dt-qwen-image-2.1

Qwen Image 2.1 weights in the Draw Things (s4nnc .ckpt) format, mirrored unmodified from the Draw Things model CDN for use by JoyFusion's on-device inference. The conversion and quantization were done by the Draw Things project; the files here are byte-identical to the upstream ones.

File	Variant	Size	SHA-256
qwen_image_2.1_q8p.ckpt	8-bit	7,387,840,512	30b79864b1b307d066b0ffb4349f2e4ba41f15b1476672876a5d68cb9f3aa247
qwen_image_2.1_q6p.ckpt	6-bit	5,643,010,048	3ee834e2c41302f6d5b4413d1056bd795132dd1bcbc145f1fbd298cfb020c6c7
qwen_image_2.1_i4x.ckpt	4-bit S	4,120,993,792	9663d3c1b1b7b7beff309c3916d7469b50d077aef49de69ae7af7dd6a4d7aa78
qwen_image_2.1_i8x.ckpt	8-bit S	7,174,447,104	28a7b9509a9d136fa61b94a4fbe5c042eecac3141d3d0cb9d049a777b82d4b70
qwen_image_2.1_vae_f16.ckpt	VAE	662,470,656	2f76345874aa8e1f36e200b9fc6a349e2aaf9024ec4513a41a09368408d7e859

The shared text encoder qwen_3_vl_8b_instruct_i8x.ckpt (Qwen3-VL-8B-Instruct, Apache-2.0) lives in JoyFusionAI/dt-components. It contains only the text model; reference-image editing needs the f16 checkpoint with the vision tower, which is not mirrored here.

License

Qwen Image 2.1 is licensed under the Qwen RESEARCH LICENSE AGREEMENT (see LICENSE), which permits use, reproduction and distribution for non-commercial (research or evaluation) purposes only. Commercial use requires a separate license from Hangzhou Tongyi Laboratory Technology Co., Ltd. These files are unmodified copies.