This is a INT8 (row-wise) quantization made with "convert_to_quant" script with --int8 --scaling_mode row --anima --low-memory flags. The purpose of this is to squeeze as much vram as possible on low-end GPUs (6gigs) like mine, at least i noticed significant speed up during loading and generating images, so i want to share this with people. 8-16 steps CFG 1-2 Sampler - Euler/ER_SDE Scheduler - Simple/Beta57 You can also use int8 qwen3_06b text encoder provided by <a target="_blank" rel="ugc" href="https://civarchive.com/user/reakaakasky">reakaakasky </a> <a target="_blank" rel="ugc" href="https://civarchive.com/models/2760490/anima-int8?modelVersionId=3109193">https://civarchive.com/models/2760490/anima-int8?modelVersionId=3109193</a> to squeeze more memory. AnimaTurbo 1.0 is identical to <a target="_blank" rel="ugc" href="https://civarchive.com/models/2458426/anima?modelVersionId=3108589">https://civarchive.com/models/2458426/anima?modelVersionId=3108589</a> AnimaYume 1.0 have "Expressive" Turbo Lora by <a target="_blank" rel="ugc" href="https://civarchive.com/user/kai">ka</a><a target="_blank" rel="ugc" href="https://civarchive.com/models/2458426/anima?modelVersionId=3108589">i</a> Thanks <a target="_blank" rel="ugc" href="https://civarchive.com/user/circlestone_labs">circlestone_labs</a> for amazing Anima model, <a target="_blank" rel="ugc" href="https://civarchive.com/user/duongve13112002">duongve13112002</a> for great AnimaYume finetune, <a target="_blank" rel="ugc" href="https://civarchive.com/user/kai">kai</a> for excellent revision of turbo lora.

Files with SHA256: e0c9acbcde1362d9a5239b8d464a40ee3c008f8efee66f6d8819362607fe6d58