{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"provenance":[{"file_id":"https://huggingface.co/codeShare/FLUX.2-klein-9b-SDNQ-4bit/blob/main/colab_notebooks/utils/bria_rembg_background_removal.ipynb","timestamp":1780096938720}],"gpuType":"T4","mount_file_id":"11YEDI8zKDFOc43YcDo0oO4tnBOp7ZRkv","authorship_tag":"ABX9TyN2RuFLR9nriHUbuHeis7jw"},"kernelspec":{"name":"python3","display_name":"Python 3"},"language_info":{"name":"python"},"accelerator":"GPU"},"cells":[{"cell_type":"markdown","metadata":{"id":"54eb8180"},"source":["### Hugging Face Login\n","\n","To access gated models like `briaai/RMBG-2.0`, you need to log in to Hugging Face. Make sure you have your Hugging Face API token stored in Colab secrets as `HF_TOKEN`. You can create a token [here](https://huggingface.co/settings/tokens)."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"69c32694","executionInfo":{"status":"ok","timestamp":1780096900756,"user_tz":-120,"elapsed":1063,"user":{"displayName":"","userId":""}},"outputId":"68bd6361-361a-429c-a6f6-27f4665f8391"},"source":["from google.colab import userdata\n","from huggingface_hub import login\n","\n","# Retrieve HF_TOKEN from Colab secrets\n","HF_TOKEN = userdata.get('HF_TOKEN')\n","\n","if HF_TOKEN:\n","    login(token=HF_TOKEN)\n","    print(\"Successfully logged into Hugging Face.\")\n","else:\n","    print(\"HF_TOKEN not found in Colab secrets. Please add it to proceed with gated models.\")"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Successfully logged into Hugging Face.\n"]}]},{"cell_type":"code","source":["!pip install -q \\\n","    transformers \\\n","    accelerate \\\n","    pillow \\\n","    tqdm \\\n","    safetensors \\\n","    kornia"],"metadata":{"id":"Y3HzjfqPsP45","executionInfo":{"status":"ok","timestamp":1780095890183,"user_tz":-120,"elapsed":15178,"user":{"displayName":"","userId":""}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"5b92f137-fcee-42d8-ae14-ca2c39726e87"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["\u001b[?25l   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m0.0/1.2 MB\u001b[0m \u001b[31m?\u001b[0m eta \u001b[36m-:--:--\u001b[0m\r\u001b[2K   \u001b[91m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[91m╸\u001b[0m \u001b[32m1.2/1.2 MB\u001b[0m \u001b[31m76.0 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\r\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.2/1.2 MB\u001b[0m \u001b[31m22.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n","\u001b[?25h\u001b[?25l   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m0.0/3.7 MB\u001b[0m \u001b[31m?\u001b[0m eta \u001b[36m-:--:--\u001b[0m\r\u001b[2K   \u001b[91m━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[90m╺\u001b[0m\u001b[90m━━━━━━━━━━━━\u001b[0m \u001b[32m2.5/3.7 MB\u001b[0m \u001b[31m72.1 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\r\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m3.7/3.7 MB\u001b[0m \u001b[31m53.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n","\u001b[?25h"]}]},{"cell_type":"code","source":["from google.colab import drive\n","drive.mount('/content/drive')"],"metadata":{"id":"WHDPLITy4Rh0"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["Note:  Bria AI remove background tool is a gated access model , meaning you will need a huggingface account , and you\n","will need to fetch your huggingface api token , the  agree to open source terms of usage on https://huggingface.co/briaai/RMBG-2.0  while logged into your huggingface account"],"metadata":{"id":"TT87LNT9uAWR"}},{"cell_type":"code","source":["# =============================================================================\n","# BRIA RMBG 2.0 ZIP Background Removal\n","# =============================================================================\n","# Features:\n","# - Google Drive ZIP input\n","# - Recursive image discovery\n","# - Removes __MACOSX metadata\n","# - Removes .DS_Store files\n","# - Skips AppleDouble files (._filename.jpg)\n","# - Preserves folder structure\n","# - Transparent PNG output\n","# - Custom output ZIP filename\n","# - GPU acceleration\n","# =============================================================================\n","\n","# Install dependencies first:\n","# !pip install -q transformers accelerate pillow tqdm safetensors\n","\n","import zipfile\n","import shutil\n","from pathlib import Path\n","\n","import torch\n","import numpy as np\n","\n","from PIL import Image\n","from tqdm import tqdm\n","\n","from transformers import AutoModelForImageSegmentation\n","\n","# =============================================================================\n","# SETTINGS\n","# =============================================================================\n","\n","zip_filepath = \"/content/drive/MyDrive/images.zip\"  #@param {type:\"string\"}\n","\n","output_zip_filename = \"transparent_images.zip\"  #@param {type:\"string\"}\n","\n","# =============================================================================\n","# MOUNT GOOGLE DRIVE\n","# =============================================================================\n","\n","from google.colab import drive\n","\n","drive.mount(\"/content/drive\")\n","\n","# =============================================================================\n","# VALIDATE INPUT ZIP\n","# =============================================================================\n","\n","zip_filepath = Path(zip_filepath)\n","\n","if not zip_filepath.exists():\n","    raise FileNotFoundError(\n","        f\"ZIP file not found:\\n{zip_filepath}\"\n","    )\n","\n","# =============================================================================\n","# PREPARE OUTPUT ZIP NAME\n","# =============================================================================\n","\n","output_zip_filename = output_zip_filename.strip()\n","\n","if not output_zip_filename:\n","    output_zip_filename = \"transparent_images.zip\"\n","\n","if not output_zip_filename.lower().endswith(\".zip\"):\n","    output_zip_filename += \".zip\"\n","\n","# =============================================================================\n","# WORKING DIRECTORIES\n","# =============================================================================\n","\n","base_dir = Path(\"/content/bria_rmbg_work\")\n","\n","if base_dir.exists():\n","    shutil.rmtree(base_dir)\n","\n","input_dir = base_dir / \"input\"\n","output_dir = base_dir / \"output\"\n","\n","input_dir.mkdir(parents=True, exist_ok=True)\n","output_dir.mkdir(parents=True, exist_ok=True)\n","\n","# =============================================================================\n","# EXTRACT ZIP\n","# =============================================================================\n","\n","print(\"Extracting ZIP...\")\n","\n","with zipfile.ZipFile(zip_filepath, \"r\") as z:\n","    z.extractall(input_dir)\n","\n","print(\"ZIP extracted\")\n","\n","# =============================================================================\n","# REMOVE MACOS METADATA\n","# =============================================================================\n","\n","print(\"Removing macOS metadata...\")\n","\n","for macos_dir in input_dir.rglob(\"__MACOSX\"):\n","\n","    if macos_dir.is_dir():\n","\n","        try:\n","            shutil.rmtree(\n","                macos_dir,\n","                ignore_errors=True\n","            )\n","        except:\n","            pass\n","\n","for ds_store in input_dir.rglob(\".DS_Store\"):\n","\n","    try:\n","        ds_store.unlink()\n","    except:\n","        pass\n","\n","# =============================================================================\n","# FIND IMAGES\n","# =============================================================================\n","\n","IMAGE_EXTENSIONS = {\n","    \".jpg\",\n","    \".jpeg\",\n","    \".png\",\n","    \".webp\",\n","    \".bmp\",\n","    \".tif\",\n","    \".tiff\"\n","}\n","\n","image_files = []\n","\n","for p in input_dir.rglob(\"*\"):\n","\n","    if not p.is_file():\n","        continue\n","\n","    if \"__MACOSX\" in p.parts:\n","        continue\n","\n","    if p.name.startswith(\"._\"):\n","        continue\n","\n","    if p.name == \".DS_Store\":\n","        continue\n","\n","    if p.suffix.lower() not in IMAGE_EXTENSIONS:\n","        continue\n","\n","    image_files.append(p)\n","\n","print(f\"Found {len(image_files)} image(s)\")\n","\n","if len(image_files) == 0:\n","    raise RuntimeError(\n","        \"No valid image files found in ZIP.\"\n","    )\n","\n","# =============================================================================\n","# DEVICE\n","# =============================================================================\n","\n","device = (\n","    \"cuda\"\n","    if torch.cuda.is_available()\n","    else \"cpu\"\n",")\n","\n","print(f\"Using device: {device}\")\n","\n","# =============================================================================\n","# LOAD BRIA RMBG MODEL\n","# =============================================================================\n","\n","print(\"Loading BRIA RMBG 2.0...\")\n","\n","model = AutoModelForImageSegmentation.from_pretrained(\n","    \"briaai/RMBG-2.0\",\n","    trust_remote_code=True\n",")\n","\n","model.to(device)\n","model.eval()\n","\n","print(\"Model loaded\")\n","\n","# =============================================================================\n","# IMAGE HELPERS\n","# =============================================================================\n","\n","MODEL_SIZE = 1024\n","\n","def preprocess_image(image):\n","\n","    image = image.convert(\"RGB\")\n","\n","    original_size = image.size\n","\n","    resized = image.resize(\n","        (MODEL_SIZE, MODEL_SIZE),\n","        Image.Resampling.LANCZOS\n","    )\n","\n","    image_np = (\n","        np.array(resized)\n","        .astype(np.float32)\n","        / 255.0\n","    )\n","\n","    image_np = image_np.transpose(\n","        2,\n","        0,\n","        1\n","    )\n","\n","    tensor = torch.from_numpy(\n","        image_np\n","    ).unsqueeze(0)\n","\n","    return tensor, original_size\n","\n","\n","def postprocess_mask(\n","    mask,\n","    original_size\n","):\n","\n","    mask = mask.squeeze()\n","\n","    mask = mask.cpu().numpy()\n","\n","    mask = (\n","        mask * 255\n","    ).astype(np.uint8)\n","\n","    mask_img = Image.fromarray(mask)\n","\n","    mask_img = mask_img.resize(\n","        original_size,\n","        Image.Resampling.LANCZOS\n","    )\n","\n","    return mask_img\n","\n","# =============================================================================\n","# PROCESS IMAGES\n","# =============================================================================\n","\n","print(\"Removing backgrounds...\")\n","\n","for image_path in tqdm(image_files):\n","\n","    try:\n","\n","        image = Image.open(\n","            image_path\n","        ).convert(\"RGB\")\n","\n","        tensor, original_size = (\n","            preprocess_image(image)\n","        )\n","\n","        tensor = tensor.to(device)\n","\n","        with torch.no_grad():\n","\n","            prediction = model(tensor)\n","\n","            if isinstance(\n","                prediction,\n","                (tuple, list)\n","            ):\n","                prediction = prediction[-1]\n","\n","            prediction = torch.sigmoid(\n","                prediction\n","            )\n","\n","        mask = postprocess_mask(\n","            prediction[0][0],\n","            original_size\n","        )\n","\n","        rgba = image.convert(\"RGBA\")\n","\n","        rgba.putalpha(mask)\n","\n","        relative_path = (\n","            image_path.relative_to(\n","                input_dir\n","            )\n","        )\n","\n","        output_path = (\n","            output_dir /\n","            relative_path.with_suffix(\n","                \".png\"\n","            )\n","        )\n","\n","        output_path.parent.mkdir(\n","            parents=True,\n","            exist_ok=True\n","        )\n","\n","        rgba.save(\n","            output_path,\n","            \"PNG\",\n","            optimize=True\n","        )\n","\n","    except Exception as e:\n","\n","        print()\n","        print(\n","            f\"Failed: {image_path.name}\"\n","        )\n","        print(e)\n","\n","# =============================================================================\n","# CREATE ZIP\n","# =============================================================================\n","\n","print(\"Creating output ZIP...\")\n","\n","temp_zip_base = Path(\n","    \"/content/bria_rmbg_output\"\n",")\n","\n","if temp_zip_base.with_suffix(\n","    \".zip\"\n",").exists():\n","    temp_zip_base.with_suffix(\n","        \".zip\"\n","    ).unlink()\n","\n","shutil.make_archive(\n","    str(temp_zip_base),\n","    \"zip\",\n","    output_dir\n",")\n","\n","# =============================================================================\n","# SAVE TO GOOGLE DRIVE\n","# =============================================================================\n","\n","final_zip_path = (\n","    Path(\"/content/drive/MyDrive\")\n","    / output_zip_filename\n",")\n","\n","if final_zip_path.exists():\n","    final_zip_path.unlink()\n","\n","shutil.move(\n","    str(\n","        temp_zip_base.with_suffix(\n","            \".zip\"\n","        )\n","    ),\n","    str(final_zip_path)\n",")\n","\n","# =============================================================================\n","# SUMMARY\n","# =============================================================================\n","\n","print()\n","print(\"=\" * 60)\n","print(\"PROCESSING COMPLETE\")\n","print(\"=\" * 60)\n","print(f\"Images processed : {len(image_files)}\")\n","print(f\"Output ZIP       : {final_zip_path}\")\n","print(\"=\" * 60)"],"metadata":{"id":"HOMl19GZxBmh"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"6973ff74"},"source":["### Disconnect Runtime"]},{"cell_type":"code","metadata":{"id":"75027cc1"},"source":["#@title Auto Disconnect Runtime upon code completion\n","disconnect_runtime = True #@param {type:'boolean'}\n","\n","if disconnect_runtime:\n","  from google.colab import runtime\n","  runtime.unassign()"],"execution_count":null,"outputs":[]}]}