Qwen Image 2.1 Turbo INT4W4A8 (int8 file) mirror

A private-use mirror of the Qwen Image 2.1 Turbo checkpoint published on Civitai (download version 3394623), kept here so rented GPU instances can pull it at full speed; the Civitai CDN throttles a single host to 11-33 MB/s after a few GB.

  • File: qwenImage21TurboINT4W4A8_turboINT8.safetensors (7,256,783,064 bytes, sha256 85d3244148cc294581844d310a7338fed26d056e470666ec84cf1525b876fdab)
  • Runs in ComfyUI through the DaSiWa image workflow chain, 8 steps, CFG 1, sampler euler, scheduler simple
  • Text encoder and VAE are not mirrored here: they come from Comfy-Org/Qwen-Image-2.1 (text_encoders/qwen3vl_8b_int8_convrot.safetensors, vae/qwen_image_2.1_vae_bf16.safetensors)

All credit for the checkpoint goes to its Civitai author; this repository only re-hosts the file under the licence it was published with.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for releaseworthy/qwen-image-2.1-turbo-int8

Base model

Qwen/Qwen-Image
Finetuned
(109)
this model