Skip to content

Commit

Permalink
Update README
Browse files Browse the repository at this point in the history
  • Loading branch information
saddam213 committed Apr 30, 2024
1 parent c8a58fe commit 6938b70
Show file tree
Hide file tree
Showing 5 changed files with 30 additions and 16 deletions.
11 changes: 8 additions & 3 deletions OnnxStack.Converter/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,12 +9,17 @@ python -m pip install -r requirements.txt

## Usage
```bash
convert.py --optimize --model_input '..\stable-diffusion-v1-5' --model_output '..\converted' --controlnet
python convert.py --model_input '..\stable-diffusion-v1-5' --controlnet
```
`--optimize` - Run the model optimization

`--model_input` - Safetensor model to convert

`--model_output` - Output for converted ONNX model (NOTE: This folder is deleted before each run)
`--model_output` - Output for converted ONNX model

`--clean` - Clear convert/optimize model cache

`--tempDir` - Directory for temp Olive files

`--only_unet` - Only convert UNET model

`--controlnet` - Create a ControlNet enabled Unet model
10 changes: 6 additions & 4 deletions OnnxStack.Converter/latent_consistency/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,15 +9,17 @@ python -m pip install -r requirements.txt

## Usage
```bash
convert.py --model_input "D:\Models\LCM_Dreamshaper_v7" --controlnet
python convert.py --model_input "D:\Models\LCM_Dreamshaper_v7" --controlnet
```

`--model_input` - Safetensor model to convert

`--model_output` - Output for converted ONNX model

`--controlnet` - Create a ControlNet enabled Unet model

`--clean` - Clear convert/optimize model cache

`--tempDir` - Directory for temp Olive files
`--tempDir` - Directory for temp Olive files

`--only_unet` - Only convert UNET model

`--controlnet` - Create a ControlNet enabled Unet model
11 changes: 8 additions & 3 deletions OnnxStack.Converter/stable_cascade/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,12 +9,17 @@ python -m pip install -r requirements.txt

## Usage
```bash
convert.py --optimize --model_input '..\stable-cascade' --image_encoder
python convert.py --model_input "D:\Models\stable-cascade" --image_encoder
```
`--optimize` - Run the model optimization

`--model_input` - Safetensor model to convert

`--model_output` - Output for converted ONNX model (NOTE: This folder is deleted before each run)
`--model_output` - Output for converted ONNX model

`--clean` - Clear convert/optimize model cache

`--tempDir` - Directory for temp Olive files

`--only_unet` - Only convert UNET model

`--image_encoder` - Convert the optional image encoder
10 changes: 6 additions & 4 deletions OnnxStack.Converter/stable_diffusion/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,15 +9,17 @@ python -m pip install -r requirements.txt

## Usage
```bash
convert.py --model_input "D:\Models\stable-diffusion-v1-5" --controlnet
python convert.py --model_input "D:\Models\stable-diffusion-v1-5" --controlnet
```

`--model_input` - Safetensor model to convert

`--model_output` - Output for converted ONNX model

`--controlnet` - Create a ControlNet enabled Unet model

`--clean` - Clear convert/optimize model cache

`--tempDir` - Directory for temp Olive files
`--tempDir` - Directory for temp Olive files

`--only_unet` - Only convert UNET model

`--controlnet` - Create a ControlNet enabled Unet model
4 changes: 2 additions & 2 deletions OnnxStack.Converter/stable_diffusion_xl/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -16,14 +16,14 @@ python convert.py --model_input "D:\Models\stable-diffusion-xl-base-1.0" --contr

`--model_output` - Output for converted ONNX model

`--controlnet` - Create a ControlNet enabled Unet model

`--clean` - Clear convert/optimize model cache

`--tempDir` - Directory for temp Olive files

`--only_unet` - Only convert UNET model

`--controlnet` - Create a ControlNet enabled Unet model

## Extra Requirements
To successfully optimize SDXL models you will need the patched `vae` from repository below otherwise you may get black image results

Expand Down

0 comments on commit 6938b70

Please sign in to comment.