fix: deprecated arguments in stable diffusion demo (#10434)

* Refactor requirements and improve Hugging Face token handling

- Updated `requirements.txt` to include specific package versions, ensuring better dependency management and compatibility.
- Enhanced `run.py` to check for the Hugging Face access token (`HUGGING_FACE_ACCESS_TOKEN`) at runtime, providing a clear error message if it's missing.
- Replaced deprecated arguments (`use_auth_token`, `revision="fp16"`) with updated ones (`token`, `variant="fp16"`) in the StableDiffusionPipeline setup.
- Added minor formatting improvements to improve readability and maintainability in the `run.py` script.

These changes enhance user clarity, ensure compatibility with updated libraries, and improve the code's overall quality.

* chore: pin versions for diffusers and torch in requirements.txt

* refactor: use standard HF_TOKEN for authentication in run.py

* notebook

---------

Co-authored-by: Abubakar Abid <[email protected]>

demo/stable-diffusion/requirements.txt

@@ -1,6 +1,3 @@
-diffusers
-transformers
-nvidia-ml-py3
-ftfy
-torch
-Pillow
+diffusers==0.32.2
+torch==2.5.1
+

demo/stable-diffusion/run.ipynb

@@ -1 +1 @@
-{"cells": [{"cell_type": "markdown", "id": "302934307671667531413257853548643485645", "metadata": {}, "source": ["# Gradio Demo: stable-diffusion\n", "### Note: This is a simplified version of the code needed to create the Stable Diffusion demo. See full code here: https://hf.co/spaces/stabilityai/stable-diffusion/tree/main\n", "        "]}, {"cell_type": "code", "execution_count": null, "id": "272996653310673477252411125948039410165", "metadata": {}, "outputs": [], "source": ["!pip install -q gradio diffusers transformers nvidia-ml-py3 ftfy torch Pillow "]}, {"cell_type": "code", "execution_count": null, "id": "288918539441861185822528903084949547379", "metadata": {}, "outputs": [], "source": ["import gradio as gr\n", "import torch\n", "from diffusers import StableDiffusionPipeline  # type: ignore\n", "from PIL import Image\n", "import os\n", "\n", "auth_token = os.getenv(\"auth_token\")\n", "model_id = \"CompVis/stable-diffusion-v1-4\"\n", "device = \"cpu\"\n", "pipe = StableDiffusionPipeline.from_pretrained(\n", "    model_id, use_auth_token=auth_token, revision=\"fp16\", torch_dtype=torch.float16\n", ")\n", "pipe = pipe.to(device)\n", "\n", "def infer(prompt, samples, steps, scale, seed):\n", "    generator = torch.Generator(device=device).manual_seed(seed)\n", "    images_list = pipe(  # type: ignore\n", "        [prompt] * samples,\n", "        num_inference_steps=steps,\n", "        guidance_scale=scale,\n", "        generator=generator,\n", "    )\n", "    images = []\n", "    safe_image = Image.open(r\"unsafe.png\")\n", "    for i, image in enumerate(images_list[\"sample\"]):  # type: ignore\n", "        if images_list[\"nsfw_content_detected\"][i]:  # type: ignore\n", "            images.append(safe_image)\n", "        else:\n", "            images.append(image)\n", "    return images\n", "\n", "block = gr.Blocks()\n", "\n", "with block:\n", "    with gr.Group():\n", "        with gr.Row():\n", "            text = gr.Textbox(\n", "                label=\"Enter your prompt\",\n", "                max_lines=1,\n", "                placeholder=\"Enter your prompt\",\n", "                container=False,\n", "            )\n", "            btn = gr.Button(\"Generate image\")\n", "        gallery = gr.Gallery(\n", "            label=\"Generated images\",\n", "            show_label=False,\n", "            elem_id=\"gallery\",\n", "            columns=[2],\n", "        )\n", "\n", "        advanced_button = gr.Button(\"Advanced options\", elem_id=\"advanced-btn\")\n", "\n", "        with gr.Row(elem_id=\"advanced-options\"):\n", "            samples = gr.Slider(label=\"Images\", minimum=1, maximum=4, value=4, step=1)\n", "            steps = gr.Slider(label=\"Steps\", minimum=1, maximum=50, value=45, step=1)\n", "            scale = gr.Slider(\n", "                label=\"Guidance Scale\", minimum=0, maximum=50, value=7.5, step=0.1\n", "            )\n", "            seed = gr.Slider(\n", "                label=\"Seed\",\n", "                minimum=0,\n", "                maximum=2147483647,\n", "                step=1,\n", "                randomize=True,\n", "            )\n", "        gr.on([text.submit, btn.click], infer, inputs=[text, samples, steps, scale, seed], outputs=gallery)\n", "        advanced_button.click(\n", "            None,\n", "            [],\n", "            text,\n", "        )\n", "\n", "block.launch()\n"]}], "metadata": {}, "nbformat": 4, "nbformat_minor": 5}
+{"cells": [{"cell_type": "markdown", "id": "302934307671667531413257853548643485645", "metadata": {}, "source": ["# Gradio Demo: stable-diffusion\n", "### Note: This is a simplified version of the code needed to create the Stable Diffusion demo. See full code here: https://hf.co/spaces/stabilityai/stable-diffusion/tree/main\n", "        "]}, {"cell_type": "code", "execution_count": null, "id": "272996653310673477252411125948039410165", "metadata": {}, "outputs": [], "source": ["!pip install -q gradio diffusers==0.32.2 torch==2.5.1  "]}, {"cell_type": "code", "execution_count": null, "id": "288918539441861185822528903084949547379", "metadata": {}, "outputs": [], "source": ["import gradio as gr\n", "import torch\n", "from diffusers import StableDiffusionPipeline  # type: ignore\n", "from PIL import Image\n", "import os\n", "\n", "auth_token = os.getenv(\"HF_TOKEN\")\n", "if not auth_token:\n", "    print(\n", "        \"ERROR: No Hugging Face access token found.\\n\"\n", "        \"Please define an environment variable 'auth_token' before running.\\n\"\n", "        \"Example:\\n\"\n", "        \"  export HF_TOKEN=XXXXXXXX\\n\"\n", "    )\n", "\n", "model_id = \"CompVis/stable-diffusion-v1-4\"\n", "device = \"cpu\"\n", "pipe = StableDiffusionPipeline.from_pretrained(\n", "    model_id, token=auth_token, variant=\"fp16\", torch_dtype=torch.float16,\n", ")\n", "pipe = pipe.to(device)\n", "\n", "\n", "def infer(prompt, samples, steps, scale, seed):\n", "    generator = torch.Generator(device=device).manual_seed(seed)\n", "    images_list = pipe(  # type: ignore\n", "        [prompt] * samples,\n", "        num_inference_steps=steps,\n", "        guidance_scale=scale,\n", "        generator=generator,\n", "    )\n", "    images = []\n", "    safe_image = Image.open(r\"unsafe.png\")\n", "    for i, image in enumerate(images_list[\"sample\"]):  # type: ignore\n", "        if images_list[\"nsfw_content_detected\"][i]:  # type: ignore\n", "            images.append(safe_image)\n", "        else:\n", "            images.append(image)\n", "    return images\n", "\n", "\n", "block = gr.Blocks()\n", "\n", "with block:\n", "    with gr.Group():\n", "        with gr.Row():\n", "            text = gr.Textbox(\n", "                label=\"Enter your prompt\",\n", "                max_lines=1,\n", "                placeholder=\"Enter your prompt\",\n", "                container=False,\n", "            )\n", "            btn = gr.Button(\"Generate image\")\n", "        gallery = gr.Gallery(\n", "            label=\"Generated images\",\n", "            show_label=False,\n", "            elem_id=\"gallery\",\n", "            columns=[2],\n", "        )\n", "\n", "        advanced_button = gr.Button(\"Advanced options\", elem_id=\"advanced-btn\")\n", "\n", "        with gr.Row(elem_id=\"advanced-options\"):\n", "            samples = gr.Slider(label=\"Images\", minimum=1, maximum=4, value=4, step=1)\n", "            steps = gr.Slider(label=\"Steps\", minimum=1, maximum=50, value=45, step=1)\n", "            scale = gr.Slider(\n", "                label=\"Guidance Scale\", minimum=0, maximum=50, value=7.5, step=0.1\n", "            )\n", "            seed = gr.Slider(\n", "                label=\"Seed\",\n", "                minimum=0,\n", "                maximum=2147483647,\n", "                step=1,\n", "                randomize=True,\n", "            )\n", "        gr.on(\n", "            [text.submit, btn.click],\n", "            infer,\n", "            inputs=[text, samples, steps, scale, seed],\n", "            outputs=gallery,\n", "        )\n", "        advanced_button.click(\n", "            None,\n", "            [],\n", "            text,\n", "        )\n", "\n", "block.launch()\n"]}], "metadata": {}, "nbformat": 4, "nbformat_minor": 5}

demo/stable-diffusion/run.py

@@ -4,14 +4,23 @@
 from PIL import Image
 import os
 
-auth_token = os.getenv("auth_token")
+auth_token = os.getenv("HF_TOKEN")
+if not auth_token:
+    print(
+        "ERROR: No Hugging Face access token found.\n"
+        "Please define an environment variable 'auth_token' before running.\n"
+        "Example:\n"
+        "  export HF_TOKEN=XXXXXXXX\n"
+    )
+
 model_id = "CompVis/stable-diffusion-v1-4"
 device = "cpu"
 pipe = StableDiffusionPipeline.from_pretrained(
-    model_id, use_auth_token=auth_token, revision="fp16", torch_dtype=torch.float16
+    model_id, token=auth_token, variant="fp16", torch_dtype=torch.float16,
 )
 pipe = pipe.to(device)
 
+
 def infer(prompt, samples, steps, scale, seed):
     generator = torch.Generator(device=device).manual_seed(seed)
     images_list = pipe(  # type: ignore
@@ -29,6 +38,7 @@ def infer(prompt, samples, steps, scale, seed):
             images.append(image)
     return images
 
+
 block = gr.Blocks()
 
 with block:
@@ -63,7 +73,12 @@ def infer(prompt, samples, steps, scale, seed):
                 step=1,
                 randomize=True,
             )
-        gr.on([text.submit, btn.click], infer, inputs=[text, samples, steps, scale, seed], outputs=gallery)
+        gr.on(
+            [text.submit, btn.click],
+            infer,
+            inputs=[text, samples, steps, scale, seed],
+            outputs=gallery,
+        )
         advanced_button.click(
             None,
             [],