v21.7.12

.gitignore

@@ -19,3 +19,5 @@ test/output
 test/logs
 test/*.json
 test/ft
+requirements_tmp_for_setup.txt
+0.13.3

Dockerfile

@@ -26,8 +26,8 @@ RUN python3 -m pip install wheel
 ## RUN python3 -m pip install -v -U git+https://github.com/facebookresearch/xformers.git@main#egg=xformers
 
 # Install requirements
-COPY requirements_unix.txt setup.py ./
-RUN python3 -m pip install --use-pep517 -r requirements_unix.txt xformers
+COPY requirements_linux.txt ./setup/setup.py ./
+RUN python3 -m pip install --use-pep517 -r requirements_linux.txt xformers
 
 # Replace pillow with pillow-simd
 RUN python3 -m pip uninstall -y pillow && \

README.md

@@ -71,7 +71,17 @@
 ## Installation
 
 ### Runpod
-Follow the instructions found in this discussion: https://github.com/bmaltais/kohya_ss/discussions/379
+- Select the pytorch 2.0.1 template
+- ssh into the runpod
+
+```
+cd /workspace
+git clone https://github.com/bmaltais/kohya_ss.git
+git checkout dev2
+./setup.sh -p
+```
+
+Connect to the public URL displayed
 
 ### Docker
 Docker is supported on Windows and Linux distributions. However this method currently only supports Nvidia GPUs. 
@@ -102,7 +112,11 @@
 
 venv support need to be pre-installed. Can be done on ubuntu 22.04 with `apt install python3.10-venv`
 
-Make sure to use a version of python >= 3.10.6 and < 3.11.0
+Install the cudaNN drivers following the instructions from: `https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64`
+
+Use a version of python >= 3.10.6 and < 3.11.0
+
+On WSL2, make sure to `export LD_LIBRARY_PATH=/usr/lib/wsl/lib/` or else AdamW8bit will not work.
 
 #### Setup
 
@@ -170,11 +184,6 @@
 .\setup.bat
 ```
 
-If this is a 1st install answer No when asked `Do you want to uninstall previous versions of torch and associated files before installing`.
-
-
-Then configure accelerate with the same answers as in the MacOS instructions when prompted.
-
 ### Optional: CUDNN 8.6
 
 This step is optional but can improve the learning speed for NVIDIA 30X0/40X0 owners. It allows for larger training batch size and faster training speed.
@@ -196,26 +205,20 @@
 ## Upgrading
 
 The following commands will work from the root directory of the project if you'd prefer to not run scripts.
-These commands will work on any OS.
-```bash
-git pull
-
-.\venv\Scripts\activate
-
-pip install --use-pep517 --upgrade -r requirements.txt
-```
 
 ### Windows Upgrade
 When a new release comes out, you can upgrade your repo with the following commands in the root directory:
 
 ```powershell
-upgrade.bat
+git pull
+.\setup.bat
 ```
 
 ### Linux and macOS Upgrade
 You can cd into the root directory and simply run
 
 ```bash
+git pull
 # Refresh and update everything
 ./setup.sh
 
@@ -279,13 +282,7 @@
 
 ## LoRA
 
-Training a LoRA currently uses the `train_network.py` code. You can create a LoRA network by using the all-in-one `gui.cmd` or by running the dedicated LoRA training GUI with:
-
-```
-.\venv\Scripts\activate
-
-python lora_gui.py
-```
+Training a LoRA currently uses the `train_network.py` code. You can create a LoRA network by using the all-in-one gui.
 
 Once you have created the LoRA network, you can generate images via auto1111 by installing [this extension](https://github.com/kohya-ss/sd-webui-additional-networks).
 
@@ -353,8 +350,15 @@
 
 ## Change History
 
+* 2023/06/24 (v21.7.12)
+- Significantly improved the setup process on all platforms
+- Better support for runpod
+* 2023/06/23 (v21.7.11)
+- This is a significant update to how setup work across different platform. It might be causing issues... especially for linux env like runpod. If you encounter problems please report them in the issues so I can try to address them. You can revert to the previous release with `git checkout v21.7.10`
+
+The setup solution is now much more modulat and will simplify requirements support across different environments... hoping this will make it easier to run on different OS.
 * 2023/06/19 (v21.7.10)
 - Quick fix for linux GUI startup where it would try to install darwin requirements on top of linux. Ugly fix but work. Hopefulle some linux user will improve via a PR.
 * 2023/06/18 (v21.7.9)
 - Implement temporary fix for validation of image dataset. Will no longer stop execution but will let training continue... this is changed to avoid stopping training on false positive... yet still raise awaireness that something might be wrong with the image dataset structure.
 * 2023/06/14 (v21.7.8)

gui.bat

@@ -1,18 +1,26 @@
 @echo off
+
 :: Deactivate the virtual environment
 call .\venv\Scripts\deactivate.bat
 
 :: Calling external python program to check for local modules
-python .\tools\check_local_modules.py --no_question
+python .\setup\check_local_modules.py --no_question
 
 :: Activate the virtual environment
 call .\venv\Scripts\activate.bat
 set PATH=%PATH%;%~dp0venv\Lib\site-packages\torch\lib
 
 :: Validate requirements
-python.exe .\tools\validate_requirements.py
+python.exe .\setup\validate_requirements.py
 
 :: If the exit code is 0, run the kohya_gui.py script with the command-line arguments
 if %errorlevel% equ 0 (
-    python.exe kohya_gui.py %*
-)
+    REM Check if the batch was started via double-click
+    IF /i "%comspec% /c %~0 " equ "%cmdcmdline:"=%" (
+        REM echo This script was started by double clicking.
+        cmd /k python.exe kohya_gui.py %*
+    ) ELSE (
+        REM echo This script was started from a command prompt.
+        python.exe kohya_gui.py %*
+    )
+)

gui.ps1

@@ -29,10 +29,10 @@ if ($pipOutput) {
 $env:PATH += ";$($MyInvocation.MyCommand.Path)\venv\Lib\site-packages\torch\lib"
 
 # Debug info about system
-# python.exe .\tools\debug_info.py
+# python.exe .\setup\debug_info.py
 
 # Validate the requirements and store the exit code
-python.exe .\tools\validate_requirements.py
+python.exe .\setup\validate_requirements.py
 
 # If the exit code is 0, read arguments from gui_parameters.txt (if it exists)
 # and run the kohya_gui.py script with the command-line arguments

gui.sh

@@ -1,8 +1,24 @@
 #!/usr/bin/env bash
 
+# Checks to see if variable is set and non-empty.
+# This is defined first, so we can use the function for some default variable values
+env_var_exists() {
+  if [[ -n "${!1}" ]]; then
+    return 0
+  else
+    return 1
+  fi
+}
+
+# Need RUNPOD to have a default value before first access
+RUNPOD=false
+if env_var_exists RUNPOD_POD_ID || env_var_exists RUNPOD_API_KEY; then
+  RUNPOD=true
+fi
+
 # If it is run with the sudo command, get the complete LD_LIBRARY_PATH environment variable of the system and assign it to the current environment,
 # because it will be used later.
-if [ -n "$SUDO_USER" ] || [ -n "$SUDO_COMMAND" ] ; then
+if [ -n "$SUDO_USER" ] || [ -n "$SUDO_COMMAND" ]; then
     echo "The sudo command resets the non-essential environment variables, we keep the LD_LIBRARY_PATH variable."
     export LD_LIBRARY_PATH=$(sudo -i printenv LD_LIBRARY_PATH)
 fi
@@ -11,12 +27,42 @@ fi
 SCRIPT_DIR=$(cd -- "$(dirname -- "$0")" && pwd)
 
 # Step into GUI local directory
-cd "$SCRIPT_DIR"
+cd "$SCRIPT_DIR" || exit 1
+
+if [ "$RUNPOD" = false ]; then
+    # Activate the virtual environment
+    source "$SCRIPT_DIR/venv/bin/activate" || exit 1
+fi
+
+# Check if LD_LIBRARY_PATH environment variable exists
+if [[ -z "${LD_LIBRARY_PATH}" ]]; then
+    # Set the ANSI escape sequence for yellow text
+    YELLOW='\033[0;33m'
+    # Set the ANSI escape sequence to reset text color
+    RESET='\033[0m'
+
+    echo -e "${YELLOW}Warning: LD_LIBRARY_PATH environment variable is not set.${RESET}"
+    echo -e "${YELLOW}Certain functionalities may not work correctly.${RESET}"
+    echo -e "${YELLOW}Please ensure that the required libraries are properly configured.${RESET}"
+    echo -e " "
+fi
 
-# Activate the virtual environment
-source "$SCRIPT_DIR/venv/bin/activate"
+# Determine the requirements file based on the system
+if [[ "$OSTYPE" == "darwin"* ]]; then
+    if [[ "$(uname -m)" == "arm64" ]]; then
+        REQUIREMENTS_FILE="$SCRIPT_DIR/requirements_macos_arm64.txt"
+    else
+        REQUIREMENTS_FILE="$SCRIPT_DIR/requirements_macos_amd64.txt"
+    fi
+else
+    if [ "$RUNPOD" = true ]; then
+        REQUIREMENTS_FILE="$SCRIPT_DIR/requirements_linux.txt"
+    else
+        REQUIREMENTS_FILE="$SCRIPT_DIR/requirements_runpod.txt"
+    fi
+fi
 
-# If the requirements are validated, run the kohya_gui.py script with the command-line arguments
-if python  "$SCRIPT_DIR"/tools/validate_requirements_unix.py -r "$SCRIPT_DIR"/requirements_unix.txt; then
+# Validate the requirements and run the script if successful
+if python "$SCRIPT_DIR/setup/validate_requirements.py" -r "$REQUIREMENTS_FILE"; then
     python "$SCRIPT_DIR/kohya_gui.py" "$@"
 fi

library/convert_model_gui.py

@@ -174,6 +174,10 @@ def gradio_convert_model_tab(headless=False):
         gr.Markdown(
             'This utility can be used to convert from one stable diffusion model format to another.'
         )
+
+        model_ext = gr.Textbox(value='*.safetensors *.ckpt', visible=False)
+        model_ext_name = gr.Textbox(value='Model types', visible=False)
+
         with gr.Row():
             source_model_input = gr.Textbox(
                 label='Source model',
@@ -198,7 +202,7 @@ def gradio_convert_model_tab(headless=False):
             )
             button_source_model_file.click(
                 get_file_path,
-                inputs=[source_model_input],
+                inputs=[source_model_input, model_ext, model_ext_name],
                 outputs=source_model_input,
                 show_progress=False,
             )

library/custom_logging.py

@@ -1,6 +1,7 @@
 import os
 import logging
 import time
+import sys
 
 from rich.theme import Theme
 from rich.logging import RichHandler
@@ -23,7 +24,10 @@ def setup_logging(clean=False, debug=False):
     except:
         pass
 
-    logging.basicConfig(level=logging.DEBUG, format='%(asctime)s | %(levelname)s | %(pathname)s | %(message)s', filename='setup.log', filemode='a', encoding='utf-8', force=True)
+    if sys.version_info >= (3, 9):
+        logging.basicConfig(level=logging.DEBUG, format='%(asctime)s | %(levelname)s | %(pathname)s | %(message)s', filename='setup.log', filemode='a', encoding='utf-8', force=True)
+    else:
+        logging.basicConfig(level=logging.DEBUG, format='%(asctime)s | %(levelname)s | %(pathname)s | %(message)s', filename='setup.log', filemode='a', force=True)
 
     console = Console(log_time=True, log_time_format='%H:%M:%S-%f', theme=Theme({
         "traceback.border": "black",

lora_gui.py

@@ -1293,7 +1293,15 @@ def update_LoRA_settings(LoRA_type):
                     'decompose_both': ({'LyCORIS/LoKr'}, gr.Slider),
                     'train_on_input': ({'LyCORIS/iA3'}, gr.Slider),
                     'scale_weight_norms': (
-                        {'Kohya DyLoRA', 'Kohya LoCon'},
+                        {
+                            'LoCon',
+                            'Kohya DyLoRA',
+                            'Kohya LoCon',
+                            'LyCORIS/DyLoRA',
+                            'LyCORIS/LoHa',
+                            'LyCORIS/LoCon',
+                            'LyCORIS/LoKr',
+                        },
                         gr.Slider,
                     ),
                     'network_dropout': (

pyproject.toml

@@ -0,0 +1,19 @@
+[build-system]
+requires = ["poetry-core>=1.0.0"]
+build-backend = "poetry.core.masonry.api"
+
+[tool.poetry]
+name = "library"
+version = "1.0.3"
+description = "Libraries required to run kohya_ss GUI"
+authors = ["Bernard Maltais <[email protected]>"]
+license = "Apache-2.0" # Apache Software License
+
+[[tool.poetry.source]]
+name = "library"
+path = "library"
+
+[tool.poetry.dependencies]
+python = ">=3.9,<3.11"
+
+[tool.poetry.dev-dependencies]

requirements_Ubuntu_20.04.txt → requirements.txt

@@ -1,32 +1,25 @@
-accelerate==0.19.0
 albumentations==1.3.0
 altair==4.2.2
-bitsandbytes==0.35.0
 dadaptation==3.1
 diffusers[torch]==0.10.2
 easygui==0.98.3
 einops==0.6.0
 fairscale==0.4.13
 ftfy==6.1.1
-gradio==3.23.0; sys_platform == 'darwin'
-gradio==3.32.0; sys_platform != 'darwin'
-huggingface-hub==0.13.3; sys_platform == 'darwin'
-huggingface-hub==0.13.3; sys_platform != 'darwin'
+gradio==3.33.1
+huggingface-hub>=0.13.3
 lion-pytorch==0.0.6
 lycoris_lora==0.1.6
 opencv-python==4.7.0.68
 prodigyopt==1.0
 pytorch-lightning==1.9.0
 rich==13.4.1
 safetensors==0.2.6
-tensorboard==2.10.1 ; sys_platform != 'darwin'
-tensorboard==2.12.1 ; sys_platform == 'darwin'
-tensorflow==2.10.1; sys_platform != 'darwin'
 timm==0.6.12
 tk==0.1.0
 toml==0.10.2
 transformers==4.26.0
 voluptuous==0.13.1
 wandb==0.15.0
 # for kohya_ss library
-.
+-e . # no_verify leave this to specify not checking this a verification stage

requirements_linux.txt

@@ -0,0 +1,4 @@
+torch==2.0.1+cu118 torchvision==0.15.2+cu118 --extra-index-url https://download.pytorch.org/whl/cu118 # no_verify leave this to specify not checking this a verification stage
+xformers==0.0.20 bitsandbytes==0.39.1
+accelerate==0.19.0 tensorboard==2.12.1 tensorflow==2.12.0
+-r requirements.txt

requirements_macos_amd64.txt

@@ -0,0 +1,4 @@
+torch==2.0.0 torchvision==0.15.1 -f https://download.pytorch.org/whl/cpu/torch_stable.html
+xformers bitsandbytes==0.35.0
+accelerate==0.19.0 tensorflow-macos tensorboard==2.12.1
+-r requirements.txt

requirements_macos_arm64.txt

@@ -0,0 +1,4 @@
+torch==2.0.0 torchvision==0.15.1 -f https://download.pytorch.org/whl/cpu/torch_stable.html
+xformers bitsandbytes==0.35.0
+accelerate==0.19.0 tensorflow-metal tensorboard==2.12.1
+-r requirements.txt

requirements_runpod.txt

@@ -0,0 +1,5 @@
+torch==2.0.1+cu117 torchvision==0.15.2+cu117 --extra-index-url https://download.pytorch.org/whl/cu117 # no_verify leave this to specify not checking this a verification stage
+xformers==0.0.20 bitsandbytes==0.39.1
+accelerate==0.19.0 tensorboard==2.12.1 tensorflow==2.12.0
+tensorrt
+-r requirements.txt