kaggle-minecraft-skin-generator-ipywidgetsui.ipynb

{"metadata":{"colab":{"provenance":[]},"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"name":"python","version":"3.10.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Minecraft Skin Generator Ipywidgets UI\n\nIpywidgets UIs should be accepted on Kaggle as its not a web ui and doesn't make you go out of the Kaggle experience\n\nCredits:\n- Author of the AI Model: Cory Spencer <cory@monadical.com>\n- Forked Improved Version & Ports: [Nick088](https://linktr.ee/Nick088)\n- daroche (helping me fix the 3d model texture isue)\n- [Brottweiler](https://gist.github.com/Brottweiler/483d0856c6692ef70cf90bf1a85ce364) (script to fix the 3d model texture\n- [meew](https://huggingface.co/spaces/meeww/Minecraft_Skin_Generator/blob/main/models/player_model.glb) (Minecraft Player 3d model)","metadata":{"id":"Va5kmdY9wd9n"}},{"cell_type":"code","source":"# Install\n# If you get a restart runtime warning, you don't have to, just click cancel\n!git clone https://github.com/Nick088Official/Minecraft_Skin_Generator.git\n%cd Minecraft_Skin_Generator\n!pip install -r /kaggle/working/Minecraft_Skin_Generator/Scripts/requirements_no_ui.txt\n!apt-get install -y imagemagick\nfrom IPython.display import clear_output\n\n# ipywidgets ui\n!pip install ipywidgets\nimport ipywidgets as widgets\nfrom IPython.display import display\n\nclear_output()\nprint(\"Installed!\")","metadata":{"id":"xVXg-SypqUxb","cellView":"form","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Run Inference UI\nfrom ipywidgets import Layout\n\n# style to fix too long descriptions\nstyle = {'description_width': 'initial'}\n\n# Define the widgets for each parameter\nprompt_widget = widgets.Text(\n    value='An agent',\n    description='Prompt:',\n    style=style\n)\n\nstable_diffusion_model_widget = widgets.Dropdown(\n    options=['2', 'xl'],\n    value='xl',\n    description='Stable Diffusion Model (xl understands better prompts):',\n    style=style,\n    layout=Layout(width='450px')\n)\n\nnum_inference_steps_widget = widgets.IntText(\n    value=25,\n    min=1,\n    description='Number of Inference Steps (more value = better image but longer process):',\n    orientation='horizontal',\n    readout=True,\n    readout_format='d',\n    style=style,\n    layout=Layout(width='550px')\n)\n\nguidance_scale_widget = widgets.FloatText(\n    value=7.5,\n    min=0.1,\n    description='Guidance Scale (more value = more adherent to the prompt):',\n    orientation='horizontal',\n    readout=True,\n    readout_format='.1f',\n    style=style,\n    layout=Layout(width='600px')\n)\n\nmodel_precision_type_widget = widgets.Dropdown(\n    options=['fp16', 'fp32'],\n    value='fp16',\n    description='Model Precision Type (fp16 is faster but fp32 is more precise but slower and requires paid colab gpu):',\n    style=style,\n    layout=Layout(width='700px')\n)\n\nseed_widget = widgets.IntText(\n    value=0,\n    description='Seed (A starting point to initate generation, put 0 for random):',\n    readout=True,\n    readout_format='d',\n    style=style,\n    layout=Layout(width='600px')\n)\n\nfilename_widget = widgets.Text(\n    value='output-skin.png',\n    description='Output Image Name:',\n    style=style,\n    layout=Layout(width='600px')\n)\n\nmodel_3d_widget = widgets.Checkbox(\n    value=True,\n    description=\"See the skin also as a 3D Model\",\n    indent=False,\n    style=style,\n    layout=Layout(width='600px')\n)\n\nverbose_widget = widgets.Checkbox(\n    value=False,\n    description=\"Verbose (more detailed output info)\",\n    indent=False,\n    style=style,\n    layout=Layout(width='600px')\n)\n\n# Create a function to handle the button click event\ndef run_inference(b):\n    prompt = prompt_widget.value\n    stable_diffusion_model = stable_diffusion_model_widget.value\n    num_inference_steps = num_inference_steps_widget.value\n    guidance_scale = guidance_scale_widget.value\n    model_precision_type = model_precision_type_widget.value\n    seed = seed_widget.value\n    filename = filename_widget.value\n    verbose = verbose_widget.value\n    model_3d = model_3d_widget.value\n\n    if verbose:\n        verbose_opt = '--verbose'\n    else:\n        verbose_opt = ''\n\n    if model_3d:\n        model_3d_opt = '--model_3d'\n    else:\n        model_3d_opt = ''\n\n    if stable_diffusion_model == '2':\n        sd_model = \"minecraft-skins\"\n    else:\n        sd_model = \"minecraft-skins-sdxl\"\n\n    command = f\"Scripts/{sd_model}.py '{prompt}' {num_inference_steps} {guidance_scale} {model_precision_type} {seed} {filename} {model_3d_opt} {verbose_opt}\"\n\n    !python $command\n\n\ndef display_asset(b):\n\n  filename = filename_widget.value\n\n  import matplotlib.pyplot as plt\n\n  img = plt.imread(f\"output_minecraft_skins/{filename}\")\n\n  plt.axis(\"off\")\n\n  plt.imshow(img)\n\n\n# Create a button widget and attach the click event handler\nrun = widgets.Button(\n    description=\"Run Inference (Generate Skin)\",\n    layout=Layout(width='600px'))\n\nrun.on_click(run_inference)\n\n\n# display asset output\ndisplay_asset_button = widgets.Button(\n    description=\"Display Skin Asset Output\",\n    layout=Layout(width='600px'))\n\ndisplay_asset_button.on_click(display_asset)\n\n\n# Display the widgets and the button\ndisplay(prompt_widget)\ndisplay(stable_diffusion_model_widget)\ndisplay(num_inference_steps_widget)\ndisplay(guidance_scale_widget)\ndisplay(model_precision_type_widget)\ndisplay(seed_widget)\ndisplay(filename_widget)\ndisplay(model_3d_widget)\ndisplay(verbose_widget)\ndisplay(run)\ndisplay(display_asset_button)","metadata":{"id":"OENv2tMOqr93","cellView":"form","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# HOW TO DOWNLOAD OUTPUT SKIN (ASSET & 3D MODEL)\nAfter you click Run Inference on the Ipywidgets UI, open the Console at the bottom and you should see the progress of the skin, when you see 'Successfully generated the skin!' it means it's done!\nAt your right you should find 'Output', expand the /kaggle/working folder and then the Minecraft_Skin_Generator folder, you will find the output asset image in the 'output_minecraft_skins' folder with the name you gave it, and also you can find the .glb 3d model if you made model_3d = True in the inference cell before, there will be a (the name you gave it)-converted.png too that was just made to fix a bug to load the texture on the 3d model in the project files (this bug won't happen in-game), just click the 3 dots at its right of it and click download","metadata":{}}]}