### Navigate to Installation Folder (Windows)

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README.md

Use this command in PowerShell or Command Prompt to change to the desired directory for worker installation. Avoid using spaces in the installation path.

```cmd
cd C:\path\to\install\folder
```

--------------------------------

### Install WSL and Ubuntu 24.04

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Installs the Windows Subsystem for Linux and the Ubuntu 24.04 distribution. This command may require running 'wsl --install' first if WSL is not already enabled.

```bash
wsl --install -d Ubuntu-24.04
```

--------------------------------

### Enable DXG Detection and Verify ROCm Setup

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Commands to set environment variables for DXG detection and ROCm path, then reload the bash profile. Finally, run rocminfo to verify the setup.

```bash
# 1. Add the environment variables to your bash profile
echo 'export HSA_ENABLE_DXG_DETECTION=1' >> ~/.bashrc
echo 'export PATH=/opt/rocm/bin:$PATH' >> ~/.bashrc

# 2. Reload your profile
source ~/.bashrc
```

```bash
rocminfo
```

--------------------------------

### rocminfo Output Example

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Example output from the rocminfo command, indicating a successful WSL environment detection and HSA system attributes.

```text
WSL environment detected.
=====================
HSA System Attributes
=====================
Runtime Version:         1.1
Runtime Ext Version:     1.6
System Timestamp Freq.:  1000.000000MHz
Sig. Max Wait Duration:  18446744073709551615 (0xFFFFFFFFFFFFFFFF) (timestamp count)
Machine Model:           LARGE
System Endianness:       LITTLE
Mwaitx:                  DISABLED
DMAbuf Support:          NO

==========
HSA Agents
==========
...
```

--------------------------------

### Start or Monitor Running Container

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Start a container named 'reGen' and attach to its output, or monitor a running container. CTRL+C detaches the container, leaving it running in the background.

```bash
docker start -ai reGen
```

--------------------------------

### Setup Python Virtual Environment

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Create and activate a Python virtual environment for the worker. This isolates dependencies and ensures consistent behavior.

```bash
python -m venv regen
```

```bash
py -3.11 -m venv regen
```

```bash
regen\Scripts\activate.bat
```

```bash
regen\Scripts\Activate.ps1
```

```bash
source regen/bin/activate
```

--------------------------------

### Clone and Build librocdxg Library

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Steps to clone the librocdxg repository, set the Windows SDK path, configure the build environment using CMake, and compile/install the library. Ensure the WIN_SDK path matches your installed Windows SDK version.

```bash
# 1. Clone the repository to your local WSL
git clone https://github.com/ROCm/librocdxg.git
cd librocdxg

# 2. Set the Windows SDK path 
# IMPORTANT: Change "10.0.26100.0" below to match the folder version you found in Step 1!
export win_sdk='/mnt/c/Program Files (x86)/Windows Kits/10/Include/10.0.26100.0/'
 
# 3. Configure the build environment
mkdir -p build
cd build
cmake .. -DWIN_SDK="${win_sdk}/shared"

# 4. Compile and install the library
make
sudo make install
```

--------------------------------

### Start a Container Detached

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Start a container named 'reGen' in detached mode, allowing it to run in the background.

```bash
docker start reGen
```

--------------------------------

### Install Ubuntu 24.04 Build Dependencies

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Installs essential build tools, CMake, Git, and SSL certificates within the Ubuntu 24.04 WSL environment. These are necessary for compiling code and managing repositories.

```bash
sudo apt update && sudo apt install -y build-essential cmake git ca-certificates wget gpg
```

--------------------------------

### Install Dependencies for RoCM

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Install the required Python packages for the worker, specifically for RoCM-enabled systems. This command should be run after cloning the repository and activating the virtual environment.

```bash
pip install -r requirements.txt --extra-index-url https://download.pytorch.org/whl/rocm6.2
```

--------------------------------

### Install Dependencies for CUDA

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Install the required Python packages for the worker, specifically for CUDA-enabled systems. This command should be run after cloning the repository and activating the virtual environment.

```bash
pip install -r requirements.txt --extra-index-url https://download.pytorch.org/whl/cu128
```

--------------------------------

### Build and Run with Docker Compose

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Build the Docker image and start the horde-worker-reGen container using Docker Compose. Replace '[cuda|rocm]' with your specific GPU type (cuda or rocm). Ensure bridgeData.yaml is configured correctly.

```bash
docker compose -f Dockerfiles/compose.[cuda|rocm].yaml build --pull
docker compose -f Dockerfiles/compose.[cuda|rocm].yaml up -dV
```

--------------------------------

### Configure ROCm Repository for Ubuntu 24.04

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

This multi-step process adds the AMD ROCm GPG key, configures the ROCm repository for Ubuntu 24.04 (Noble), and sets an APT priority pin to ensure AMD's packages are preferred. Finally, it updates the package list and installs the ROCm stack.

```bash
# 1. Download and add the AMD GPG key
sudo mkdir --parents --mode=0755 /etc/apt/keyrings
wget https://repo.radeon.com/rocm/rocm.gpg.key -O - | gpg --dearmor | sudo tee /etc/apt/keyrings/rocm.gpg > /dev/null

# 2. Add the ROCm repository specifically for Ubuntu 24.04 (Noble)
sudo tee /etc/apt/sources.list.d/rocm.list << EOF
deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/rocm/apt/latest noble main
EOF

# 3. Set the Apt priority pin to prefer AMD's repo over Ubuntu's defaults
echo -e "Package: *
Pin: release o=repo.radeon.com
Pin-Priority: 600" | sudo tee /etc/apt/preferences.d/rocm-pin-600

# 4. Update your package list and install the ROCm stack
sudo apt update
sudo apt install -y rocm rocm-dev
```

--------------------------------

### Run the Worker

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Start the Horde Worker Regen process. Ensure your configuration file (bridgeData.yaml) is set up with an API key and worker name.

```bash
python run_worker.py
```

--------------------------------

### Clone Horde Worker Repository (Windows/Linux)

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README.md

Clone the Horde worker repository using Git. This is the recommended method for installation on both Windows and Linux systems.

```git
git clone https://github.com/Haidra-Org/horde-worker-reGen.git
cd horde-worker-reGen
```

--------------------------------

### Clone Horde-Worker-ReGen Repository

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Clone the specific directory for horde-worker-reGen using sparse checkout. This is the initial step to get the necessary Dockerfiles and configuration templates.

```bash
git clone --sparse https://github.com/Haidra-Org/horde-worker-reGen.git
cd horde-worker-reGen
git sparse-checkout set --no-cone Dockerfiles /bridgeData_template.yaml
```

--------------------------------

### Start Multiple GPU Workers on Linux

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README.md

When using multiple GPUs on Linux, start a separate worker instance for each GPU by specifying the CUDA_VISIBLE_DEVICES environment variable. Ensure sufficient RAM is available, as queue_size and max_threads impact memory usage.

```bash
CUDA_VISIBLE_DEVICES=0 ./horde-bridge.sh -n "Instance 1"
CUDA_VISIBLE_DEVICES=1 ./horde-bridge.sh -n "Instance 2"
```

--------------------------------

### Download Models

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Execute the script to download necessary models. This is a critical first step before running the worker.

```bash
python download_models.py
```

--------------------------------

### Build Arguments for Working with Git Forks

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Configure Git owner and branch to easily test changes from forks or personal branches.

```bash
--build-arg GIT_OWNER=your-github-username \
--build-arg GIT_BRANCH=your-feature-branch
```

--------------------------------

### Build Arguments for CI/CD Git Integration

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Automate Git branch and owner configuration in CI/CD pipelines using environment variables.

```bash
--build-arg GIT_BRANCH=${CI_COMMIT_BRANCH} \
--build-arg GIT_OWNER=${CI_PROJECT_NAMESPACE}
```

--------------------------------

### Build NVIDIA (CUDA) Docker Image

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Builds the Docker image for NVIDIA GPUs using Dockerfile.cuda. Customize with build arguments like CUDA_VERSION and PYTHON_VERSION.

```bash
docker build -f Dockerfile.cuda \
  --build-arg CUDA_VERSION=12.4.1 \
  --build-arg PYTHON_VERSION=3.11 \
  --build-arg GIT_BRANCH=main \
  --build-arg GIT_OWNER=Haidra-Org \
  --build-arg USE_PIP_CACHE=true \
  -t horde-worker-regen:cuda .
```

--------------------------------

### System RAM Configuration

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README.md

Settings for systems with less than 32GB of System RAM. Focuses on using SD15 models and managing memory.

```yaml
- load_large_models: false
```

--------------------------------

### Build AMD (ROCm) Docker Image

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Builds the Docker image for AMD GPUs using Dockerfile.rocm. Customize with build arguments like ROCM_VERSION and PYTHON_VERSION.

```bash
docker build -f Dockerfile.rocm \
  --build-arg ROCM_VERSION=6.1.2 \
  --build-arg PYTHON_VERSION=3.11 \
  --build-arg GIT_BRANCH=main \
  --build-arg GIT_OWNER=Haidra-Org \
  --build-arg USE_PIP_CACHE=true \
  -t horde-worker-regen:rocm .
```

--------------------------------

### Generate .env file from bridgeData.yaml (Windows)

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Use the `convert_config_to_env.py` script with `runtime.cmd` to convert a `bridgeData.yaml` file into a `.env` file suitable for Docker.

```batch
.\runtime.cmd python -s -m convert_config_to_env --file .\bridgeData.yaml
```

--------------------------------

### Generate .env file from bridgeData.yaml (Linux)

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Use the `convert_config_to_env.py` script with `runtime.sh` to convert a `bridgeData.yaml` file into a `.env` file suitable for Docker.

```shell
./runtime.sh python -s -m convert_config_to_env --file ./bridgeData.yaml
```

--------------------------------

### Generate .env file from bridgeData.yaml (venv)

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Execute the `convert_config_to_env.py` script directly to convert a `bridgeData.yaml` file into a `.env` file. This method is suitable for users within a virtual environment.

```python
python -m convert_config_to_env --file ./bridgeData.yaml
```

--------------------------------

### Build Argument for Pull Request Review

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Allows reviewers to build and test changes from a contributor's specific branch and username.

```bash
--build-arg GIT_OWNER=contributor-username \
--build-arg GIT_BRANCH=feature-branch
```

--------------------------------

### Run NVIDIA (CUDA) Container

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Runs the Horde Worker Regen container with CUDA support, enabling all available GPUs.

```bash
docker run -it --gpus all horde-worker-regen:cuda
```

--------------------------------

### 24GB+ VRAM Worker Configuration

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README.md

Recommended settings for GPUs with 24GB or more VRAM (e.g., 3090, 4090). Adjust max_threads and related settings for optimal performance.

```yaml
- queue_size: 1 # <32GB RAM: 0, 32GB: 1, >32GB: 2
- safety_on_gpu: true
- high_memory_mode: true
- high_performance_mode: true
- unload_models_from_vram_often: false
- max_threads: 1 # 2 is often viable for xx90 cards
- post_process_job_overlap: true
- queue_size: 2 # Set to 1 if max_threads: 2
- max_power: 64 # Reduce if max_threads: 2
- max_batch: 8 # Increase if max_threads: 1, decrease if max_threads: 2
- allow_sdxl_controlnet: true
```

--------------------------------

### Rebuild Docker Image for Worker Update

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Rebuild the Docker image using the appropriate Dockerfile for CUDA or AMD, and tag it accordingly.

```bash
docker build -f Dockerfile.[cuda|amd] -t horde-worker-regen:[cuda|rocm] .
```

--------------------------------

### Run AMD (ROCm) Container

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Runs the Horde Worker Regen container with ROCm support, mapping necessary devices and group for AMD GPU access.

```bash
docker run -it --device=/dev/kfd --device=/dev/dri --group-add video horde-worker-regen:rocm
```

--------------------------------

### Unregister Existing Ubuntu 24.04 WSL Image

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Resets the Ubuntu 24.04 WSL image to a clean state. This action will delete all data within the WSL image, so ensure any important data is backed up beforehand.

```bash
wsl --unregister Ubuntu-24.04
```

--------------------------------

### Clone Horde Worker Repository

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Clone the worker repository to your local system using Git.

```bash
git clone https://github.com/Haidra-Org/horde-worker-reGen.git
```

--------------------------------

### Lower-end GPU Worker Settings

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README.md

Options for lower-end GPUs or under-performing workers. `extra_slow_worker` requires user opt-in and is for very low MPS/S or kudos/hr.

```yaml
- extra_slow_worker: true
- limit_max_steps: true
- preload_timeout: 120
```

--------------------------------

### Build Argument to Disable Pip Cache

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Use this argument to troubleshoot pip cache-related errors by building the image without using the cache.

```bash
--build-arg USE_PIP_CACHE=false
```

--------------------------------

### Supported Stable Diffusion Baselines

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README.md

List of Stable Diffusion baseline categories currently supported for custom models. Flux.schnell series models are permitted, but Flux.dev models are not.

```text
stable_diffusion_1
stable_diffusion_2_768
stable_diffusion_2_512
stable_diffusion_xl
stable_cascade
flux_1
```

--------------------------------

### Set Horde Worker Configuration via Environment Variables

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Configure the Horde worker by setting environment variables. Prepend 'AIWORKER_' to any option found in `bridgeData_template.yaml`. Ensure sensitive information like API keys and worker names are updated.

```shell
AIWORKER_API_KEY=your_api_key_here          # Important
AIWORKER_CACHE_HOME=/workspace/models       # Important
AIWORKER_DREAMER_NAME=your_worker_name_here # Important
AIWORKER_ALLOW_CONTROLNET=True
AIWORKER_ALLOW_LORA=True
AIWORKER_MAX_LORA_CACHE_SIZE=50
AIWORKER_ALLOW_PAINTING=True
AIWORKER_MAX_POWER=38
AIWORKER_MAX_THREADS=1 # Only set to 2 on high end or xx90 machines
AIWORKER_MODELS_TO_LOAD=['TOP 3', 'AlbedoBase XL (SDXL)'] # Be mindful of download times; each model average 2-8 gb
AIWORKER_MODELS_TO_SKIP=['pix2pix', 'SDXL_beta::stability.ai#6901']
AIWORKER_QUEUE_SIZE=2
AIWORKER_MAX_BATCH=4
AIWORKER_SAFETY_ON_GPU=True
AIWORKER_CIVITAI_API_TOKEN=your_token_here
```

--------------------------------

### 8-10GB VRAM Worker Configuration

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README.md

Configuration for GPUs with 8-10GB VRAM (e.g., 2080, 3060, 4060). Disable post-processing and SDXL ControlNet if not needed.

```yaml
- queue_size: 1 # <32GB RAM: 0, 32GB: 1, >32GB: 2
- safety_on_gpu: false
- max_threads: 1
- max_power: 32 # No higher
- max_batch: 4 # No higher
- allow_post_processing: false # If using SDXL/Flux, else can be true
- allow_sdxl_controlnet: false
```

--------------------------------

### Build Argument for Testing Specific Git Branches

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Use this build argument to test new features from a specific Git branch before merging it into the main branch.

```bash
--build-arg GIT_BRANCH=feature-new-model
```

--------------------------------

### Reference Custom Model in bridgeData.yaml

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README.md

Define custom models by specifying their name, baseline, and local filepath in the bridgeData.yaml configuration file. Ensure the model name does not conflict with existing models.

```yaml
custom_models:
  - name: My Custom Model
    baseline: stable_diffusion_xl
    filepath: /path/to/model/file.safetensors
```

--------------------------------

### 12-16GB VRAM Worker Configuration

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README.md

Settings suitable for GPUs with 12-16GB VRAM (e.g., 3080 Ti, 4070 Ti, 4080). Consider disabling safety_on_gpu for Cascade/Flux.

```yaml
- queue_size: 1 # <32GB RAM: 0, 32GB: 1, >32GB: 2
- safety_on_gpu: true # Consider false if using Cascade/Flux
- moderate_performance_mode: true
- unload_models_from_vram_often: false
- max_threads: 1
- max_power: 50
- max_batch: 4 # Or higher
```

--------------------------------

### Update Dependencies with CUDA

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Manually update Python dependencies, including PyTorch for CUDA, after pulling changes from the repository. This ensures you have the latest required packages.

```bash
python -m pip install -r requirements.txt -U https://download.pytorch.org/whl/cu128
```

--------------------------------

### Update Dependencies with RoCM

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/README_advanced.md

Manually update Python dependencies, including PyTorch for RoCM, after pulling changes from the repository. This ensures you have the latest required packages.

```bash
python -m pip install -r requirements.txt -U https://download.pytorch.org/whl/rocm6.2
```

--------------------------------

### Update Worker with Docker Compose

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Update the horde-worker-reGen application by pulling the latest changes, rebuilding the Docker image, and recreating the container using Docker Compose. This process ensures you are running the most recent version.

```bash
cd horde-worker-reGen
git pull
docker compose -f Dockerfiles/compose.[cuda|rocm].yaml build --pull
docker compose -f Dockerfiles/compose.[cuda|rocm].yaml up -dV
```

--------------------------------

### Stop a Running Container

Source: https://github.com/haidra-org/horde-worker-regen/blob/main/Dockerfiles/README.md

Stop a running horde-worker-reGen container named 'reGen'. It is recommended to set the worker to maintenance mode via the API before stopping to avoid job loss.

```bash
docker stop reGen
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.