error": "Please provide a prompt and base64 encoded image."}
+        
+        # Check if a new controlnet is provided
+        if controlnet_type is not None and controlnet_type != self.control_type:
+            print(f"changing controlnet from {self.control_type} to {controlnet_type} using {CONTROLNET_MAPPING[controlnet_type]['model_id']} model")
+            self.control_type = controlnet_type
+            self.controlnet = ControlNetModel.from_pretrained(CONTROLNET_MAPPING[self.control_type]["model_id"],
+                                                              torch_dtype=dtype).to(device)
+            self.pipe.controlnet = self.controlnet
+        
+        
+        # hyperparamters
+        negatice_prompt = data.pop("negative_prompt", None)
+        num_inference_steps = data.pop("num_inference_steps", 30)
+        guidance_scale = data.pop("guidance_scale", 7.5)
+        negative_prompt = data.pop("negative_prompt", None)
+        height = data.pop("height", None)
+        width = data.pop("width", None)
+        controlnet_conditioning_scale = data.pop("controlnet_conditioning_scale", 1.0)
+        
+        # process image
+
+        image = self.decode_base64_image(image)
+        image = inference(image)
+        control_image = CONTROLNET_MAPPING[self.control_type]["hinter"](image)
+        
+        # run inference pipeline
+        out = self.pipe(
+            prompt=prompt, 
+            negative_prompt=negative_prompt,
+            image=control_image,
+            num_inference_steps=num_inference_steps, 
+            guidance_scale=guidance_scale,
+            num_images_per_prompt=1,
+            height=height,
+            width=width,
+            controlnet_conditioning_scale=controlnet_conditioning_scale,
+            generator=self.generator
+        )
+        result_image = paste_transparent_image(out.images[0], image)
+        
+        # convert the resulting PIL image to a base64-encoded string
+        
+        # return first generate PIL image
+        return  result_image
+
+    
+    # helper to decode input image
+
+    def decode_base64_image(self, image_string):
+        base64_image = base64.b64decode(image_string)
+        buffer = BytesIO(base64_image)
+        image = Image.open(buffer)
+        return image