| unsloth/Z-Image-Turbo-ComfyUI |
|
|
| A mirror of the two single-file Z-Image assets that Comfy-Org/z_image_turbo publishes |
| (https://huggingface.co/Comfy-Org/z_image_turbo). |
| Unsloth has not modified the weights: every file here is byte for byte the file of the same name in the source repository (verified on the Hub's LFS sha256). |
|
|
| Both files reach here through Z-Image-Turbo |
| (https://huggingface.co/Tongyi-MAI/Z-Image-Turbo), released by Tongyi MAI, Alibaba Group |
| under the Apache License, Version 2.0, whose full text is in LICENSE. Neither the source |
| repack nor Tongyi-MAI ships a NOTICE file, so there is nothing to propagate under |
| Apache-2.0 section 4(d); the attributions below are the ones section 4(c) says to retain. |
|
|
| split_files/vae/ae.safetensors |
| The FLUX.1 autoencoder, which Z-Image reuses. Copyright Black Forest Labs Inc., |
| licensed under the Apache License, Version 2.0. This is not an inference: the file's |
| own safetensors metadata says so, and is preserved intact by this byte-identical |
| mirror |
| modelspec.title = "Flux.1 Autoencoder" |
| modelspec.author = "Black Forest Labs" |
| modelspec.license = "Apache License 2.0" |
| modelspec.implementation = https://github.com/black-forest-labs/flux |
| It is the same autoencoder black-forest-labs/FLUX.1-schnell ships as ae.safetensors |
| under Apache-2.0. |
|
|
| split_files/text_encoders/qwen_3_4b.safetensors |
| The conditioner Z-Image pairs with, derived from Qwen3-4B |
| (https://huggingface.co/Qwen/Qwen3-4B, Apache-2.0, Copyright Alibaba Cloud) as |
| modified and redistributed by Tongyi MAI inside Z-Image-Turbo. |
|
|
| Changes relative to Tongyi-MAI/Z-Image-Turbo (Apache-2.0 section 4(b)): the components are |
| extracted into standalone safetensors files with ComfyUI key names. That repack is |
| Comfy-Org's work, not Unsloth's. Not an official Alibaba Group, Black Forest Labs or |
| ComfyUI product, and not endorsed by any of them. |
|
|
