training scripts

training-scripts/musubi/ernie/dataset.currentset.toml

@@ -0,0 +1,15 @@
+# Image-only ERNIE-Image dataset config for staged wrapper runs.
+# The wrapper copies a selected source dataset into datasets/currentset,
+# creates empty .txt sidecars there, then uses this config for caching/training.
+
+[general]
+resolution = [512, 512]
+caption_extension = ".txt"
+batch_size = 1
+enable_bucket = true
+bucket_no_upscale = false
+
+[[datasets]]
+image_directory = "C:/Development/trainers/musubi-tuner/datasets/currentset"
+cache_directory = "C:/Development/trainers/musubi-tuner/datasets/cache"
+num_repeats = 1

training-scripts/musubi/ernie/prepare_dataset_ernie.bat

@@ -0,0 +1,117 @@
+@echo off
+setlocal EnableExtensions
+
+cd /d "%~dp0\..\.."
+
+if "%~1"=="" (
+  echo Usage: %~nx0 ^<dataset_name^>
+  echo Example: %~nx0 aubreyplaza
+  exit /b 1
+)
+
+set "DATASET_NAME=%~1"
+set "SOURCE_DATASET_DIR=C:\Development\ai-toolkit\datasets\%DATASET_NAME%"
+set "CURRENTSET_DIR=C:\Development\trainers\musubi-tuner\datasets\currentset"
+set "CACHE_DIR=C:\Development\trainers\musubi-tuner\datasets\cache"
+set "STATUS_FILE=C:\Development\trainers\musubi-tuner\datasets\ernie_%DATASET_NAME%.txt"
+set "DATASET_CONFIG=C:\Development\trainers\musubi-tuner\templates\ernie\dataset.currentset.toml"
+set "PYTHON_EXE=C:\Development\trainers\musubi-tuner\venv\Scripts\python.exe"
+set "ERNIE_VAE=C:/Development/ComfyUI/models/vae/flux2_vae.safetensors"
+set "ERNIE_TEXT_ENCODER=C:/Development/ComfyUI/models/text_encoders/ministral-3-3b.safetensors"
+set "COUNT_FILE=%TEMP%\musubi_ernie_image_count.txt"
+set "EXIT_CODE=0"
+set "IMAGE_COUNT=0"
+
+if not exist "%SOURCE_DATASET_DIR%" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Dataset directory not found: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+if not exist "%PYTHON_EXE%" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Python executable not found: %PYTHON_EXE%
+  exit /b 1
+)
+
+"%PYTHON_EXE%" -c "from pathlib import Path; exts={'.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl'}; print(sum(1 for f in Path(r'%SOURCE_DATASET_DIR%').iterdir() if f.is_file() and f.suffix.lower() in exts))" > "%COUNT_FILE%"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to count supported image files in: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+set /p IMAGE_COUNT=<"%COUNT_FILE%"
+del /q "%COUNT_FILE%" >nul 2>nul
+
+if "%IMAGE_COUNT%"=="0" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] No supported image files found in: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+> "%STATUS_FILE%" echo starting
+
+echo [INFO] Dataset name   : %DATASET_NAME%
+echo [INFO] Source dataset : %SOURCE_DATASET_DIR%
+echo [INFO] Staging dir    : %CURRENTSET_DIR%
+echo [INFO] Cache dir      : %CACHE_DIR%
+echo [INFO] Status file    : %STATUS_FILE%
+echo [INFO] Dataset config : %DATASET_CONFIG%
+echo [INFO] Image count    : %IMAGE_COUNT%
+
+if not exist "%CURRENTSET_DIR%" mkdir "%CURRENTSET_DIR%"
+if not exist "%CACHE_DIR%" mkdir "%CACHE_DIR%"
+
+echo [INFO] Clearing staging and cache directories...
+powershell -NoProfile -Command "Get-ChildItem -LiteralPath '%CURRENTSET_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue; Get-ChildItem -LiteralPath '%CACHE_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to clear staging or cache directories.
+  exit /b 1
+)
+
+echo [INFO] Copying dataset images to currentset...
+powershell -NoProfile -Command "Copy-Item -LiteralPath (Get-ChildItem -LiteralPath '%SOURCE_DATASET_DIR%' -File | Where-Object { $_.Extension -in '.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl' } | Select-Object -ExpandProperty FullName) -Destination '%CURRENTSET_DIR%' -Force"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to copy dataset images into currentset.
+  exit /b 1
+)
+
+echo [INFO] Creating empty caption files...
+powershell -NoProfile -Command "Get-ChildItem -LiteralPath '%CURRENTSET_DIR%' -File | Where-Object { $_.Extension -in '.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl' } | ForEach-Object { $caption = [System.IO.Path]::ChangeExtension($_.FullName, '.txt'); if (-not (Test-Path -LiteralPath $caption)) { New-Item -ItemType File -Path $caption | Out-Null } }"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to create empty caption files in currentset.
+  exit /b 1
+)
+
+echo [INFO] Running latent caching...
+"%PYTHON_EXE%" src\musubi_tuner\ernie_image_cache_latents.py ^
+  --dataset_config "%DATASET_CONFIG%" ^
+  --vae "%ERNIE_VAE%" ^
+  --device cuda ^
+  --skip_existing
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Latent caching failed.
+  exit /b 1
+)
+
+echo [INFO] Running text encoder caching...
+"%PYTHON_EXE%" src\musubi_tuner\ernie_image_cache_text_encoder_outputs.py ^
+  --dataset_config "%DATASET_CONFIG%" ^
+  --text_encoder "%ERNIE_TEXT_ENCODER%" ^
+  --device cuda ^
+  --batch_size 1 ^
+  --skip_existing
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Text encoder caching failed.
+  exit /b 1
+)
+
+> "%STATUS_FILE%" echo prepared
+echo [INFO] Dataset prepared. Run train_only_ernie.bat %DATASET_NAME% to start training.
+exit /b 0

training-scripts/musubi/ernie/train_dataset_ernie.bat

@@ -0,0 +1,164 @@
+@echo off
+setlocal EnableExtensions
+
+cd /d "%~dp0\..\.."
+
+if "%~1"=="" (
+  echo Usage: %~nx0 ^<dataset_name^>
+  echo Example: %~nx0 aubreyplaza
+  exit /b 1
+)
+
+set "DATASET_NAME=%~1"
+set "SOURCE_DATASET_DIR=C:\Development\ai-toolkit\datasets\%DATASET_NAME%"
+set "CURRENTSET_DIR=C:\Development\trainers\musubi-tuner\datasets\currentset"
+set "CACHE_DIR=C:\Development\trainers\musubi-tuner\datasets\cache"
+set "STATUS_FILE=C:\Development\trainers\musubi-tuner\datasets\ernie_%DATASET_NAME%.txt"
+set "DATASET_CONFIG=C:\Development\trainers\musubi-tuner\templates\ernie\dataset.currentset.toml"
+set "PYTHON_EXE=C:\Development\trainers\musubi-tuner\venv\Scripts\python.exe"
+set "ERNIE_DIT=C:/Development/ComfyUI/models/diffusion_models/Ernie/ernie-image.safetensors"
+set "ERNIE_VAE=C:/Development/ComfyUI/models/vae/flux2_vae.safetensors"
+set "ERNIE_TEXT_ENCODER=C:/Development/ComfyUI/models/text_encoders/ministral-3-3b.safetensors"
+set "COUNT_FILE=%TEMP%\musubi_ernie_image_count.txt"
+set "EXIT_CODE=0"
+set "IMAGE_COUNT=0"
+set "OUTPUT_NAME=ernie_%DATASET_NAME%_v1"
+
+if not exist "%SOURCE_DATASET_DIR%" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Dataset directory not found: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+if not exist "%PYTHON_EXE%" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Python executable not found: %PYTHON_EXE%
+  exit /b 1
+)
+
+"%PYTHON_EXE%" -c "from pathlib import Path; exts={'.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl'}; print(sum(1 for f in Path(r'%SOURCE_DATASET_DIR%').iterdir() if f.is_file() and f.suffix.lower() in exts))" > "%COUNT_FILE%"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to count supported image files in: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+set /p IMAGE_COUNT=<"%COUNT_FILE%"
+del /q "%COUNT_FILE%" >nul 2>nul
+
+if "%IMAGE_COUNT%"=="0" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] No supported image files found in: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+> "%STATUS_FILE%" echo starting
+
+echo [INFO] Dataset name   : %DATASET_NAME%
+echo [INFO] Source dataset : %SOURCE_DATASET_DIR%
+echo [INFO] Staging dir    : %CURRENTSET_DIR%
+echo [INFO] Cache dir      : %CACHE_DIR%
+echo [INFO] Status file    : %STATUS_FILE%
+echo [INFO] Dataset config : %DATASET_CONFIG%
+echo [INFO] Output prefix  : %OUTPUT_NAME%
+echo [INFO] Image count    : %IMAGE_COUNT%
+
+if not exist "%CURRENTSET_DIR%" mkdir "%CURRENTSET_DIR%"
+if not exist "%CACHE_DIR%" mkdir "%CACHE_DIR%"
+
+powershell -NoProfile -Command "Get-ChildItem -LiteralPath '%CURRENTSET_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue; Get-ChildItem -LiteralPath '%CACHE_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to clear staging or cache directories.
+  exit /b 1
+)
+
+powershell -NoProfile -Command "Copy-Item -LiteralPath (Get-ChildItem -LiteralPath '%SOURCE_DATASET_DIR%' -File | Where-Object { $_.Extension -in '.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl' } | Select-Object -ExpandProperty FullName) -Destination '%CURRENTSET_DIR%' -Force"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to copy dataset images into currentset.
+  exit /b 1
+)
+
+powershell -NoProfile -Command "Get-ChildItem -LiteralPath '%CURRENTSET_DIR%' -File | Where-Object { $_.Extension -in '.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl' } | ForEach-Object { $caption = [System.IO.Path]::ChangeExtension($_.FullName, '.txt'); if (-not (Test-Path -LiteralPath $caption)) { New-Item -ItemType File -Path $caption | Out-Null } }"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to create empty caption files in currentset.
+  set "EXIT_CODE=1"
+  goto cleanup
+)
+
+echo [INFO] Running latent caching...
+"%PYTHON_EXE%" src\musubi_tuner\ernie_image_cache_latents.py ^
+  --dataset_config "%DATASET_CONFIG%" ^
+  --vae "%ERNIE_VAE%" ^
+  --device cuda ^
+  --skip_existing
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Latent caching failed.
+  set "EXIT_CODE=1"
+  goto cleanup
+)
+
+echo [INFO] Running text encoder caching...
+"%PYTHON_EXE%" src\musubi_tuner\ernie_image_cache_text_encoder_outputs.py ^
+  --dataset_config "%DATASET_CONFIG%" ^
+  --text_encoder "%ERNIE_TEXT_ENCODER%" ^
+  --device cuda ^
+  --batch_size 1 ^
+  --skip_existing
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Text encoder caching failed.
+  set "EXIT_CODE=1"
+  goto cleanup
+)
+
+> "%STATUS_FILE%" echo running
+echo [INFO] Running ERNIE-Image LoRA training...
+"%PYTHON_EXE%" -m accelerate.commands.launch --num_cpu_threads_per_process 1 --mixed_precision bf16 src\musubi_tuner\ernie_image_train_network.py ^
+  --mixed_precision bf16 ^
+  --dataset_config "%DATASET_CONFIG%" ^
+  --dit "%ERNIE_DIT%" ^
+  --vae "%ERNIE_VAE%" ^
+  --text_encoder "%ERNIE_TEXT_ENCODER%" ^
+  --sdpa ^
+  --timestep_sampling shift ^
+  --weighting_scheme none ^
+  --discrete_flow_shift 4.0 ^
+  --optimizer_type adamw8bit ^
+  --learning_rate 1e-4 ^
+  --gradient_checkpointing ^
+  --max_data_loader_n_workers 2 ^
+  --persistent_data_loader_workers ^
+  --fp8_base --fp8_scaled ^
+  --fp8_text_encoder ^
+  --blocks_to_swap 4 ^
+  --use_pinned_memory_for_block_swap ^
+  --network_module networks.lora_ernie_image ^
+  --network_dim 32 ^
+  --network_alpha 32 ^
+  --max_train_epochs 100 ^
+  --save_every_n_epochs 100 ^
+  --seed 42 ^
+  --output_dir output ^
+  --output_name "%OUTPUT_NAME%"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Training failed.
+  set "EXIT_CODE=1"
+)
+
+:cleanup
+echo [INFO] Cleaning staging and cache directories...
+powershell -NoProfile -Command "Get-ChildItem -LiteralPath '%CURRENTSET_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue; Get-ChildItem -LiteralPath '%CACHE_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue"
+
+if not "%EXIT_CODE%"=="0" (
+  echo [ERROR] Wrapper finished with failures.
+  exit /b %EXIT_CODE%
+)
+
+> "%STATUS_FILE%" echo finished
+echo [INFO] Completed successfully.
+exit /b 0

training-scripts/musubi/ernie/train_only_ernie.bat

@@ -0,0 +1,68 @@
+@echo off
+setlocal EnableExtensions
+
+cd /d "%~dp0\..\.."
+
+if "%~1"=="" (
+  echo Usage: %~nx0 ^<dataset_name^>
+  echo Example: %~nx0 aubreyplaza
+  exit /b 1
+)
+
+set "DATASET_NAME=%~1"
+set "STATUS_FILE=C:\Development\trainers\musubi-tuner\datasets\ernie_%DATASET_NAME%.txt"
+set "DATASET_CONFIG=C:\Development\trainers\musubi-tuner\templates\ernie\dataset.currentset.toml"
+set "PYTHON_EXE=C:\Development\trainers\musubi-tuner\venv\Scripts\python.exe"
+set "ERNIE_DIT=C:/Development/ComfyUI/models/diffusion_models/Ernie/ernie-image.safetensors"
+set "ERNIE_VAE=C:/Development/ComfyUI/models/vae/flux2_vae.safetensors"
+set "ERNIE_TEXT_ENCODER=C:/Development/ComfyUI/models/text_encoders/ministral-3-3b.safetensors"
+set "OUTPUT_NAME=ernie_%DATASET_NAME%_v1"
+
+if not exist "%PYTHON_EXE%" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Python executable not found: %PYTHON_EXE%
+  exit /b 1
+)
+
+echo [INFO] Dataset name   : %DATASET_NAME%
+echo [INFO] Dataset config : %DATASET_CONFIG%
+echo [INFO] Output prefix  : %OUTPUT_NAME%
+
+> "%STATUS_FILE%" echo running
+echo [INFO] Running ERNIE-Image LoRA training...
+"%PYTHON_EXE%" -m accelerate.commands.launch --num_cpu_threads_per_process 1 --mixed_precision bf16 src\musubi_tuner\ernie_image_train_network.py ^
+  --mixed_precision bf16 ^
+  --dataset_config "%DATASET_CONFIG%" ^
+  --dit "%ERNIE_DIT%" ^
+  --vae "%ERNIE_VAE%" ^
+  --text_encoder "%ERNIE_TEXT_ENCODER%" ^
+  --sdpa ^
+  --timestep_sampling shift ^
+  --weighting_scheme none ^
+  --discrete_flow_shift 4.0 ^
+  --optimizer_type adamw8bit ^
+  --learning_rate 1e-4 ^
+  --gradient_checkpointing ^
+  --max_data_loader_n_workers 2 ^
+  --persistent_data_loader_workers ^
+  --fp8_base --fp8_scaled ^
+  --fp8_text_encoder ^
+  --blocks_to_swap 4 ^
+  --use_pinned_memory_for_block_swap ^
+  --network_module networks.lora_ernie_image ^
+  --network_dim 32 ^
+  --network_alpha 32 ^
+  --max_train_epochs 100 ^
+  --save_every_n_epochs 100 ^
+  --seed 42 ^
+  --output_dir output ^
+  --output_name "%OUTPUT_NAME%"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Training failed.
+  exit /b 1
+)
+
+> "%STATUS_FILE%" echo finished
+echo [INFO] Completed successfully.
+exit /b 0

training-scripts/musubi/ltx23/dataset.currentset.toml

@@ -0,0 +1,15 @@
+# Image-only LTX 2.3 dataset config for staged wrapper runs.
+# The wrapper copies a selected source dataset into datasets/currentset,
+# creates empty .txt sidecars there, then uses this config for caching/training.
+
+[general]
+resolution = [512, 512]
+caption_extension = ".txt"
+batch_size = 1
+enable_bucket = true
+bucket_no_upscale = false
+
+[[datasets]]
+image_directory = "C:/Development/trainers/musubi-tuner/datasets/currentset"
+cache_directory = "C:/Development/trainers/musubi-tuner/datasets/cache"
+num_repeats = 1

training-scripts/musubi/ltx23/prepare_dataset_ltx23.bat

@@ -0,0 +1,121 @@
+@echo off
+setlocal EnableExtensions
+
+cd /d "%~dp0\..\.."
+
+if "%~1"=="" (
+  echo Usage: %~nx0 ^<dataset_name^>
+  echo Example: %~nx0 aubreyplaza
+  exit /b 1
+)
+
+set "DATASET_NAME=%~1"
+set "SOURCE_DATASET_DIR=C:\Development\ai-toolkit\datasets\%DATASET_NAME%"
+set "CURRENTSET_DIR=C:\Development\trainers\musubi-tuner\datasets\currentset"
+set "CACHE_DIR=C:\Development\trainers\musubi-tuner\datasets\cache"
+set "STATUS_FILE=C:\Development\trainers\musubi-tuner\datasets\ltx23_%DATASET_NAME%.txt"
+set "DATASET_CONFIG=C:\Development\trainers\musubi-tuner\templates\ltx23\dataset.currentset.toml"
+set "PYTHON_EXE=C:\Development\trainers\musubi-tuner\venv\Scripts\python.exe"
+set "LTX2_CHECKPOINT=C:/Development/ComfyUI/models/checkpoints/LTX2/ltx-2.3-22b-dev.safetensors"
+set "GEMMA_SAFETENSORS=C:/Development/ComfyUI/models/text_encoders/gemma_3_12B_it_fp8_e4m3fn.safetensors"
+set "COUNT_FILE=%TEMP%\musubi_ltx23_image_count.txt"
+set "IMAGE_COUNT=0"
+
+if not exist "%SOURCE_DATASET_DIR%" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Dataset directory not found: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+if not exist "%PYTHON_EXE%" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Python executable not found: %PYTHON_EXE%
+  exit /b 1
+)
+
+"%PYTHON_EXE%" -c "from pathlib import Path; exts={'.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl'}; print(sum(1 for f in Path(r'%SOURCE_DATASET_DIR%').iterdir() if f.is_file() and f.suffix.lower() in exts))" > "%COUNT_FILE%"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to count supported image files in: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+set /p IMAGE_COUNT=<"%COUNT_FILE%"
+del /q "%COUNT_FILE%" >nul 2>nul
+
+if "%IMAGE_COUNT%"=="0" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] No supported image files found in: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+> "%STATUS_FILE%" echo starting
+
+echo [INFO] Dataset name   : %DATASET_NAME%
+echo [INFO] Source dataset : %SOURCE_DATASET_DIR%
+echo [INFO] Staging dir    : %CURRENTSET_DIR%
+echo [INFO] Cache dir      : %CACHE_DIR%
+echo [INFO] Status file    : %STATUS_FILE%
+echo [INFO] Dataset config : %DATASET_CONFIG%
+echo [INFO] Image count    : %IMAGE_COUNT%
+
+if not exist "%CURRENTSET_DIR%" mkdir "%CURRENTSET_DIR%"
+if not exist "%CACHE_DIR%" mkdir "%CACHE_DIR%"
+
+echo [INFO] Clearing staging and cache directories...
+powershell -NoProfile -Command "Get-ChildItem -LiteralPath '%CURRENTSET_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue; Get-ChildItem -LiteralPath '%CACHE_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to clear staging or cache directories.
+  exit /b 1
+)
+
+echo [INFO] Copying dataset images to currentset...
+powershell -NoProfile -Command "Copy-Item -LiteralPath (Get-ChildItem -LiteralPath '%SOURCE_DATASET_DIR%' -File | Where-Object { $_.Extension -in '.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl' } | Select-Object -ExpandProperty FullName) -Destination '%CURRENTSET_DIR%' -Force"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to copy dataset images into currentset.
+  exit /b 1
+)
+
+echo [INFO] Creating empty caption files...
+powershell -NoProfile -Command "Get-ChildItem -LiteralPath '%CURRENTSET_DIR%' -File | Where-Object { $_.Extension -in '.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl' } | ForEach-Object { $caption = [System.IO.Path]::ChangeExtension($_.FullName, '.txt'); if (-not (Test-Path -LiteralPath $caption)) { New-Item -ItemType File -Path $caption | Out-Null } }"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to create empty caption files in currentset.
+  exit /b 1
+)
+
+echo [INFO] Running latent caching...
+"%PYTHON_EXE%" ltx2_cache_latents.py ^
+  --dataset_config "%DATASET_CONFIG%" ^
+  --ltx2_checkpoint "%LTX2_CHECKPOINT%" ^
+  --device cuda ^
+  --vae_dtype bf16 ^
+  --ltx2_mode video ^
+  --skip_existing
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Latent caching failed.
+  exit /b 1
+)
+
+echo [INFO] Running text encoder caching...
+"%PYTHON_EXE%" ltx2_cache_text_encoder_outputs.py ^
+  --dataset_config "%DATASET_CONFIG%" ^
+  --ltx2_checkpoint "%LTX2_CHECKPOINT%" ^
+  --gemma_safetensors "%GEMMA_SAFETENSORS%" ^
+  --device cuda ^
+  --mixed_precision bf16 ^
+  --ltx2_mode video ^
+  --batch_size 1 ^
+  --skip_existing
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Text encoder caching failed.
+  exit /b 1
+)
+
+> "%STATUS_FILE%" echo prepared
+echo [INFO] Dataset prepared. Run train_only_ltx23.bat %DATASET_NAME% to start training.
+exit /b 0

training-scripts/musubi/ltx23/train_dataset_ltx23.bat

@@ -0,0 +1,142 @@
+@echo off
+setlocal EnableExtensions
+
+cd /d "%~dp0\..\.."
+
+if "%~1"=="" (
+  echo Usage: %~nx0 ^<dataset_name^>
+  echo Example: %~nx0 aubreyplaza
+  exit /b 1
+)
+
+set "DATASET_NAME=%~1"
+set "SOURCE_DATASET_DIR=C:\Development\ai-toolkit\datasets\%DATASET_NAME%"
+set "CURRENTSET_DIR=C:\Development\trainers\musubi-tuner\datasets\currentset"
+set "CACHE_DIR=C:\Development\trainers\musubi-tuner\datasets\cache"
+set "STATUS_FILE=C:\Development\trainers\musubi-tuner\datasets\ltx23_%DATASET_NAME%.txt"
+set "DATASET_CONFIG=C:\Development\trainers\musubi-tuner\templates\ltx23\dataset.currentset.toml"
+set "PYTHON_EXE=C:\Development\trainers\musubi-tuner\venv\Scripts\python.exe"
+set "LTX2_CHECKPOINT=C:/Development/ComfyUI/models/checkpoints/LTX2/ltx-2.3-22b-dev.safetensors"
+set "GEMMA_SAFETENSORS=C:/Development/ComfyUI/models/text_encoders/gemma_3_12B_it_fp8_e4m3fn.safetensors"
+set "COUNT_FILE=%TEMP%\musubi_ltx23_image_count.txt"
+set "EXIT_CODE=0"
+set "IMAGE_COUNT=0"
+set "OUTPUT_NAME=ltx23_%DATASET_NAME%_v1"
+
+if not exist "%SOURCE_DATASET_DIR%" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Dataset directory not found: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+if not exist "%PYTHON_EXE%" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Python executable not found: %PYTHON_EXE%
+  exit /b 1
+)
+
+"%PYTHON_EXE%" -c "from pathlib import Path; exts={'.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl'}; print(sum(1 for f in Path(r'%SOURCE_DATASET_DIR%').iterdir() if f.is_file() and f.suffix.lower() in exts))" > "%COUNT_FILE%"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to count supported image files in: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+set /p IMAGE_COUNT=<"%COUNT_FILE%"
+del /q "%COUNT_FILE%" >nul 2>nul
+
+if "%IMAGE_COUNT%"=="0" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] No supported image files found in: %SOURCE_DATASET_DIR%
+  exit /b 1
+)
+
+> "%STATUS_FILE%" echo starting
+
+echo [INFO] Dataset name   : %DATASET_NAME%
+echo [INFO] Source dataset : %SOURCE_DATASET_DIR%
+echo [INFO] Staging dir    : %CURRENTSET_DIR%
+echo [INFO] Cache dir      : %CACHE_DIR%
+echo [INFO] Status file    : %STATUS_FILE%
+echo [INFO] Dataset config : %DATASET_CONFIG%
+echo [INFO] Output prefix  : %OUTPUT_NAME%
+echo [INFO] Image count    : %IMAGE_COUNT%
+
+if not exist "%CURRENTSET_DIR%" mkdir "%CURRENTSET_DIR%"
+if not exist "%CACHE_DIR%" mkdir "%CACHE_DIR%"
+
+powershell -NoProfile -Command "Get-ChildItem -LiteralPath '%CURRENTSET_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue; Get-ChildItem -LiteralPath '%CACHE_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to clear staging or cache directories.
+  exit /b 1
+)
+
+powershell -NoProfile -Command "Copy-Item -LiteralPath (Get-ChildItem -LiteralPath '%SOURCE_DATASET_DIR%' -File | Where-Object { $_.Extension -in '.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl' } | Select-Object -ExpandProperty FullName) -Destination '%CURRENTSET_DIR%' -Force"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to copy dataset images into currentset.
+  exit /b 1
+)
+
+powershell -NoProfile -Command "Get-ChildItem -LiteralPath '%CURRENTSET_DIR%' -File | Where-Object { $_.Extension -in '.png','.jpg','.jpeg','.webp','.bmp','.avif','.jxl' } | ForEach-Object { $caption = [System.IO.Path]::ChangeExtension($_.FullName, '.txt'); if (-not (Test-Path -LiteralPath $caption)) { New-Item -ItemType File -Path $caption | Out-Null } }"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Failed to create empty caption files in currentset.
+  set "EXIT_CODE=1"
+  goto cleanup
+)
+
+echo [INFO] Running latent caching...
+"%PYTHON_EXE%" ltx2_cache_latents.py ^
+  --dataset_config "%DATASET_CONFIG%" ^
+  --ltx2_checkpoint "%LTX2_CHECKPOINT%" ^
+  --device cuda ^
+  --vae_dtype bf16 ^
+  --ltx2_mode video ^
+  --skip_existing
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Latent caching failed.
+  set "EXIT_CODE=1"
+  goto cleanup
+)
+
+echo [INFO] Running text encoder caching...
+"%PYTHON_EXE%" ltx2_cache_text_encoder_outputs.py ^
+  --dataset_config "%DATASET_CONFIG%" ^
+  --ltx2_checkpoint "%LTX2_CHECKPOINT%" ^
+  --gemma_safetensors "%GEMMA_SAFETENSORS%" ^
+  --device cuda ^
+  --mixed_precision bf16 ^
+  --ltx2_mode video ^
+  --batch_size 1 ^
+  --skip_existing
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Text encoder caching failed.
+  set "EXIT_CODE=1"
+  goto cleanup
+)
+
+> "%STATUS_FILE%" echo running
+echo [INFO] Running optimized training...
+"%PYTHON_EXE%" -m accelerate.commands.launch --num_cpu_threads_per_process 1 --mixed_precision bf16 ltx2_train_network.py --mixed_precision bf16 --dataset_config "%DATASET_CONFIG%" --gemma_safetensors "%GEMMA_SAFETENSORS%" --ltx2_checkpoint "%LTX2_CHECKPOINT%" --ltx_version 2.3 --ltx_version_check_mode error --ltx2_mode video --lora_target_preset video_sa_ca_ff --fp8_base --fp8_scaled --blocks_to_swap 4 --use_pinned_memory_for_block_swap --sdpa --gradient_checkpointing --learning_rate 1e-4 --optimizer_type AdamW --lr_scheduler constant_with_warmup --lr_warmup_steps 100 --max_train_epochs 120 --max_data_loader_n_workers 8 --persistent_data_loader_workers --save_every_n_epochs 120 --network_module networks.lora_ltx2 --network_dim 32 --network_alpha 32 --no_save_original_lora --timestep_sampling shifted_logit_normal --output_dir output --output_name "%OUTPUT_NAME%"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Training failed.
+  set "EXIT_CODE=1"
+)
+
+:cleanup
+echo [INFO] Cleaning staging and cache directories...
+powershell -NoProfile -Command "Get-ChildItem -LiteralPath '%CURRENTSET_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue; Get-ChildItem -LiteralPath '%CACHE_DIR%' -Force -ErrorAction SilentlyContinue | Remove-Item -Recurse -Force -ErrorAction SilentlyContinue"
+
+if not "%EXIT_CODE%"=="0" (
+  echo [ERROR] Wrapper finished with failures.
+  exit /b %EXIT_CODE%
+)
+
+> "%STATUS_FILE%" echo finished
+echo [INFO] Completed successfully.
+exit /b 0

training-scripts/musubi/ltx23/train_only_ltx23.bat

@@ -0,0 +1,41 @@
+@echo off
+setlocal EnableExtensions
+
+cd /d "%~dp0\..\.."
+
+if "%~1"=="" (
+  echo Usage: %~nx0 ^<dataset_name^>
+  echo Example: %~nx0 aubreyplaza
+  exit /b 1
+)
+
+set "DATASET_NAME=%~1"
+set "STATUS_FILE=C:\Development\trainers\musubi-tuner\datasets\ltx23_%DATASET_NAME%.txt"
+set "DATASET_CONFIG=C:\Development\trainers\musubi-tuner\templates\ltx23\dataset.currentset.toml"
+set "PYTHON_EXE=C:\Development\trainers\musubi-tuner\venv\Scripts\python.exe"
+set "LTX2_CHECKPOINT=C:/Development/ComfyUI/models/checkpoints/LTX2/ltx-2.3-22b-dev.safetensors"
+set "GEMMA_SAFETENSORS=C:/Development/ComfyUI/models/text_encoders/gemma_3_12B_it_fp8_e4m3fn.safetensors"
+set "OUTPUT_NAME=ltx23_%DATASET_NAME%_v1"
+
+if not exist "%PYTHON_EXE%" (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Python executable not found: %PYTHON_EXE%
+  exit /b 1
+)
+
+echo [INFO] Dataset name   : %DATASET_NAME%
+echo [INFO] Dataset config : %DATASET_CONFIG%
+echo [INFO] Output prefix  : %OUTPUT_NAME%
+
+> "%STATUS_FILE%" echo running
+echo [INFO] Running optimized training...
+"%PYTHON_EXE%" -m accelerate.commands.launch --num_cpu_threads_per_process 1 --mixed_precision bf16 ltx2_train_network.py --mixed_precision bf16 --dataset_config "%DATASET_CONFIG%" --gemma_safetensors "%GEMMA_SAFETENSORS%" --ltx2_checkpoint "%LTX2_CHECKPOINT%" --ltx_version 2.3 --ltx_version_check_mode error --ltx2_mode video --lora_target_preset video_sa_ca_ff --fp8_base --fp8_scaled --blocks_to_swap 4 --use_pinned_memory_for_block_swap --sdpa --gradient_checkpointing --learning_rate 1e-4 --optimizer_type AdamW --lr_scheduler constant_with_warmup --lr_warmup_steps 100 --max_train_epochs 120 --max_data_loader_n_workers 8 --persistent_data_loader_workers --save_every_n_epochs 120 --network_module networks.lora_ltx2 --network_dim 32 --network_alpha 32 --no_save_original_lora --timestep_sampling shifted_logit_normal --output_dir output --output_name "%OUTPUT_NAME%"
+if errorlevel 1 (
+  > "%STATUS_FILE%" echo errored
+  echo [ERROR] Training failed.
+  exit /b 1
+)
+
+> "%STATUS_FILE%" echo finished
+echo [INFO] Completed successfully.
+exit /b 0

training-scripts/onetrainer/ernie_template.json

@@ -0,0 +1,525 @@
+{
+  "__version": 10,
+  "training_method": "LORA",
+  "model_type": "ERNIE",
+  "debug_mode": false,
+  "debug_dir": "debug",
+  "workspace_dir": "workspace/run",
+  "cache_dir": "workspace-cache/run",
+  "tensorboard": true,
+  "tensorboard_expose": false,
+  "tensorboard_always_on": false,
+  "tensorboard_port": 6006,
+  "validation": false,
+  "validate_after": 1,
+  "validate_after_unit": "EPOCH",
+  "continue_last_backup": false,
+  "prevent_overwrites": false,
+  "include_train_config": "NONE",
+  "multi_gpu": false,
+  "device_indexes": "",
+  "gradient_reduce_precision": "FLOAT_32_STOCHASTIC",
+  "fused_gradient_reduce": true,
+  "async_gradient_reduce": true,
+  "async_gradient_reduce_buffer": 100,
+  "base_model_name": "baidu/ERNIE-Image",
+  "output_dtype": "BFLOAT_16",
+  "output_model_format": "SAFETENSORS",
+  "output_model_destination": "models/ernie_[person_to_train]_v1_onetrainer.safetensors",
+  "gradient_checkpointing": "ON",
+  "enable_async_offloading": true,
+  "enable_activation_offloading": true,
+  "layer_offload_fraction": 0.0,
+  "force_circular_padding": false,
+  "compile": true,
+  "concept_file_name": "training_concepts/concepts.json",
+  "concepts": [
+    {
+      "__version": 2,
+      "image": {
+        "__version": 0,
+        "enable_crop_jitter": true,
+        "enable_random_flip": false,
+        "enable_fixed_flip": false,
+        "enable_random_rotate": false,
+        "enable_fixed_rotate": false,
+        "random_rotate_max_angle": 0.0,
+        "enable_random_brightness": false,
+        "enable_fixed_brightness": false,
+        "random_brightness_max_strength": 0.0,
+        "enable_random_contrast": false,
+        "enable_fixed_contrast": false,
+        "random_contrast_max_strength": 0.0,
+        "enable_random_saturation": false,
+        "enable_fixed_saturation": false,
+        "random_saturation_max_strength": 0.0,
+        "enable_random_hue": false,
+        "enable_fixed_hue": false,
+        "random_hue_max_strength": 0.0,
+        "enable_resolution_override": false,
+        "resolution_override": "512",
+        "enable_random_circular_mask_shrink": false,
+        "enable_random_mask_rotate_crop": false
+      },
+      "text": {
+        "__version": 0,
+        "prompt_source": "sample",
+        "prompt_path": "",
+        "enable_tag_shuffling": false,
+        "tag_delimiter": ",",
+        "keep_tags_count": 1,
+        "tag_dropout_enable": false,
+        "tag_dropout_mode": "FULL",
+        "tag_dropout_probability": 0.0,
+        "tag_dropout_special_tags_mode": "NONE",
+        "tag_dropout_special_tags": "",
+        "tag_dropout_special_tags_regex": false,
+        "caps_randomize_enable": false,
+        "caps_randomize_mode": "capslock, title, first, random",
+        "caps_randomize_probability": 0.0,
+        "caps_randomize_lowercase": false
+      },
+      "name": "",
+      "path": "C:/Development/ai-toolkit/datasets/[person_to_train]",
+      "seed": -469069486,
+      "enabled": true,
+      "type": "STANDARD",
+      "include_subdirectories": false,
+      "image_variations": 1,
+      "text_variations": 1,
+      "balancing": 1.0,
+      "balancing_strategy": "REPEATS",
+      "loss_weight": 1.0,
+      "concept_stats": {}
+    }
+  ],
+  "aspect_ratio_bucketing": true,
+  "latent_caching": true,
+  "clear_cache_before_training": true,
+  "learning_rate_scheduler": "CONSTANT",
+  "custom_learning_rate_scheduler": null,
+  "scheduler_params": [],
+  "learning_rate": 0.0003,
+  "learning_rate_warmup_steps": 200.0,
+  "learning_rate_cycles": 1.0,
+  "learning_rate_min_factor": 0.0,
+  "epochs": 120,
+  "batch_size": 2,
+  "gradient_accumulation_steps": 1,
+  "ema": "OFF",
+  "ema_decay": 0.999,
+  "ema_update_step_interval": 5,
+  "dataloader_threads": 1,
+  "train_device": "cuda",
+  "temp_device": "cpu",
+  "train_dtype": "BFLOAT_16",
+  "fallback_train_dtype": "BFLOAT_16",
+  "enable_autocast_cache": true,
+  "only_cache": false,
+  "resolution": "512",
+  "frames": "25",
+  "mse_strength": 1.0,
+  "mae_strength": 0.0,
+  "log_cosh_strength": 0.0,
+  "huber_strength": 0.0,
+  "huber_delta": 1.0,
+  "vb_loss_strength": 1.0,
+  "loss_weight_fn": "CONSTANT",
+  "loss_weight_strength": 5.0,
+  "dropout_probability": 0.0,
+  "loss_scaler": "NONE",
+  "learning_rate_scaler": "NONE",
+  "clip_grad_norm": 1.0,
+  "offset_noise_weight": 0.0,
+  "generalized_offset_noise": false,
+  "perturbation_noise_weight": 0.0,
+  "rescale_noise_scheduler_to_zero_terminal_snr": false,
+  "force_v_prediction": false,
+  "force_epsilon_prediction": false,
+  "min_noising_strength": 0.0,
+  "max_noising_strength": 1.0,
+  "timestep_distribution": "LOGIT_NORMAL",
+  "noising_weight": 0.0,
+  "noising_bias": 0.0,
+  "timestep_shift": 1.0,
+  "dynamic_timestep_shifting": false,
+  "unet": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 0,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "prior": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 0,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "transformer": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 0,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "INT_W8A8",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "quantization": {
+    "__version": 0,
+    "layer_filter": "layers",
+    "layer_filter_preset": "blocks",
+    "layer_filter_regex": false,
+    "svd_dtype": "NONE",
+    "svd_rank": 16,
+    "cache_dir": null
+  },
+  "text_encoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": false,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_8",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_layer_skip": 0,
+  "text_encoder_sequence_length": 512,
+  "text_encoder_2": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_2_layer_skip": 0,
+  "text_encoder_2_sequence_length": 77,
+  "text_encoder_3": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_3_layer_skip": 0,
+  "text_encoder_4": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_4_layer_skip": 0,
+  "vae": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "effnet_encoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "decoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "decoder_text_encoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "decoder_vqgan": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "masked_training": false,
+  "unmasked_probability": 0.1,
+  "unmasked_weight": 0.1,
+  "normalize_masked_area_loss": false,
+  "masked_prior_preservation_weight": 0.0,
+  "custom_conditioning_image": false,
+  "layer_filter": "self_attention,mlp",
+  "layer_filter_preset": "attn-mlp",
+  "layer_filter_regex": false,
+  "embedding_learning_rate": null,
+  "preserve_embedding_norm": false,
+  "embedding": {
+    "__version": 0,
+    "uuid": "55a5c9af-2ee6-4098-97a3-cc66342dc887",
+    "model_name": "",
+    "placeholder": "<embedding>",
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "token_count": 1,
+    "initial_embedding_text": "*",
+    "is_output_embedding": false
+  },
+  "additional_embeddings": [],
+  "embedding_weight_dtype": "FLOAT_32",
+  "cloud": {
+    "__version": 0,
+    "enabled": false,
+    "type": "RUNPOD",
+    "file_sync": "NATIVE_SCP",
+    "create": true,
+    "name": "OneTrainer",
+    "tensorboard_tunnel": true,
+    "sub_type": "",
+    "gpu_type": "",
+    "volume_size": 100,
+    "min_download": 0,
+    "remote_dir": "/workspace",
+    "huggingface_cache_dir": "/workspace/huggingface_cache",
+    "onetrainer_dir": "/workspace/OneTrainer",
+    "install_cmd": "git clone https://github.com/Nerogar/OneTrainer",
+    "install_onetrainer": true,
+    "update_onetrainer": true,
+    "detach_trainer": false,
+    "run_id": "job1",
+    "download_samples": true,
+    "download_output_model": true,
+    "download_saves": true,
+    "download_backups": false,
+    "download_tensorboard": false,
+    "delete_workspace": false,
+    "on_finish": "NONE",
+    "on_error": "NONE",
+    "on_detached_finish": "NONE",
+    "on_detached_error": "NONE"
+  },
+  "peft_type": "LORA",
+  "lora_model_name": "",
+  "lora_rank": 16,
+  "lora_alpha": 1.0,
+  "lora_decompose": false,
+  "lora_decompose_norm_epsilon": true,
+  "lora_decompose_output_axis": false,
+  "lora_weight_dtype": "FLOAT_32",
+  "bundle_additional_embeddings": true,
+  "oft_block_size": 32,
+  "oft_coft": false,
+  "coft_eps": 0.0001,
+  "oft_block_share": false,
+  "optimizer": {
+    "__version": 0,
+    "optimizer": "ADAMW",
+    "adam_w_mode": false,
+    "alpha": null,
+    "amsgrad": false,
+    "beta1": 0.9,
+    "beta2": 0.999,
+    "beta3": null,
+    "bias_correction": false,
+    "block_wise": false,
+    "capturable": false,
+    "centered": false,
+    "clip_threshold": null,
+    "d0": null,
+    "d_coef": null,
+    "dampening": null,
+    "decay_rate": null,
+    "decouple": false,
+    "differentiable": false,
+    "eps": 1e-8,
+    "eps2": null,
+    "foreach": false,
+    "fsdp_in_use": false,
+    "fused": true,
+    "fused_back_pass": false,
+    "growth_rate": null,
+    "initial_accumulator_value": null,
+    "initial_accumulator": null,
+    "is_paged": false,
+    "log_every": null,
+    "lr_decay": null,
+    "max_unorm": null,
+    "maximize": false,
+    "min_8bit_size": null,
+    "quant_block_size": null,
+    "momentum": null,
+    "nesterov": false,
+    "no_prox": false,
+    "optim_bits": null,
+    "percentile_clipping": null,
+    "r": null,
+    "relative_step": false,
+    "safeguard_warmup": false,
+    "scale_parameter": false,
+    "stochastic_rounding": false,
+    "use_bias_correction": false,
+    "use_triton": false,
+    "warmup_init": false,
+    "weight_decay": 0.01,
+    "weight_lr_power": null,
+    "decoupled_decay": false,
+    "fixed_decay": false,
+    "rectify": false,
+    "degenerated_to_sgd": false,
+    "k": null,
+    "xi": null,
+    "n_sma_threshold": null,
+    "ams_bound": false,
+    "adanorm": false,
+    "adam_debias": false,
+    "slice_p": null,
+    "cautious": false,
+    "weight_decay_by_lr": true,
+    "prodigy_steps": null,
+    "use_speed": false,
+    "split_groups": true,
+    "split_groups_mean": true,
+    "factored": true,
+    "factored_fp32": true,
+    "use_stableadamw": true,
+    "use_cautious": false,
+    "use_grams": false,
+    "use_adopt": false,
+    "d_limiter": true,
+    "use_schedulefree": true,
+    "use_orthograd": false,
+    "nnmf_factor": false,
+    "orthogonal_gradient": false,
+    "use_atan2": false,
+    "use_AdEMAMix": false,
+    "beta3_ema": null,
+    "alpha_grad": null,
+    "beta1_warmup": null,
+    "min_beta1": null,
+    "Simplified_AdEMAMix": false,
+    "kourkoutas_beta": false,
+    "schedulefree_c": null,
+    "ns_steps": null,
+    "MuonWithAuxAdam": false,
+    "muon_hidden_layers": null,
+    "muon_adam_regex": false,
+    "muon_adam_lr": null,
+    "muon_te1_adam_lr": null,
+    "muon_te2_adam_lr": null,
+    "muon_adam_config": {},
+    "rms_rescaling": true,
+    "normuon_variant": false,
+    "beta2_normuon": null,
+    "low_rank_ortho": false,
+    "ortho_rank": null,
+    "accelerated_ns": false,
+    "cautious_wd": false,
+    "approx_mars": false,
+    "auto_kappa_p": false,
+    "compile": false
+  },
+  "optimizer_defaults": {},
+  "sample_definition_file_name": "training_samples/samples.json",
+  "samples": [],
+  "sample_after": 10,
+  "sample_after_unit": "MINUTE",
+  "sample_skip_first": 0,
+  "sample_image_format": "JPG",
+  "sample_video_format": "MP4",
+  "sample_audio_format": "MP3",
+  "samples_to_tensorboard": true,
+  "non_ema_sampling": true,
+  "backup_after": 30,
+  "backup_after_unit": "MINUTE",
+  "rolling_backup": true,
+  "rolling_backup_count": 2,
+  "backup_before_save": true,
+  "save_every": 0,
+  "save_every_unit": "NEVER",
+  "save_skip_first": 0,
+  "save_filename_prefix": ""
+}

training-scripts/onetrainer/fk9_prodigy_template.json

@@ -0,0 +1,530 @@
+{
+  "__version": 10,
+  "training_method": "LORA",
+  "model_type": "FLUX_2",
+  "debug_mode": false,
+  "debug_dir": "debug",
+  "workspace_dir": "workspace/run",
+  "cache_dir": "workspace-cache/run",
+  "tensorboard": true,
+  "tensorboard_expose": false,
+  "tensorboard_always_on": false,
+  "tensorboard_port": 6006,
+  "validation": false,
+  "validate_after": 1,
+  "validate_after_unit": "EPOCH",
+  "continue_last_backup": false,
+  "prevent_overwrites": false,
+  "include_train_config": "NONE",
+  "multi_gpu": false,
+  "device_indexes": "",
+  "gradient_reduce_precision": "FLOAT_32_STOCHASTIC",
+  "fused_gradient_reduce": true,
+  "async_gradient_reduce": true,
+  "async_gradient_reduce_buffer": 100,
+  "base_model_name": "black-forest-labs/FLUX.2-klein-base-9B",
+  "output_dtype": "BFLOAT_16",
+  "output_model_format": "SAFETENSORS",
+  "output_model_destination": "models/fk9_[person_to_train]_v1_prodigy.safetensors",
+  "gradient_checkpointing": "ON",
+  "enable_async_offloading": true,
+  "enable_activation_offloading": true,
+  "layer_offload_fraction": 0.0,
+  "force_circular_padding": false,
+  "compile": true,
+  "concept_file_name": "training_concepts/concepts.json",
+  "concepts": [
+    {
+      "__version": 2,
+      "image": {
+        "__version": 0,
+        "enable_crop_jitter": true,
+        "enable_random_flip": false,
+        "enable_fixed_flip": false,
+        "enable_random_rotate": false,
+        "enable_fixed_rotate": false,
+        "random_rotate_max_angle": 0.0,
+        "enable_random_brightness": false,
+        "enable_fixed_brightness": false,
+        "random_brightness_max_strength": 0.0,
+        "enable_random_contrast": false,
+        "enable_fixed_contrast": false,
+        "random_contrast_max_strength": 0.0,
+        "enable_random_saturation": false,
+        "enable_fixed_saturation": false,
+        "random_saturation_max_strength": 0.0,
+        "enable_random_hue": false,
+        "enable_fixed_hue": false,
+        "random_hue_max_strength": 0.0,
+        "enable_resolution_override": false,
+        "resolution_override": "512",
+        "enable_random_circular_mask_shrink": false,
+        "enable_random_mask_rotate_crop": false
+      },
+      "text": {
+        "__version": 0,
+        "prompt_source": "sample",
+        "prompt_path": "",
+        "enable_tag_shuffling": false,
+        "tag_delimiter": ",",
+        "keep_tags_count": 1,
+        "tag_dropout_enable": false,
+        "tag_dropout_mode": "FULL",
+        "tag_dropout_probability": 0.0,
+        "tag_dropout_special_tags_mode": "NONE",
+        "tag_dropout_special_tags": "",
+        "tag_dropout_special_tags_regex": false,
+        "caps_randomize_enable": false,
+        "caps_randomize_mode": "capslock, title, first, random",
+        "caps_randomize_probability": 0.0,
+        "caps_randomize_lowercase": false
+      },
+      "name": "",
+      "path": "C:/Development/ai-toolkit/datasets/[person_to_train]",
+      "seed": -469069486,
+      "enabled": true,
+      "type": "STANDARD",
+      "include_subdirectories": false,
+      "image_variations": 1,
+      "text_variations": 1,
+      "balancing": 1.0,
+      "balancing_strategy": "REPEATS",
+      "loss_weight": 1.0,
+      "concept_stats": {}
+    }
+  ],
+  "aspect_ratio_bucketing": true,
+  "latent_caching": true,
+  "clear_cache_before_training": true,
+  "learning_rate_scheduler": "CONSTANT",
+  "custom_learning_rate_scheduler": null,
+  "scheduler_params": [],
+  "learning_rate": 1.0,
+  "learning_rate_warmup_steps": 200.0,
+  "learning_rate_cycles": 1.0,
+  "learning_rate_min_factor": 0.0,
+  "epochs": 100,
+  "batch_size": 2,
+  "gradient_accumulation_steps": 1,
+  "ema": "OFF",
+  "ema_decay": 0.999,
+  "ema_update_step_interval": 5,
+  "dataloader_threads": 1,
+  "train_device": "cuda",
+  "temp_device": "cpu",
+  "train_dtype": "BFLOAT_16",
+  "fallback_train_dtype": "BFLOAT_16",
+  "enable_autocast_cache": true,
+  "only_cache": false,
+  "resolution": "512",
+  "frames": "25",
+  "mse_strength": 1.0,
+  "mae_strength": 0.0,
+  "log_cosh_strength": 0.0,
+  "huber_strength": 0.0,
+  "huber_delta": 1.0,
+  "vb_loss_strength": 1.0,
+  "loss_weight_fn": "CONSTANT",
+  "loss_weight_strength": 5.0,
+  "dropout_probability": 0.0,
+  "loss_scaler": "NONE",
+  "learning_rate_scaler": "NONE",
+  "clip_grad_norm": 1.0,
+  "offset_noise_weight": 0.0,
+  "generalized_offset_noise": false,
+  "perturbation_noise_weight": 0.0,
+  "rescale_noise_scheduler_to_zero_terminal_snr": false,
+  "force_v_prediction": false,
+  "force_epsilon_prediction": false,
+  "min_noising_strength": 0.0,
+  "max_noising_strength": 1.0,
+  "timestep_distribution": "LOGIT_NORMAL",
+  "noising_weight": 0.0,
+  "noising_bias": 0.0,
+  "timestep_shift": 1.0,
+  "dynamic_timestep_shifting": false,
+  "unet": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 0,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "prior": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 0,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "transformer": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 0,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_8",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "quantization": {
+    "__version": 0,
+    "layer_filter": "transformer_block",
+    "layer_filter_preset": "blocks",
+    "layer_filter_regex": false,
+    "svd_dtype": "NONE",
+    "svd_rank": 16,
+    "cache_dir": "workspace-cache/run/quantization"
+  },
+  "text_encoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": false,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_8",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_layer_skip": 0,
+  "text_encoder_sequence_length": 512,
+  "text_encoder_2": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_2_layer_skip": 0,
+  "text_encoder_2_sequence_length": 77,
+  "text_encoder_3": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_3_layer_skip": 0,
+  "text_encoder_4": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_4_layer_skip": 0,
+  "vae": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "effnet_encoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "decoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "decoder_text_encoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "decoder_vqgan": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "masked_training": false,
+  "unmasked_probability": 0.1,
+  "unmasked_weight": 0.1,
+  "normalize_masked_area_loss": false,
+  "masked_prior_preservation_weight": 0.0,
+  "custom_conditioning_image": false,
+  "layer_filter": "transformer_blocks",
+  "layer_filter_preset": "transformer_block",
+  "layer_filter_regex": false,
+  "embedding_learning_rate": null,
+  "preserve_embedding_norm": false,
+  "embedding": {
+    "__version": 0,
+    "uuid": "32cefb93-f7d2-4083-be65-f367c7ebccb5",
+    "model_name": "",
+    "placeholder": "<embedding>",
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "token_count": 1,
+    "initial_embedding_text": "*",
+    "is_output_embedding": false
+  },
+  "additional_embeddings": [],
+  "embedding_weight_dtype": "FLOAT_32",
+  "cloud": {
+    "__version": 0,
+    "enabled": false,
+    "type": "RUNPOD",
+    "file_sync": "NATIVE_SCP",
+    "create": true,
+    "name": "OneTrainer",
+    "tensorboard_tunnel": true,
+    "sub_type": "",
+    "gpu_type": "",
+    "volume_size": 100,
+    "min_download": 0,
+    "remote_dir": "/workspace",
+    "huggingface_cache_dir": "/workspace/huggingface_cache",
+    "onetrainer_dir": "/workspace/OneTrainer",
+    "install_cmd": "git clone https://github.com/Nerogar/OneTrainer",
+    "install_onetrainer": true,
+    "update_onetrainer": true,
+    "detach_trainer": false,
+    "run_id": "job1",
+    "download_samples": true,
+    "download_output_model": true,
+    "download_saves": true,
+    "download_backups": false,
+    "download_tensorboard": false,
+    "delete_workspace": false,
+    "on_finish": "NONE",
+    "on_error": "NONE",
+    "on_detached_finish": "NONE",
+    "on_detached_error": "NONE"
+  },
+  "peft_type": "LORA",
+  "lora_model_name": "",
+  "lora_rank": 16,
+  "lora_alpha": 1.0,
+  "lora_decompose": false,
+  "lora_decompose_norm_epsilon": true,
+  "lora_decompose_output_axis": false,
+  "lora_weight_dtype": "FLOAT_32",
+  "bundle_additional_embeddings": true,
+  "oft_block_size": 32,
+  "oft_coft": false,
+  "coft_eps": 0.0001,
+  "oft_block_share": false,
+  "optimizer": {
+    "__version": 0,
+    "optimizer": "PRODIGY_ADV",
+    "adam_w_mode": false,
+    "alpha": 5.0,
+    "amsgrad": false,
+    "beta1": 0.9,
+    "beta2": 0.99,
+    "beta3": null,
+    "bias_correction": false,
+    "block_wise": false,
+    "capturable": false,
+    "centered": false,
+    "clip_threshold": null,
+    "d0": 1e-6,
+    "d_coef": 1.0,
+    "dampening": null,
+    "decay_rate": null,
+    "decouple": false,
+    "differentiable": false,
+    "eps": 1e-8,
+    "eps2": null,
+    "foreach": false,
+    "fsdp_in_use": false,
+    "fused": false,
+    "fused_back_pass": false,
+    "growth_rate": "inf",
+    "initial_accumulator_value": null,
+    "initial_accumulator": null,
+    "is_paged": false,
+    "log_every": null,
+    "lr_decay": null,
+    "max_unorm": null,
+    "maximize": false,
+    "min_8bit_size": null,
+    "quant_block_size": null,
+    "momentum": null,
+    "nesterov": false,
+    "no_prox": false,
+    "optim_bits": null,
+    "percentile_clipping": null,
+    "r": null,
+    "relative_step": false,
+    "safeguard_warmup": false,
+    "scale_parameter": false,
+    "stochastic_rounding": true,
+    "use_bias_correction": false,
+    "use_triton": false,
+    "warmup_init": false,
+    "weight_decay": 0.0,
+    "weight_lr_power": null,
+    "decoupled_decay": false,
+    "fixed_decay": false,
+    "rectify": false,
+    "degenerated_to_sgd": false,
+    "k": null,
+    "xi": null,
+    "n_sma_threshold": null,
+    "ams_bound": false,
+    "adanorm": false,
+    "adam_debias": false,
+    "slice_p": 11,
+    "cautious": false,
+    "weight_decay_by_lr": true,
+    "prodigy_steps": 0,
+    "use_speed": false,
+    "split_groups": true,
+    "split_groups_mean": true,
+    "factored": true,
+    "factored_fp32": true,
+    "use_stableadamw": true,
+    "use_cautious": false,
+    "use_grams": false,
+    "use_adopt": false,
+    "d_limiter": false,
+    "use_schedulefree": true,
+    "use_orthograd": false,
+    "nnmf_factor": false,
+    "orthogonal_gradient": false,
+    "use_atan2": false,
+    "use_AdEMAMix": false,
+    "beta3_ema": 0.9999,
+    "alpha_grad": 100.0,
+    "beta1_warmup": null,
+    "min_beta1": null,
+    "Simplified_AdEMAMix": false,
+    "cautious_mask": false,
+    "grams_moment": false,
+    "kourkoutas_beta": false,
+    "k_warmup_steps": null,
+    "schedulefree_c": null,
+    "ns_steps": null,
+    "MuonWithAuxAdam": false,
+    "muon_hidden_layers": null,
+    "muon_adam_regex": false,
+    "muon_adam_lr": null,
+    "muon_te1_adam_lr": null,
+    "muon_te2_adam_lr": null,
+    "muon_adam_config": null,
+    "rms_rescaling": true,
+    "normuon_variant": false,
+    "beta2_normuon": null,
+    "normuon_eps": null,
+    "low_rank_ortho": false,
+    "ortho_rank": null,
+    "accelerated_ns": false,
+    "cautious_wd": false,
+    "approx_mars": false,
+    "kappa_p": null,
+    "auto_kappa_p": false,
+    "compile": false
+  },
+  "optimizer_defaults": {},
+  "sample_definition_file_name": "training_samples/samples.json",
+  "samples": null,
+  "sample_after": 10,
+  "sample_after_unit": "MINUTE",
+  "sample_skip_first": 0,
+  "sample_image_format": "JPG",
+  "sample_video_format": "MP4",
+  "sample_audio_format": "MP3",
+  "samples_to_tensorboard": true,
+  "non_ema_sampling": true,
+  "backup_after": 30,
+  "backup_after_unit": "MINUTE",
+  "rolling_backup": true,
+  "rolling_backup_count": 2,
+  "backup_before_save": true,
+  "save_every": 0,
+  "save_every_unit": "NEVER",
+  "save_skip_first": 0,
+  "save_filename_prefix": ""
+}

training-scripts/onetrainer/fk9_template.json

@@ -0,0 +1,530 @@
+{
+  "__version": 10,
+  "training_method": "LORA",
+  "model_type": "FLUX_2",
+  "debug_mode": false,
+  "debug_dir": "debug",
+  "workspace_dir": "workspace/run",
+  "cache_dir": "workspace-cache/run",
+  "tensorboard": true,
+  "tensorboard_expose": false,
+  "tensorboard_always_on": false,
+  "tensorboard_port": 6006,
+  "validation": false,
+  "validate_after": 1,
+  "validate_after_unit": "EPOCH",
+  "continue_last_backup": false,
+  "prevent_overwrites": false,
+  "include_train_config": "NONE",
+  "multi_gpu": false,
+  "device_indexes": "",
+  "gradient_reduce_precision": "FLOAT_32_STOCHASTIC",
+  "fused_gradient_reduce": true,
+  "async_gradient_reduce": true,
+  "async_gradient_reduce_buffer": 100,
+  "base_model_name": "black-forest-labs/FLUX.2-klein-base-9B",
+  "output_dtype": "BFLOAT_16",
+  "output_model_format": "SAFETENSORS",
+  "output_model_destination": "models/fk9_[person_to_train]_v2_onetrainer.safetensors",
+  "gradient_checkpointing": "ON",
+  "enable_async_offloading": true,
+  "enable_activation_offloading": true,
+  "layer_offload_fraction": 0.0,
+  "force_circular_padding": false,
+  "compile": true,
+  "concept_file_name": "training_concepts/concepts.json",
+  "concepts": [
+    {
+      "__version": 2,
+      "image": {
+        "__version": 0,
+        "enable_crop_jitter": true,
+        "enable_random_flip": false,
+        "enable_fixed_flip": false,
+        "enable_random_rotate": false,
+        "enable_fixed_rotate": false,
+        "random_rotate_max_angle": 0.0,
+        "enable_random_brightness": false,
+        "enable_fixed_brightness": false,
+        "random_brightness_max_strength": 0.0,
+        "enable_random_contrast": false,
+        "enable_fixed_contrast": false,
+        "random_contrast_max_strength": 0.0,
+        "enable_random_saturation": false,
+        "enable_fixed_saturation": false,
+        "random_saturation_max_strength": 0.0,
+        "enable_random_hue": false,
+        "enable_fixed_hue": false,
+        "random_hue_max_strength": 0.0,
+        "enable_resolution_override": false,
+        "resolution_override": "512",
+        "enable_random_circular_mask_shrink": false,
+        "enable_random_mask_rotate_crop": false
+      },
+      "text": {
+        "__version": 0,
+        "prompt_source": "sample",
+        "prompt_path": "",
+        "enable_tag_shuffling": false,
+        "tag_delimiter": ",",
+        "keep_tags_count": 1,
+        "tag_dropout_enable": false,
+        "tag_dropout_mode": "FULL",
+        "tag_dropout_probability": 0.0,
+        "tag_dropout_special_tags_mode": "NONE",
+        "tag_dropout_special_tags": "",
+        "tag_dropout_special_tags_regex": false,
+        "caps_randomize_enable": false,
+        "caps_randomize_mode": "capslock, title, first, random",
+        "caps_randomize_probability": 0.0,
+        "caps_randomize_lowercase": false
+      },
+      "name": "",
+      "path": "C:/Development/ai-toolkit/datasets/[person_to_train]",
+      "seed": -469069486,
+      "enabled": true,
+      "type": "STANDARD",
+      "include_subdirectories": false,
+      "image_variations": 1,
+      "text_variations": 1,
+      "balancing": 1.0,
+      "balancing_strategy": "REPEATS",
+      "loss_weight": 1.0,
+      "concept_stats": {}
+    }
+  ],
+  "aspect_ratio_bucketing": true,
+  "latent_caching": true,
+  "clear_cache_before_training": true,
+  "learning_rate_scheduler": "CONSTANT",
+  "custom_learning_rate_scheduler": null,
+  "scheduler_params": [],
+  "learning_rate": 3e-5,
+  "learning_rate_warmup_steps": 200.0,
+  "learning_rate_cycles": 1.0,
+  "learning_rate_min_factor": 0.0,
+  "epochs": 100,
+  "batch_size": 2,
+  "gradient_accumulation_steps": 1,
+  "ema": "OFF",
+  "ema_decay": 0.999,
+  "ema_update_step_interval": 5,
+  "dataloader_threads": 1,
+  "train_device": "cuda",
+  "temp_device": "cpu",
+  "train_dtype": "BFLOAT_16",
+  "fallback_train_dtype": "BFLOAT_16",
+  "enable_autocast_cache": true,
+  "only_cache": false,
+  "resolution": "512",
+  "frames": "25",
+  "mse_strength": 1.0,
+  "mae_strength": 0.0,
+  "log_cosh_strength": 0.0,
+  "huber_strength": 0.0,
+  "huber_delta": 1.0,
+  "vb_loss_strength": 1.0,
+  "loss_weight_fn": "CONSTANT",
+  "loss_weight_strength": 5.0,
+  "dropout_probability": 0.0,
+  "loss_scaler": "NONE",
+  "learning_rate_scaler": "NONE",
+  "clip_grad_norm": 1.0,
+  "offset_noise_weight": 0.0,
+  "generalized_offset_noise": false,
+  "perturbation_noise_weight": 0.0,
+  "rescale_noise_scheduler_to_zero_terminal_snr": false,
+  "force_v_prediction": false,
+  "force_epsilon_prediction": false,
+  "min_noising_strength": 0.0,
+  "max_noising_strength": 1.0,
+  "timestep_distribution": "LOGIT_NORMAL",
+  "noising_weight": 0.0,
+  "noising_bias": 0.0,
+  "timestep_shift": 1.0,
+  "dynamic_timestep_shifting": false,
+  "unet": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 0,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "prior": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 0,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "transformer": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 0,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "INT_W8A8",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "quantization": {
+    "__version": 0,
+    "layer_filter": "transformer_block",
+    "layer_filter_preset": "blocks",
+    "layer_filter_regex": false,
+    "svd_dtype": "NONE",
+    "svd_rank": 16,
+    "cache_dir": "workspace-cache/run/quantization"
+  },
+  "text_encoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": false,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_8",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_layer_skip": 0,
+  "text_encoder_sequence_length": 512,
+  "text_encoder_2": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_2_layer_skip": 0,
+  "text_encoder_2_sequence_length": 77,
+  "text_encoder_3": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_3_layer_skip": 0,
+  "text_encoder_4": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": 30,
+    "stop_training_after_unit": "EPOCH",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "text_encoder_4_layer_skip": 0,
+  "vae": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "effnet_encoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "decoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "decoder_text_encoder": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "decoder_vqgan": {
+    "__version": 0,
+    "model_name": "",
+    "include": true,
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "learning_rate": null,
+    "weight_dtype": "FLOAT_32",
+    "dropout_probability": 0.0,
+    "train_embedding": true,
+    "attention_mask": false,
+    "guidance_scale": 1.0
+  },
+  "masked_training": false,
+  "unmasked_probability": 0.1,
+  "unmasked_weight": 0.1,
+  "normalize_masked_area_loss": false,
+  "masked_prior_preservation_weight": 0.0,
+  "custom_conditioning_image": false,
+  "layer_filter": "transformer_blocks",
+  "layer_filter_preset": "transformer_block",
+  "layer_filter_regex": false,
+  "embedding_learning_rate": null,
+  "preserve_embedding_norm": false,
+  "embedding": {
+    "__version": 0,
+    "uuid": "32cefb93-f7d2-4083-be65-f367c7ebccb5",
+    "model_name": "",
+    "placeholder": "<embedding>",
+    "train": true,
+    "stop_training_after": null,
+    "stop_training_after_unit": "NEVER",
+    "token_count": 1,
+    "initial_embedding_text": "*",
+    "is_output_embedding": false
+  },
+  "additional_embeddings": [],
+  "embedding_weight_dtype": "FLOAT_32",
+  "cloud": {
+    "__version": 0,
+    "enabled": false,
+    "type": "RUNPOD",
+    "file_sync": "NATIVE_SCP",
+    "create": true,
+    "name": "OneTrainer",
+    "tensorboard_tunnel": true,
+    "sub_type": "",
+    "gpu_type": "",
+    "volume_size": 100,
+    "min_download": 0,
+    "remote_dir": "/workspace",
+    "huggingface_cache_dir": "/workspace/huggingface_cache",
+    "onetrainer_dir": "/workspace/OneTrainer",
+    "install_cmd": "git clone https://github.com/Nerogar/OneTrainer",
+    "install_onetrainer": true,
+    "update_onetrainer": true,
+    "detach_trainer": false,
+    "run_id": "job1",
+    "download_samples": true,
+    "download_output_model": true,
+    "download_saves": true,
+    "download_backups": false,
+    "download_tensorboard": false,
+    "delete_workspace": false,
+    "on_finish": "NONE",
+    "on_error": "NONE",
+    "on_detached_finish": "NONE",
+    "on_detached_error": "NONE"
+  },
+  "peft_type": "LORA",
+  "lora_model_name": "",
+  "lora_rank": 16,
+  "lora_alpha": 1.0,
+  "lora_decompose": false,
+  "lora_decompose_norm_epsilon": true,
+  "lora_decompose_output_axis": false,
+  "lora_weight_dtype": "FLOAT_32",
+  "bundle_additional_embeddings": true,
+  "oft_block_size": 32,
+  "oft_coft": false,
+  "coft_eps": 0.0001,
+  "oft_block_share": false,
+  "optimizer": {
+    "__version": 0,
+    "optimizer": "ADAMW",
+    "adam_w_mode": false,
+    "alpha": null,
+    "amsgrad": false,
+    "beta1": 0.9,
+    "beta2": 0.999,
+    "beta3": null,
+    "bias_correction": false,
+    "block_wise": false,
+    "capturable": false,
+    "centered": false,
+    "clip_threshold": null,
+    "d0": null,
+    "d_coef": null,
+    "dampening": null,
+    "decay_rate": null,
+    "decouple": false,
+    "differentiable": false,
+    "eps": 1e-8,
+    "eps2": null,
+    "foreach": false,
+    "fsdp_in_use": false,
+    "fused": true,
+    "fused_back_pass": false,
+    "growth_rate": null,
+    "initial_accumulator_value": null,
+    "initial_accumulator": null,
+    "is_paged": false,
+    "log_every": null,
+    "lr_decay": null,
+    "max_unorm": null,
+    "maximize": false,
+    "min_8bit_size": null,
+    "quant_block_size": null,
+    "momentum": null,
+    "nesterov": false,
+    "no_prox": false,
+    "optim_bits": null,
+    "percentile_clipping": null,
+    "r": null,
+    "relative_step": false,
+    "safeguard_warmup": false,
+    "scale_parameter": false,
+    "stochastic_rounding": false,
+    "use_bias_correction": false,
+    "use_triton": false,
+    "warmup_init": false,
+    "weight_decay": 0.01,
+    "weight_lr_power": null,
+    "decoupled_decay": false,
+    "fixed_decay": false,
+    "rectify": false,
+    "degenerated_to_sgd": false,
+    "k": null,
+    "xi": null,
+    "n_sma_threshold": null,
+    "ams_bound": false,
+    "adanorm": false,
+    "adam_debias": false,
+    "slice_p": null,
+    "cautious": false,
+    "weight_decay_by_lr": true,
+    "prodigy_steps": null,
+    "use_speed": false,
+    "split_groups": true,
+    "split_groups_mean": true,
+    "factored": true,
+    "factored_fp32": true,
+    "use_stableadamw": true,
+    "use_cautious": false,
+    "use_grams": false,
+    "use_adopt": false,
+    "d_limiter": true,
+    "use_schedulefree": true,
+    "use_orthograd": false,
+    "nnmf_factor": false,
+    "orthogonal_gradient": false,
+    "use_atan2": false,
+    "use_AdEMAMix": false,
+    "beta3_ema": null,
+    "alpha_grad": null,
+    "beta1_warmup": null,
+    "min_beta1": null,
+    "Simplified_AdEMAMix": false,
+    "cautious_mask": false,
+    "grams_moment": false,
+    "kourkoutas_beta": false,
+    "k_warmup_steps": null,
+    "schedulefree_c": null,
+    "ns_steps": null,
+    "MuonWithAuxAdam": false,
+    "muon_hidden_layers": null,
+    "muon_adam_regex": false,
+    "muon_adam_lr": null,
+    "muon_te1_adam_lr": null,
+    "muon_te2_adam_lr": null,
+    "muon_adam_config": {},
+    "rms_rescaling": true,
+    "normuon_variant": false,
+    "beta2_normuon": null,
+    "normuon_eps": null,
+    "low_rank_ortho": false,
+    "ortho_rank": null,
+    "accelerated_ns": false,
+    "cautious_wd": false,
+    "approx_mars": false,
+    "kappa_p": null,
+    "auto_kappa_p": false,
+    "compile": false
+  },
+  "optimizer_defaults": {},
+  "sample_definition_file_name": "training_samples/samples.json",
+  "samples": null,
+  "sample_after": 10,
+  "sample_after_unit": "MINUTE",
+  "sample_skip_first": 0,
+  "sample_image_format": "JPG",
+  "sample_video_format": "MP4",
+  "sample_audio_format": "MP3",
+  "samples_to_tensorboard": true,
+  "non_ema_sampling": true,
+  "backup_after": 30,
+  "backup_after_unit": "MINUTE",
+  "rolling_backup": true,
+  "rolling_backup_count": 2,
+  "backup_before_save": true,
+  "save_every": 0,
+  "save_every_unit": "NEVER",
+  "save_skip_first": 0,
+  "save_filename_prefix": ""
+}