Update

gradio_app.py

@@ -8,20 +8,9 @@ import shutil
 from typing import Generator, Optional, Tuple
 import logging
 
-try:
-    import gradio as gr
-except ImportError:
-    print("❌ Gradio not found. Please install it with: pip install gradio>=4.0.0")
-    sys.exit(1)
-
-try:
-    from huggingface_hub import hf_hub_download, snapshot_download
-except ImportError:
-    print(
-        "❌ huggingface_hub not found. Please install it with: pip install huggingface_hub"
-    )
-    sys.exit(1)
-
+import gradio as gr
+import spaces
+from huggingface_hub import hf_hub_download
 import torch
 from PIL import Image
 
@@ -57,9 +46,6 @@ training_state = TrainingState()
 
 def ensure_models_available():
     """Download models from HuggingFace if they're not available locally"""
-    models_dir = "models"
-
-    # Check if models directory exists and has the required files
     required_files = [
         "models/emlnet/res_decoder.pth",
         "models/emlnet/res_imagenet.pth",
@@ -115,6 +101,11 @@ def ensure_models_available():
         print("✅ Model files are already available locally.")
 
 
+# Initialize models and setup at module level for ZeroGPU packing
+ensure_models_available()
+torch.hub.set_dir("models/torch")
+
+
 def create_args_from_config(
     image_path: str,
     exp_name: str,
@@ -177,8 +168,9 @@ def create_args_from_config(
     return args
 
 
+@spaces.GPU(duration=300)  # Request GPU for up to 300 seconds (5 minutes)
 def train_model(args: argparse.Namespace) -> None:
-    """Training function that runs in a separate thread"""
+    """Training function that runs with ZeroGPU allocation"""
     try:
         # Create and train model with streaming results
         training_state.model = GradioGaussianSplatting2D(args, training_state.results)
@@ -351,21 +343,6 @@ def start_training_and_stream(
                         f"Loss: {metrics['loss']:.4f}"
                     )
                     logs_text += status_line
-
-                    # Add image status info for debugging
-                    if training_state.results.current_render is not None:
-                        logs_text += f"\n📸 Current render: {training_state.results.current_render.size}"
-                    else:
-                        logs_text += "\n📸 Current render: None"
-
-                    if training_state.results.current_gaussian_id is not None:
-                        logs_text += f"\n🆔 Gaussian ID: {training_state.results.current_gaussian_id.size}"
-                    else:
-                        logs_text += "\n🆔 Gaussian ID: None"
-
-                    logs_text += (
-                        f"\n💾 Stored steps: {len(training_state.results.step_renders)}"
-                    )
             else:
                 logs_text = "Waiting for training to start..."
 
@@ -509,307 +486,281 @@ def update_step_slider_after_training() -> gr.Slider:
     )
 
 
-def create_interface():
-    """Create the Gradio interface"""
-
-    with gr.Blocks(
-        title="Image-GS: 2D Gaussian Splatting", theme=gr.themes.Soft()
-    ) as demo:
-        gr.Markdown("""
-        # Image-GS: Content-Adaptive Image Representation via 2D Gaussians
-        
-        Upload an image and configure parameters to train a 2D Gaussian Splatting representation.
-        """)
-
-        with gr.Row():
-            with gr.Column(scale=1):
-                gr.Markdown("## Configuration")
+# Create Gradio interface at top level (best practice for Spaces)
+with gr.Blocks(title="Image-GS: 2D Gaussian Splatting", theme=gr.themes.Soft()) as demo:
+    gr.Markdown("""
+    # Image-GS: Content-Adaptive Image Representation via 2D Gaussians
+    
+    Upload an image and configure parameters to train a 2D Gaussian Splatting representation.
+    """)
+
+    with gr.Row():
+        with gr.Column(scale=1):
+            gr.Markdown("## Configuration")
+
+            # Image upload
+            image_input = gr.Image(
+                label="Input Image",
+                type="pil",
+                height=300,
+                sources=["upload"],
+                show_label=True,
+            )
 
-                # Image upload
-                image_input = gr.Image(
-                    label="Input Image",
-                    type="pil",
-                    height=300,
-                    sources=["upload"],
-                    show_label=True,
+            # Basic parameters
+            with gr.Group():
+                gr.Markdown("### Basic Parameters")
+                exp_name = gr.Textbox(
+                    label="Experiment Name",
+                    value="gradio_experiment",
+                    info="Name for this training run",
+                )
+                num_gaussians = gr.Slider(
+                    minimum=100,
+                    maximum=50000,
+                    value=10000,
+                    step=1000,
+                    label="Number of Gaussians",
+                    info="Number of Gaussians (for compression rate control). More = higher quality but slower training",
+                )
+                max_steps = gr.Slider(
+                    minimum=100,
+                    maximum=20000,
+                    value=10000,
+                    step=100,
+                    label="Maximum Training Steps",
+                    info="Maximum number of optimization steps. Default: 10000",
                 )
 
-                # Basic parameters
-                with gr.Group():
-                    gr.Markdown("### Basic Parameters")
-                    exp_name = gr.Textbox(
-                        label="Experiment Name",
-                        value="gradio_experiment",
-                        info="Name for this training run",
-                    )
-                    num_gaussians = gr.Slider(
-                        minimum=100,
-                        maximum=50000,
-                        value=10000,
-                        step=1000,
-                        label="Number of Gaussians",
-                        info="Number of Gaussians (for compression rate control). More = higher quality but slower training",
-                    )
-                    max_steps = gr.Slider(
-                        minimum=100,
-                        maximum=20000,
-                        value=10000,
-                        step=100,
-                        label="Maximum Training Steps",
-                        info="Maximum number of optimization steps. Default: 10000",
-                    )
-
-                # Quantization parameters
-                with gr.Group():
-                    gr.Markdown("### Quantization")
-                    quantize = gr.Checkbox(
-                        label="Enable Quantization",
-                        value=False,
-                        info="Enable bit precision control of Gaussian parameters. Reduces memory usage.",
-                    )
-                    with gr.Row():
-                        pos_bits = gr.Slider(
-                            4,
-                            32,
-                            16,
-                            step=1,
-                            label="Position Bits",
-                            info="Bit precision of individual coordinate dimension",
-                        )
-                        scale_bits = gr.Slider(
-                            4,
-                            32,
-                            16,
-                            step=1,
-                            label="Scale Bits",
-                            info="Bit precision of individual scale dimension",
-                        )
-                    with gr.Row():
-                        rot_bits = gr.Slider(
-                            4,
-                            32,
-                            16,
-                            step=1,
-                            label="Rotation Bits",
-                            info="Bit precision of Gaussian orientation angle",
-                        )
-                        feat_bits = gr.Slider(
-                            4,
-                            32,
-                            16,
-                            step=1,
-                            label="Feature Bits",
-                            info="Bit precision of individual feature dimension",
-                        )
-
-                # Initialization parameters
-                with gr.Group():
-                    gr.Markdown("### Initialization")
-                    init_mode = gr.Radio(
-                        choices=["gradient", "saliency", "random"],
-                        value="saliency",
-                        label="Initialization Mode",
-                        info="Gaussian position initialization mode. Gradient uses image gradients, saliency uses attention maps.",
+            # Quantization parameters
+            with gr.Group():
+                gr.Markdown("### Quantization")
+                quantize = gr.Checkbox(
+                    label="Enable Quantization",
+                    value=False,
+                    info="Enable bit precision control of Gaussian parameters. Reduces memory usage.",
+                )
+                with gr.Row():
+                    pos_bits = gr.Slider(
+                        4,
+                        32,
+                        16,
+                        step=1,
+                        label="Position Bits",
+                        info="Bit precision of individual coordinate dimension",
                     )
-                    init_random_ratio = gr.Slider(
-                        minimum=0.0,
-                        maximum=1.0,
-                        value=0.3,
-                        step=0.1,
-                        label="Random Ratio",
-                        info="Ratio of Gaussians with randomly initialized position (default: 0.3)",
+                    scale_bits = gr.Slider(
+                        4,
+                        32,
+                        16,
+                        step=1,
+                        label="Scale Bits",
+                        info="Bit precision of individual scale dimension",
                     )
-
-                # Advanced parameters (collapsible)
-                with gr.Accordion("Advanced Parameters", open=False):
-                    # Loss parameters
-                    gr.Markdown("#### Loss Weights")
-                    with gr.Row():
-                        l1_loss_ratio = gr.Slider(
-                            0.0, 2.0, 1.0, step=0.1, label="L1 Loss"
-                        )
-                        l2_loss_ratio = gr.Slider(
-                            0.0, 2.0, 0.0, step=0.1, label="L2 Loss"
-                        )
-                        ssim_loss_ratio = gr.Slider(
-                            0.0, 1.0, 0.1, step=0.01, label="SSIM Loss"
-                        )
-
-                    # Learning rates
-                    gr.Markdown("#### Learning Rates")
-                    with gr.Row():
-                        pos_lr = gr.Number(value=5e-4, label="Position LR", precision=6)
-                        scale_lr = gr.Number(value=2e-3, label="Scale LR", precision=6)
-                    with gr.Row():
-                        rot_lr = gr.Number(value=2e-3, label="Rotation LR", precision=6)
-                        feat_lr = gr.Number(value=5e-3, label="Feature LR", precision=6)
-
-                    # Optimization options
-                    gr.Markdown("#### Optimization")
-                    disable_lr_schedule = gr.Checkbox(
-                        label="Disable LR Schedule",
-                        value=False,
-                        info="Keep learning rate constant",
+                with gr.Row():
+                    rot_bits = gr.Slider(
+                        4,
+                        32,
+                        16,
+                        step=1,
+                        label="Rotation Bits",
+                        info="Bit precision of Gaussian orientation angle",
                     )
-                    disable_prog_optim = gr.Checkbox(
-                        label="Disable Progressive Optimization",
-                        value=False,
-                        info="Don't add Gaussians during training",
+                    feat_bits = gr.Slider(
+                        4,
+                        32,
+                        16,
+                        step=1,
+                        label="Feature Bits",
+                        info="Bit precision of individual feature dimension",
                     )
 
-                # Visualization parameters
-                with gr.Group():
-                    gr.Markdown("### Visualization")
-                    vis_gaussians = gr.Checkbox(
-                        label="Visualize Gaussians",
-                        value=True,
-                        info="Visualize Gaussians during optimization (default: True)",
-                    )
-                    save_image_steps = gr.Slider(
-                        minimum=200,
-                        maximum=10000,
-                        value=200,
-                        step=100,
-                        label="Save Image Every N Steps",
-                        info="Frequency of rendering intermediate results during optimization (default: 100)",
-                    )
+            # Initialization parameters
+            with gr.Group():
+                gr.Markdown("### Initialization")
+                init_mode = gr.Radio(
+                    choices=["gradient", "saliency", "random"],
+                    value="saliency",
+                    label="Initialization Mode",
+                    info="Gaussian position initialization mode. Gradient uses image gradients, saliency uses attention maps.",
+                )
+                init_random_ratio = gr.Slider(
+                    minimum=0.0,
+                    maximum=1.0,
+                    value=0.3,
+                    step=0.1,
+                    label="Random Ratio",
+                    info="Ratio of Gaussians with randomly initialized position (default: 0.3)",
+                )
 
-                # Control buttons
+            # Advanced parameters (collapsible)
+            with gr.Accordion("Advanced Parameters", open=False):
+                # Loss parameters
+                gr.Markdown("#### Loss Weights")
                 with gr.Row():
-                    start_btn = gr.Button(
-                        "Start Training", variant="primary", size="lg"
+                    l1_loss_ratio = gr.Slider(0.0, 2.0, 1.0, step=0.1, label="L1 Loss")
+                    l2_loss_ratio = gr.Slider(0.0, 2.0, 0.0, step=0.1, label="L2 Loss")
+                    ssim_loss_ratio = gr.Slider(
+                        0.0, 1.0, 0.1, step=0.01, label="SSIM Loss"
                     )
-                    stop_btn = gr.Button("Stop Training", variant="stop", size="lg")
-
-                status_text = gr.Textbox(label="Status", interactive=False, lines=2)
-
-            with gr.Column(scale=2):
-                gr.Markdown("## Training Progress")
 
-                # Progress logs (streaming)
-                progress_logs = gr.Textbox(
-                    label="Training Logs",
-                    lines=10,
-                    max_lines=15,
-                    interactive=False,
-                    autoscroll=True,
+                # Learning rates
+                gr.Markdown("#### Learning Rates")
+                with gr.Row():
+                    pos_lr = gr.Number(value=5e-4, label="Position LR", precision=6)
+                    scale_lr = gr.Number(value=2e-3, label="Scale LR", precision=6)
+                with gr.Row():
+                    rot_lr = gr.Number(value=2e-3, label="Rotation LR", precision=6)
+                    feat_lr = gr.Number(value=5e-3, label="Feature LR", precision=6)
+
+                # Optimization options
+                gr.Markdown("#### Optimization")
+                disable_lr_schedule = gr.Checkbox(
+                    label="Disable LR Schedule",
+                    value=False,
+                    info="Keep learning rate constant",
                 )
-
-                # Initial map (computed at start based on initialization mode)
-                gr.Markdown("### Initialization Map")
-                initialization_map = gr.Image(
-                    label="Initialization Map",
-                    type="pil",
-                    height=200,
+                disable_prog_optim = gr.Checkbox(
+                    label="Disable Progressive Optimization",
+                    value=False,
+                    info="Don't add Gaussians during training",
                 )
 
-                # Training images (streaming)
-                gr.Markdown("### Current Training Results")
-                with gr.Row():
-                    current_render = gr.Image(
-                        label="Current Render",
-                        type="pil",
-                        height=300,
-                        show_label=True,
-                        show_download_button=True,
-                    )
-                    current_gaussian_id = gr.Image(
-                        label="Gaussian ID",
-                        type="pil",
-                        height=300,
-                        show_label=True,
-                        show_download_button=True,
-                    )
-
-                # Step slider for interactive browsing (will be updated dynamically)
-                step_slider = gr.Slider(
-                    minimum=0,
+            # Visualization parameters
+            with gr.Group():
+                gr.Markdown("### Visualization")
+                vis_gaussians = gr.Checkbox(
+                    label="Visualize Gaussians",
+                    value=True,
+                    info="Visualize Gaussians during optimization (default: True)",
+                )
+                save_image_steps = gr.Slider(
+                    minimum=200,
                     maximum=10000,
-                    value=0,
+                    value=200,
                     step=100,
-                    label="Browse Training Steps",
-                    info="Slide to view results from different training steps (disabled during training)",
-                    interactive=False,
+                    label="Save Image Every N Steps",
+                    info="Frequency of rendering intermediate results during optimization (default: 100)",
                 )
 
-                gr.Markdown("## Final Results")
-                with gr.Row():
-                    final_render = gr.Image(
-                        label="Final Render", type="pil", height=300
-                    )
-                    final_checkpoint = gr.File(label="Download Final Checkpoint (.pt)")
+            # Control buttons
+            with gr.Row():
+                start_btn = gr.Button("Start Training", variant="primary", size="lg")
+                stop_btn = gr.Button("Stop Training", variant="stop", size="lg")
 
-                # Results buttons
-                with gr.Row():
-                    results_btn = gr.Button("Load Final Results", size="lg")
-                    enable_slider_btn = gr.Button(
-                        "Enable Step Browsing", size="lg", variant="secondary"
-                    )
+            status_text = gr.Textbox(label="Status", interactive=False, lines=2)
 
-        # Event handlers
-        start_btn.click(
-            fn=start_training_and_stream,
-            inputs=[
-                image_input,
-                exp_name,
-                num_gaussians,
-                quantize,
-                pos_bits,
-                scale_bits,
-                rot_bits,
-                feat_bits,
-                init_mode,
-                init_random_ratio,
-                max_steps,
-                vis_gaussians,
-                save_image_steps,
-                l1_loss_ratio,
-                l2_loss_ratio,
-                ssim_loss_ratio,
-                pos_lr,
-                scale_lr,
-                rot_lr,
-                feat_lr,
-                disable_lr_schedule,
-                disable_prog_optim,
-            ],
-            outputs=[
-                status_text,
-                progress_logs,
-                initialization_map,
-                current_render,
-                current_gaussian_id,
-                start_btn,
-                stop_btn,
-            ],
-        )
+        with gr.Column(scale=2):
+            gr.Markdown("## Training Progress")
 
-        stop_btn.click(fn=stop_training, outputs=status_text)
+            # Progress logs (streaming)
+            progress_logs = gr.Textbox(
+                label="Training Logs",
+                lines=10,
+                max_lines=15,
+                interactive=False,
+                autoscroll=True,
+            )
 
-        results_btn.click(
-            fn=get_final_results, outputs=[final_render, final_checkpoint]
-        )
+            # Initial map (computed at start based on initialization mode)
+            gr.Markdown("### Initialization Map")
+            initialization_map = gr.Image(
+                label="Initialization Map",
+                type="pil",
+                height=200,
+            )
 
-        enable_slider_btn.click(
-            fn=update_step_slider_after_training, outputs=[step_slider]
-        )
+            # Training images (streaming)
+            gr.Markdown("### Current Training Results")
+            with gr.Row():
+                current_render = gr.Image(
+                    label="Current Render",
+                    type="pil",
+                    height=300,
+                    show_label=True,
+                    show_download_button=True,
+                )
+                current_gaussian_id = gr.Image(
+                    label="Gaussian ID",
+                    type="pil",
+                    height=300,
+                    show_label=True,
+                    show_download_button=True,
+                )
 
-        step_slider.change(
-            fn=browse_step_results,
-            inputs=[step_slider],
-            outputs=[current_render, current_gaussian_id],
-        )
+            # Step slider for interactive browsing (will be updated dynamically)
+            step_slider = gr.Slider(
+                minimum=0,
+                maximum=10000,
+                value=0,
+                step=100,
+                label="Browse Training Steps",
+                info="Slide to view results from different training steps (disabled during training)",
+                interactive=False,
+            )
 
-    return demo
+            gr.Markdown("## Final Results")
+            with gr.Row():
+                final_render = gr.Image(label="Final Render", type="pil", height=300)
+                final_checkpoint = gr.File(label="Download Final Checkpoint (.pt)")
 
+            # Results buttons
+            with gr.Row():
+                results_btn = gr.Button("Load Final Results", size="lg")
+                enable_slider_btn = gr.Button(
+                    "Enable Step Browsing", size="lg", variant="secondary"
+                )
 
-if __name__ == "__main__":
-    # Ensure model files are available (download from HF if needed)
-    ensure_models_available()
+    # Event handlers
+    start_btn.click(
+        fn=start_training_and_stream,
+        inputs=[
+            image_input,
+            exp_name,
+            num_gaussians,
+            quantize,
+            pos_bits,
+            scale_bits,
+            rot_bits,
+            feat_bits,
+            init_mode,
+            init_random_ratio,
+            max_steps,
+            vis_gaussians,
+            save_image_steps,
+            l1_loss_ratio,
+            l2_loss_ratio,
+            ssim_loss_ratio,
+            pos_lr,
+            scale_lr,
+            rot_lr,
+            feat_lr,
+            disable_lr_schedule,
+            disable_prog_optim,
+        ],
+        outputs=[
+            status_text,
+            progress_logs,
+            initialization_map,
+            current_render,
+            current_gaussian_id,
+            start_btn,
+            stop_btn,
+        ],
+    )
+
+    stop_btn.click(fn=stop_training, outputs=status_text)
+
+    results_btn.click(fn=get_final_results, outputs=[final_render, final_checkpoint])
 
-    # Set torch hub directory
-    torch.hub.set_dir("models/torch")
+    enable_slider_btn.click(fn=update_step_slider_after_training, outputs=[step_slider])
 
-    # Create and launch the interface
-    demo = create_interface()
-    demo.launch(server_name="0.0.0.0", server_port=7860, share=False, debug=True)
+    step_slider.change(
+        fn=browse_step_results,
+        inputs=[step_slider],
+        outputs=[current_render, current_gaussian_id],
+    )
+
+
+if __name__ == "__main__":
+    demo.queue(max_size=20).launch(server_name="0.0.0.0", server_port=7860, share=False)

pyproject.docker.toml

@@ -19,6 +19,7 @@ dependencies = [
     "gsplat",
     "gradio>=4.0.0",
     "huggingface_hub>=0.24.0",
+    "spaces>=0.28.0",
 ]
 
 # We use python 3.13 and cu124 with PyTorch 2.6.0

pyproject.toml

@@ -19,6 +19,7 @@ dependencies = [
     "gsplat",
     "gradio>=4.0.0",
     "huggingface_hub>=0.24.0",
+    "spaces>=0.28.0",
 ]
 
 # We use python 3.10 and cu124

uv.lock

@@ -759,6 +759,7 @@ dependencies = [
     { name = "scikit-image" },
     { name = "scipy", version = "1.15.3", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" },
     { name = "scipy", version = "1.16.2", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11'" },
+    { name = "spaces" },
     { name = "torch", version = "2.6.0+cu124", source = { registry = "https://download.pytorch.org/whl/cu124" }, marker = "sys_platform == 'linux'" },
     { name = "torch", version = "2.8.0", source = { registry = "https://pypi.org/simple" }, marker = "sys_platform != 'linux'" },
     { name = "torchmetrics" },
@@ -784,6 +785,7 @@ requires-dist = [
     { name = "pyyaml", specifier = ">=6.0.2" },
     { name = "scikit-image", specifier = ">=0.25.2" },
     { name = "scipy", specifier = ">=1.15.3" },
+    { name = "spaces", specifier = ">=0.28.0" },
     { name = "torch", marker = "sys_platform != 'linux'", specifier = ">=2.6.0" },
     { name = "torch", marker = "sys_platform == 'linux'", specifier = ">=2.6.0", index = "https://download.pytorch.org/whl/cu124" },
     { name = "torchmetrics", specifier = ">=1.8.2" },
@@ -1723,6 +1725,20 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/84/03/0d3ce49e2505ae70cf43bc5bb3033955d2fc9f932163e84dc0779cc47f48/prompt_toolkit-3.0.52-py3-none-any.whl", hash = "sha256:9aac639a3bbd33284347de5ad8d68ecc044b91a762dc39b7c21095fcd6a19955", size = 391431, upload-time = "2025-08-27T15:23:59.498Z" },
 ]
 
+[[package]]
+name = "psutil"
+version = "5.9.8"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/90/c7/6dc0a455d111f68ee43f27793971cf03fe29b6ef972042549db29eec39a2/psutil-5.9.8.tar.gz", hash = "sha256:6be126e3225486dff286a8fb9a06246a5253f4c7c53b475ea5f5ac934e64194c", size = 503247, upload-time = "2024-01-19T20:47:09.517Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/e7/e3/07ae864a636d70a8a6f58da27cb1179192f1140d5d1da10886ade9405797/psutil-5.9.8-cp36-abi3-macosx_10_9_x86_64.whl", hash = "sha256:aee678c8720623dc456fa20659af736241f575d79429a0e5e9cf88ae0605cc81", size = 248702, upload-time = "2024-01-19T20:47:36.303Z" },
+    { url = "https://files.pythonhosted.org/packages/b3/bd/28c5f553667116b2598b9cc55908ec435cb7f77a34f2bff3e3ca765b0f78/psutil-5.9.8-cp36-abi3-manylinux_2_12_i686.manylinux2010_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:8cb6403ce6d8e047495a701dc7c5bd788add903f8986d523e3e20b98b733e421", size = 285242, upload-time = "2024-01-19T20:47:39.65Z" },
+    { url = "https://files.pythonhosted.org/packages/c5/4f/0e22aaa246f96d6ac87fe5ebb9c5a693fbe8877f537a1022527c47ca43c5/psutil-5.9.8-cp36-abi3-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d06016f7f8625a1825ba3732081d77c94589dca78b7a3fc072194851e88461a4", size = 288191, upload-time = "2024-01-19T20:47:43.078Z" },
+    { url = "https://files.pythonhosted.org/packages/6e/f5/2aa3a4acdc1e5940b59d421742356f133185667dd190b166dbcfcf5d7b43/psutil-5.9.8-cp37-abi3-win32.whl", hash = "sha256:bc56c2a1b0d15aa3eaa5a60c9f3f8e3e565303b465dbf57a1b730e7a2b9844e0", size = 251252, upload-time = "2024-01-19T20:47:52.88Z" },
+    { url = "https://files.pythonhosted.org/packages/93/52/3e39d26feae7df0aa0fd510b14012c3678b36ed068f7d78b8d8784d61f0e/psutil-5.9.8-cp37-abi3-win_amd64.whl", hash = "sha256:8db4c1b57507eef143a15a6884ca10f7c73876cdf5d51e713151c1236a0e68cf", size = 255090, upload-time = "2024-01-19T20:47:56.019Z" },
+    { url = "https://files.pythonhosted.org/packages/05/33/2d74d588408caedd065c2497bdb5ef83ce6082db01289a1e1147f6639802/psutil-5.9.8-cp38-abi3-macosx_11_0_arm64.whl", hash = "sha256:d16bbddf0693323b8c6123dd804100241da461e41d6e332fb0ba6058f630f8c8", size = 249898, upload-time = "2024-01-19T20:47:59.238Z" },
+]
+
 [[package]]
 name = "pydantic"
 version = "2.11.9"
@@ -2232,6 +2248,24 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/e9/44/75a9c9421471a6c4805dbf2356f7c181a29c1879239abab1ea2cc8f38b40/sniffio-1.3.1-py3-none-any.whl", hash = "sha256:2f6da418d1f1e0fddd844478f41680e794e6051915791a034ff65e5f100525a2", size = 10235, upload-time = "2024-02-25T23:20:01.196Z" },
 ]
 
+[[package]]
+name = "spaces"
+version = "0.42.1"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "gradio" },
+    { name = "httpx" },
+    { name = "packaging" },
+    { name = "psutil" },
+    { name = "pydantic" },
+    { name = "requests" },
+    { name = "typing-extensions" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/9a/aa/84a43df8ce2e006e742b3f568720dca6c5e7862bca5062d80b0b4b414449/spaces-0.42.1.tar.gz", hash = "sha256:cf706325dc0d22b07629463bdb0c7dec747c7f419de50f58ece1058b7b283750", size = 28260, upload-time = "2025-09-25T16:30:09.15Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/c3/16/8b16ce2bb5170b2ab332ecb6e6f010eda69cfae44cd8644f71df53adad89/spaces-0.42.1-py3-none-any.whl", hash = "sha256:cbd2b1df62a154c215ea912f88c1ad480937cf509af967c3f339116c455222d6", size = 36397, upload-time = "2025-09-25T16:30:08.052Z" },
+]
+
 [[package]]
 name = "starlette"
 version = "0.47.3"