Update app.py

app.py

@@ -14,19 +14,30 @@ from annotator.openpose import apply_openpose
 from cldm.model import create_model, load_state_dict
 
 from huggingface_hub import hf_hub_url, cached_download
+
 REPO_ID = "lllyasviel/ControlNet"
-FILENAME = "models/control_sd15_canny.pth"
+canny_checkpoint = "models/control_sd15_canny.pth"
+pose_checkpoint = "models/control_sd15_openpose.pth"
 
-model = create_model('./models/cldm_v15.yaml')
-model.load_state_dict(load_state_dict(cached_download(
-    hf_hub_url(REPO_ID, FILENAME)
+canny_model = create_model('./models/cldm_v15.yaml')
+canny_model.load_state_dict(load_state_dict(cached_download(
+    hf_hub_url(REPO_ID, canny_checkpoint)
 ), location='cpu'))
-ddim_sampler = DDIMSampler(model)
+ddim_sampler_canny = DDIMSampler(canny_model)
+
+
+pose_model = create_model('./models/cldm_v15.yaml')
+pose_model.load_state_dict(load_state_dict(cached_download(
+    hf_hub_url(REPO_ID, pose_checkpoint)
+), location='cpu'))
+ddim_sampler_pose = DDIMSampler(pose_model)
 
 def process(input_image, prompt, input_control, a_prompt, n_prompt, num_samples, image_resolution, ddim_steps, scale, seed, eta, low_threshold, high_threshold):
     # TODO: Add other control tasks
-    return process_canny(input_image, prompt, a_prompt, n_prompt, num_samples, image_resolution, ddim_steps, scale, seed, eta, low_threshold, high_threshold)
-    
+    if input_control == "Canny Edge Map":
+        return process_canny(input_image, prompt, a_prompt, n_prompt, num_samples, image_resolution, ddim_steps, scale, seed, eta, low_threshold, high_threshold)
+    else:
+        return process_pose(input_image, prompt, a_prompt, n_prompt, num_samples, image_resolution, ddim_steps, scale, seed, eta)
 
 def process_canny(input_image, prompt, a_prompt, n_prompt, num_samples, image_resolution, ddim_steps, scale, seed, eta, low_threshold, high_threshold):
     with torch.no_grad():
@@ -42,24 +53,24 @@ def process_canny(input_image, prompt, a_prompt, n_prompt, num_samples, image_re
 
         seed_everything(seed)
 
-        cond = {"c_concat": [control], "c_crossattn": [model.get_learned_conditioning([prompt + ', ' + a_prompt] * num_samples)]}
-        un_cond = {"c_concat": [control], "c_crossattn": [model.get_learned_conditioning([n_prompt] * num_samples)]}
+        cond = {"c_concat": [control], "c_crossattn": [canny_model.get_learned_conditioning([prompt + ', ' + a_prompt] * num_samples)]}
+        un_cond = {"c_concat": [control], "c_crossattn": [canny_model.get_learned_conditioning([n_prompt] * num_samples)]}
         shape = (4, H // 8, W // 8)
 
-        samples, intermediates = ddim_sampler.sample(ddim_steps, num_samples,
+        samples, intermediates = ddim_sampler_canny.sample(ddim_steps, num_samples,
                                                      shape, cond, verbose=False, eta=eta,
                                                      unconditional_guidance_scale=scale,
                                                      unconditional_conditioning=un_cond)
-        x_samples = model.decode_first_stage(samples)
+        x_samples = canny_model.decode_first_stage(samples)
         x_samples = (einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 + 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
 
         results = [x_samples[i] for i in range(num_samples)]
     return [255 - detected_map] + results
 
-def process_pose(input_image, prompt, a_prompt, n_prompt, num_samples, image_resolution, detect_resolution, ddim_steps, scale, seed, eta):
+def process_pose(input_image, prompt, a_prompt, n_prompt, num_samples, image_resolution, ddim_steps, scale, seed, eta):
     with torch.no_grad():
         input_image = HWC3(input_image)
-        detected_map, _ = apply_openpose(resize_image(input_image, detect_resolution))
+        detected_map, _ = apply_openpose(resize_image(input_image, image_resolution))
         detected_map = HWC3(detected_map)
         img = resize_image(input_image, image_resolution)
         H, W, C = img.shape
@@ -72,15 +83,15 @@ def process_pose(input_image, prompt, a_prompt, n_prompt, num_samples, image_res
 
         seed_everything(seed)
 
-        cond = {"c_concat": [control], "c_crossattn": [model.get_learned_conditioning([prompt + ', ' + a_prompt] * num_samples)]}
-        un_cond = {"c_concat": [control], "c_crossattn": [model.get_learned_conditioning([n_prompt] * num_samples)]}
+        cond = {"c_concat": [control], "c_crossattn": [pose_model.get_learned_conditioning([prompt + ', ' + a_prompt] * num_samples)]}
+        un_cond = {"c_concat": [control], "c_crossattn": [pose_model.get_learned_conditioning([n_prompt] * num_samples)]}
         shape = (4, H // 8, W // 8)
 
-        samples, intermediates = ddim_sampler.sample(ddim_steps, num_samples,
+        samples, intermediates = ddim_sampler_pose.sample(ddim_steps, num_samples,
                                                      shape, cond, verbose=False, eta=eta,
                                                      unconditional_guidance_scale=scale,
                                                      unconditional_conditioning=un_cond)
-        x_samples = model.decode_first_stage(samples)
+        x_samples = pose_model.decode_first_stage(samples)
         x_samples = (einops.rearrange(x_samples, 'b c h w -> b h w c') * 127.5 + 127.5).cpu().numpy().clip(0, 255).astype(np.uint8)
 
         results = [x_samples[i] for i in range(num_samples)]