calcuis/ace-gguf

gguf quantized ace-step-v1-3.5b

• base model from ace-step
• as need to keep some key tensors (in f32 status) to make it works; file size might not decrease that much; but load faster than the model from checkpoint itself a lot
• umt5_base tokenizer has problem, before it was resolved, take the encoder inside safetensors checkpoint instead recently
• dry running

setup (once)

• drag gguf to > ./ComfyUI/models/diffusion_models
• drag safetensors to > ./ComfyUI/models/checkpoints
• drag pig to > ./ComfyUI/models/vae

extra: fp8/16/32 scaled stable-audio-open-1.0 with gguf quantized t5_base encoder

• base model from stabilityai
• dry running

setup (once)

• drag t5-base to > ./ComfyUI/models/text_encoders
• drag safetensors to > ./ComfyUI/models/checkpoints
• drag pig to > ./ComfyUI/models/vae

reference

• comfyui from comfyanonymous
• pig architecture from connector
• gguf-node (pypi|repo|pack)