update features
This commit is contained in:
parent
3622ac4bfb
commit
01d8aabc42
14
README.md
14
README.md
|
@ -23,6 +23,7 @@ details](https://github.com/ssube/onnx-web/blob/main/docs/user-guide.md).
|
||||||
|
|
||||||
This is an incomplete list of new and interesting features, with links to the user guide:
|
This is an incomplete list of new and interesting features, with links to the user guide:
|
||||||
|
|
||||||
|
- SDXL support
|
||||||
- hardware acceleration on both AMD and Nvidia
|
- hardware acceleration on both AMD and Nvidia
|
||||||
- tested on CUDA, DirectML, and ROCm
|
- tested on CUDA, DirectML, and ROCm
|
||||||
- [half-precision support for low-memory GPUs](docs/user-guide.md#optimizing-models-for-lower-memory-usage) on both
|
- [half-precision support for low-memory GPUs](docs/user-guide.md#optimizing-models-for-lower-memory-usage) on both
|
||||||
|
@ -37,6 +38,7 @@ This is an incomplete list of new and interesting features, with links to the us
|
||||||
- [txt2img](docs/user-guide.md#txt2img-tab)
|
- [txt2img](docs/user-guide.md#txt2img-tab)
|
||||||
- [img2img](docs/user-guide.md#img2img-tab)
|
- [img2img](docs/user-guide.md#img2img-tab)
|
||||||
- [inpainting](docs/user-guide.md#inpaint-tab), with mask drawing and upload
|
- [inpainting](docs/user-guide.md#inpaint-tab), with mask drawing and upload
|
||||||
|
- [panorama](docs/user-guide.md#panorama-pipeline)
|
||||||
- [upscaling](docs/user-guide.md#upscale-tab), with ONNX acceleration
|
- [upscaling](docs/user-guide.md#upscale-tab), with ONNX acceleration
|
||||||
- [add and use your own models](docs/user-guide.md#adding-your-own-models)
|
- [add and use your own models](docs/user-guide.md#adding-your-own-models)
|
||||||
- [convert models from diffusers and SD checkpoints](docs/converting-models.md)
|
- [convert models from diffusers and SD checkpoints](docs/converting-models.md)
|
||||||
|
@ -45,20 +47,24 @@ This is an incomplete list of new and interesting features, with links to the us
|
||||||
- [permanent and prompt-based blending](docs/user-guide.md#permanently-blending-additional-networks)
|
- [permanent and prompt-based blending](docs/user-guide.md#permanently-blending-additional-networks)
|
||||||
- [supports LoRA and LyCORIS weights](docs/user-guide.md#lora-tokens)
|
- [supports LoRA and LyCORIS weights](docs/user-guide.md#lora-tokens)
|
||||||
- [supports Textual Inversion concepts and embeddings](docs/user-guide.md#textual-inversion-tokens)
|
- [supports Textual Inversion concepts and embeddings](docs/user-guide.md#textual-inversion-tokens)
|
||||||
|
- each layer of the embeddings can be controlled and used individually
|
||||||
- ControlNet
|
- ControlNet
|
||||||
- image filters for edge detection and other methods
|
- image filters for edge detection and other methods
|
||||||
- with ONNX acceleration
|
- with ONNX acceleration
|
||||||
- highres mode
|
- highres mode
|
||||||
- runs img2img on the results of the other pipelines
|
- runs img2img on the results of the other pipelines
|
||||||
- multiple iterations can produce 8k images and larger
|
- multiple iterations can produce 8k images and larger
|
||||||
|
- [multi-stage](docs/user-guide.md#prompt-stages) and [region prompts](docs/user-guide.md#region-tokens)
|
||||||
|
- combine multiple prompts in the same image
|
||||||
|
- provide prompts for different areas in the image and blend them together
|
||||||
|
- change the prompt for highres mode and refine details without recursion
|
||||||
- infinite prompt length
|
- infinite prompt length
|
||||||
- [with long prompt weighting](docs/user-guide.md#long-prompt-weighting)
|
- [with long prompt weighting](docs/user-guide.md#long-prompt-weighting)
|
||||||
- expand and control Textual Inversions per-layer
|
|
||||||
- [image blending mode](docs/user-guide.md#blend-tab)
|
- [image blending mode](docs/user-guide.md#blend-tab)
|
||||||
- combine images from history
|
- combine images from history
|
||||||
- upscaling and face correction
|
- upscaling and correction
|
||||||
- upscaling with Real ESRGAN or Stable Diffusion
|
- upscaling with Real ESRGAN, SwinIR, and Stable Diffusion
|
||||||
- face correction with CodeFormer or GFPGAN
|
- face correction with CodeFormer and GFPGAN
|
||||||
- [API server can be run remotely](docs/server-admin.md)
|
- [API server can be run remotely](docs/server-admin.md)
|
||||||
- REST API can be served over HTTPS or HTTP
|
- REST API can be served over HTTPS or HTTP
|
||||||
- background processing for all image pipelines
|
- background processing for all image pipelines
|
||||||
|
|
|
@ -460,7 +460,19 @@ This makes your prompt less specific and some models have been trained to work b
|
||||||
|
|
||||||
#### Prompt stages
|
#### Prompt stages
|
||||||
|
|
||||||
TODO: explain `first stage || hires prompt` syntax
|
You can provide a different prompt for the highres and upscaling stages of an image using prompt stages. Each stage
|
||||||
|
of a prompt is separated by `||` and can include its own LoRAs, embeddings, and regions. If you are using multiple
|
||||||
|
iterations of highres, each iteration can have its own prompt stage. This can help you avoid recursive body parts
|
||||||
|
and some other weird mutations that can be caused by iterating over a subject prompt.
|
||||||
|
|
||||||
|
For example, a prompt like `human being sitting on wet grass, outdoors, bright sunny day` is likely to produce many
|
||||||
|
small people mixed in with the grass when used with highres. This becomes even worse with 2+ iterations. However,
|
||||||
|
changing that prompt to `human being sitting on wet grass, outdoors, bright sunny day || outdoors, bright sunny day, detailed, intricate, HDR`
|
||||||
|
will use the second stage as the prompt for highres: `outdoors, bright sunny day, detailed, intricate, HDR`.
|
||||||
|
|
||||||
|
This allows you to add and refine details, textures, and even the style of the image during the highres pass.
|
||||||
|
|
||||||
|
Prompt stages are only used during upscaling if you are using the Stable Diffusion upscaling model.
|
||||||
|
|
||||||
### Long prompt weighting syntax
|
### Long prompt weighting syntax
|
||||||
|
|
||||||
|
|
Loading…
Reference in New Issue