README.md

Attention:

the dir under this root, is defined used by clitools/examples/.sh scripts.

• [io-test] dir: contain clitools input & output, by using clitools/examples/.sh scripts.
• [sd-base-model] dir: you can found target model from HuggingFace, and clone it to this dir, clitools/examples/.sh scripts will rely on it.
• [sd-dictionary] dir: put Tokenizer reference Vocabulary-Dictionary under here.

so, if you run command tools on you own, just be careful about the path setting.

but if you want to do it by self, there's 2 ways to do it.

• Method-1: by simply executing auto_script auto_prepare_sd_models.sh:

cd ./sd
bash auto_prepare_sd_models.sh

• Method-2: or download SD manually from HuggingFace-SD-Official, there's an example for sd-turbo official. However, not all models have an official .onnx format provided.:

cd ./sd/sd-base-model/
git clone https://huggingface.co/runwayml/stable-diffusion-v1-5 -b onnx onnx-official-sd-v15/

• Method-3: or download SD manually from HuggingFace-Our-Converted, there's an example for sd-turbo official:

cd ./sd/sd-base-model/
git clone https://huggingface.co/Windsander/onnx-sd-v15 onnx-sd-v15/

Bash Tools:

• [auto_ready_model] at: auto_prepare_sd_models.sh
  This script tool helps you quickly prepare an environment for self-hosted HuggingFace models,
  which needs converting from <model>.sfaetensors to <model>.onnx format.
  Making them compatible with more platforms and tools, by using optimum-cli from readied env.
  You can find more information about optimum-cli from: Optimum-Official: ONNX Runtime If you want to prepare convert environment & sd-series at same time, then just bash this script.
  The models we support, will download by script completely.