Change Model

app.py

@@ -43,59 +43,54 @@ def make_qr(url: str = "http://www.mybirdfire.com", size: int = 512, border: int
     img = qr.make_image(fill_color="black", back_color="white").convert("RGB")
     return img.resize((size, size), resample=Image.NEAREST)
 
-# ========= SDXL dual ControlNet stylizer (canny + softedge) =========
-from diffusers import StableDiffusionXLControlNetPipeline, ControlNetModel
+# ========= SD1.5 ControlNet stylizer (canny + tile) =========
+from diffusers import StableDiffusionControlNetPipeline, ControlNetModel
 from diffusers.schedulers.scheduling_euler_discrete import EulerDiscreteScheduler
 from controlnet_aux import CannyDetector
 
-SDXL_MODEL = "stabilityai/stable-diffusion-xl-base-1.0"   # swap to your SDXL anime model if desired
-CN_CANNY   = "diffusers/controlnet-canny-sdxl-1.0"
-CN_SOFT    = "diffusers/controlnet-softedge-sdxl-1.0"     # <-- replaces non-existent tile SDXL
-
-_sdxl = {"pipe": None}
-def _load_sdxl_dual():
-    if _sdxl["pipe"] is None:
-        cn1 = ControlNetModel.from_pretrained(CN_CANNY, torch_dtype=dtype)
-        cn2 = ControlNetModel.from_pretrained(CN_SOFT,  torch_dtype=dtype)
-        pipe = StableDiffusionXLControlNetPipeline.from_pretrained(
-            SDXL_MODEL, controlnet=[cn1, cn2], torch_dtype=dtype, safety_checker=None
+BASE_15 = "runwayml/stable-diffusion-v1-5"
+CN_CANNY_15 = "lllyasviel/sd-controlnet-canny"
+CN_TILE_15  = "lllyasviel/control_v11f1e_sd15_tile"
+
+_cn = {"pipe": None}
+def _load_sd15_dual():
+    if _cn["pipe"] is None:
+        canny = ControlNetModel.from_pretrained(CN_CANNY_15, torch_dtype=dtype)
+        tile  = ControlNetModel.from_pretrained(CN_TILE_15,  torch_dtype=dtype)
+        pipe = StableDiffusionControlNetPipeline.from_pretrained(
+            BASE_15, controlnet=[canny, tile], torch_dtype=dtype, safety_checker=None
         ).to(device)
-        pipe.scheduler = EulerDiscreteScheduler.from_config(pipe.scheduler.config)
+        pipe.scheduler = EulerDiscreteScheduler.from_config(pipe.scheduler.config)  # Sampler: Euler
+        pipe.enable_attention_slicing()
         pipe.enable_vae_slicing()
-        _sdxl["pipe"] = pipe
-        _sdxl["canny"] = CannyDetector()
-    return _sdxl["pipe"], _sdxl["canny"]
+        _cn["pipe"] = pipe
+        _cn["canny_aux"] = CannyDetector()
+    return _cn["pipe"], _cn["canny_aux"]
 
-NEG = "lowres, low contrast, blurry, jpeg artifacts, worst quality, extra digits, bad anatomy"
+NEG_DEFAULT = "lowres, low contrast, blurry, jpeg artifacts, worst quality, bad anatomy, extra digits"
 
-def stylize_qr_sdxl(prompt: str, steps: int=28, guidance: float=7.0, seed: int=1470713301,
-                    canny_low: int=80, canny_high: int=160):
-    # 1) make a strong QR @1024
-    qr = make_qr("http://www.mybirdfire.com", size=1024, border=6)
-    # 2) edges for canny CN
-    pipe, canny = _load_sdxl_dual()
-    edges = canny(qr, low_threshold=int(canny_low), high_threshold=int(canny_high))
+def stylize_qr_sd15(prompt: str, negative: str, steps: int, guidance: float, seed: int,
+                    canny_low: int, canny_high: int, border: int):
+    # Make fresh QR each time
+    qr_img = make_qr("http://www.mybirdfire.com", size=512, border=int(border))
+    pipe, canny = _load_sd15_dual()
+    edges = canny(qr_img, low_threshold=int(canny_low), high_threshold=int(canny_high))
 
-    gen = torch.Generator(device=device).manual_seed(int(seed)) if int(seed)!=0 else None
+    gen = torch.Generator(device=device).manual_seed(int(seed)) if int(seed) != 0 else None
 
-    # Control weights + schedule (canny, softedge)
-    cn_scales  = [1.1, 0.6]
-    cn_start   = [0.25, 0.00]
-    cn_end     = [0.95, 1.00]
+    # Control weights (canny, tile). Tune if too “lego” or too artsy.
+    cn_scales = [1.2, 0.6]
 
     def run():
-        img = pipe(
-            prompt=prompt,
-            negative_prompt=NEG,
-            image=[edges, qr],                               # canny first, softedge second
+        return pipe(
+            prompt=str(prompt),
+            negative_prompt=negative or NEG_DEFAULT,
+            image=[edges, qr_img],                              # canny first, tile second
             controlnet_conditioning_scale=cn_scales,
-            control_guidance_start=cn_start,
-            control_guidance_end=cn_end,
             num_inference_steps=int(steps),
             guidance_scale=float(guidance),
             generator=gen
         ).images[0]
-        return img
 
     if device in ("cuda", "mps"):
         with torch.autocast(device):
@@ -104,7 +99,7 @@ def stylize_qr_sdxl(prompt: str, steps: int=28, guidance: float=7.0, seed: int=1
 
 # ========= UI =========
 with gr.Blocks() as demo:
-    gr.Markdown("## Stable Diffusion + QR Code + ControlNet")
+    gr.Markdown("## Stable Diffusion + QR Code + ControlNet (SD1.5)")
 
     with gr.Tab("Stable Diffusion (prompt → image)"):
         prompt = gr.Textbox(label="Prompt", value="Sky, Moon, Bird, Blue, In the dark, Goddess, Sweet, Beautiful, Fantasy, Art, Anime")
@@ -121,16 +116,22 @@ with gr.Blocks() as demo:
         quiet = gr.Slider(0, 8, value=4, step=1, label="Border (quiet zone)")
         out_qr = gr.Image(label="QR Code", type="pil")
         gr.Button("Generate QR").click(make_qr, [url, size, quiet], out_qr)
-        
-    with gr.Tab("QR Stylizer (SDXL canny + softedge)"):
-        p = gr.Textbox(label="Prompt", value="Sky, Moon, Bird, Blue, In the dark, Goddess, Sweet, Beautiful, Fantasy, Art, Anime")
-        st = gr.Slider(20, 40, 28, step=1, label="Steps")
-        cfg = gr.Slider(4.5, 9.0, 7.0, step=0.1, label="CFG")
-        sd = gr.Number(value=1470713301, label="Seed", precision=0)
-        cl = gr.Slider(0, 255, 80, step=1, label="Canny low")
-        ch = gr.Slider(0, 255, 160, step=1, label="Canny high")
-        out = gr.Image(label="Stylized QR (SDXL)")
-        gr.Button("Stylize").click(stylize_qr_sdxl, [p, st, cfg, sd, cl, ch], out)
+
+    with gr.Tab("QR Stylizer (SD1.5 canny + tile, Euler)"):
+        s_prompt = gr.Textbox(label="Style Prompt", value="Sky, Moon, Bird, Blue, In the dark, Goddess, Sweet, Beautiful, Fantasy, Art, Anime")
+        s_negative = gr.Textbox(label="Negative Prompt", value=NEG_DEFAULT)
+        s_steps  = gr.Slider(10, 50, value=28, label="Steps", step=1)
+        s_cfg    = gr.Slider(1, 12, value=7.0, label="CFG", step=0.1)
+        s_seed   = gr.Number(value=1470713301, label="Seed", precision=0)
+        canny_l  = gr.Slider(0, 255, value=80, step=1, label="Canny low")
+        canny_h  = gr.Slider(0, 255, value=160, step=1, label="Canny high")
+        s_border = gr.Slider(2, 10, value=6, step=1, label="QR border")
+        out_styl = gr.Image(label="Stylized QR")
+        gr.Button("Stylize").click(
+            stylize_qr_sd15,
+            [s_prompt, s_negative, s_steps, s_cfg, s_seed, canny_l, canny_h, s_border],
+            out_styl
+        )
 
 if __name__ == "__main__":
     demo.launch()

requirements.txt

@@ -6,3 +6,5 @@ safetensors
 gradio
 qrcode[pil]
 controlnet-aux
+mediapipe
+timm