Int8 qwen3 0.6b, for Anima.I heard Civitai added ComfyUI support. This file is here so it can be loaded by Civitai's ComfyUI cloud/backend (and should be loaded faster for the potato server). Otherwise you can't load the TE for Anima.EDIT: And the ComfyUI backend SUCKS. Loading models is extremely difficult. The queuing time is very long. Decoding vae always gets stuck at 0% and the whole workflow costs ~20 buzz because of it (which should be ~4 for Anima).No wonder you can't even find the relevant entry point for ComfyUI backend on the website. And development seems to have been put on hold.This page will be self-destructed in probably 1 week.

qwen3 0.6b int8

qwen_3_06b_base_int8convrot_mixed.safetensors

Int8convrot, converted from <a target="_blank" rel="ugc" href="https://huggingface.co/circlestone-labs/Anima/blob/main/split_files/diffusion_models/anima-base-v1.0.safetensors">https://huggingface.co/circlestone-labs/Anima/blob/main/split_files/diffusion_models/anima-base-v1.0.safetensors</a><hr />Int8convrot qwen3 0.6b, mainly for fun. Because why not.qwen3 0.6b (b): 200MiB smaller than previous version. Embeddings also quantized to int8.

Files with SHA256: ab973b953301bc6402dc0fc2832edef72db3169cb39563f8b6e82fa5987fa1cc