Yet another Hacker News

Qwen-Image-2.1: Compact, efficient, and unified image creation

105 points jmillikin 1 hour ago 33 comments

fishfasell 56 minutes ago | parent

The capabilities of local LLM text-to-image is honestly pretty damn impressive. IMO, I think local image generation is currently ahead of local code generation. I can get an image in seconds locally with the quality being way higher than what I'd expect from a local model. However with coding it's much slower and much less impressive. I'm sure there's a reason for this and I'm not an AI expert so I'll let the smarter folks tell me why, but that's just been my observation thus far.

victorbjorklund 54 minutes ago | parent

I mean I’m sure it’s the reverse for an artist. They would be less impressed with the image and more impressed with the code quality

gedy 26 minutes ago | parent

To generalize, LLMs are great at what you are not skilled at.

fishfasell 12 minutes ago | parent

That's a fair statement, I agree. I'm quite an abysmal artist so I could be a victim of my own bias here

mft_ 38 minutes ago | parent

I've played with diffusion models on and off since the first release of Stable Diffusion - just for amusement, without a particular goal.

Recently, I've been helping a friend's wife with some basic vector images for her sewing hobby (she has what is essentially a CNC sewing machine) and have been super-impressed with FLUX.1-Kontext, which I've been running on my Macbook Pro with mflux. Its ability to (for example) take a photo of a human or an animal and return a line drawing which is recognisably them (rather than just a generic similarish image as I've experienced with other models) is excellent.

It's an older model now, but (AIUI) has the text-handling features baked in, and in my various testing is very reliable at giving me the outputs that I want, without the randomness I've experienced previously. It's big and relatively slow (~3 mins per 512x512 image edit on my M1 Max Mac) but excellent to work with. It's also very straightforward to set up, without the harness complexity of e.g. comfyui.

hn45e7pbij 55 minutes ago | parent

Image gen you eyeball one frame and stop, code needs hundreds of tokens all correct in sequence, one bad line and the whole thing fails.

Hard_Space 49 minutes ago | parent

Interesting in the example of assembling the Cheers team how the otherwise great result genericizes Shelley Long.

hughc 24 minutes ago | parent

The result seems a pretty good representation given the source image wasn't that great. I think that Woody Harrelson comes across much worse.

TomGarden 48 minutes ago | parent

Very impressive, and kind of worrying a 7B model can have such capabilities. The implications are huge. And Qwen does no watermarking (yet) yeah?

trentor 22 minutes ago | parent

They always had a fourier space mark in their models even without the VAEs are usually pretty easy to detect.

TomGarden 21 minutes ago | parent

Ah I wasn't aware, thank you

d2kx 46 minutes ago | parent

God I love the Qwen team. Easily the most diverse set of models from all the Chinese labs. Only Gemini/DeepMind comes close.

mdp2021 42 minutes ago | parent

How do you use this model locally, similarly to using `llama-server -m <model>`?

(Of course I mean: outside direct use of Python, and in the most efficient way.)

embedding-shape 39 minutes ago | parent

Probably ComfyUI is one of the easiest way to get started with local image/video models. Or perhaps vLLM, if they have support for it already, would be something like `vllm serve <model> --omni --port 9080`

utopiah 35 minutes ago | parent

why not just as you suggested i.e. https://qwen.readthedocs.io/en/latest/run_locally/llama.cpp.... then get the result either via a UI or wget/curl it back?

mdp2021 30 minutes ago | parent

I am not sure that llama.cpp also supports image generation models.

utopiah 27 minutes ago | parent

it's multimodal, see https://github.com/ggml-org/llama.cpp/blob/master/docs/multi...

exe34 22 minutes ago | parent

Multimodal doesn't guarantee input and output.

> Currently, we support image, audio and video input.

fp64 34 minutes ago | parent

on the linked GitHub page they list support Diffusers, ComfyUI, vLLM-Omni, SGLang, and LightX2V with links to each

Iolaum 24 minutes ago | parent

There is difussion.cpp which is intended for those types of models. I set up krea-2-turbo with the help of ChatGPT 2 months ago, if you have a capable computer that's what I would suggest once it becomes supported.

spottedmarley 40 minutes ago | parent

Boy do I love waking up to find a new awesome toy from the Qwen team waiting for me to play with! Pulling it now

trains39472 37 minutes ago | parent

A 7B diffusion model can now render CJK text better than Microsoft Windows.

tomjen3 26 minutes ago | parent

Just think about how recently we got that feature in the official ChatGPT image gen. And now we have that running locally — assuming that is, I can figure out how to get this running on my Mac — blows my mind.

hgufj 34 minutes ago | parent

I am really grateful to the Chinese Labs for open sourcing their best models. If it was left to the Americans, we would be forced to pay obscene API fees to use them.

jfoster 13 minutes ago | parent

Note that the license on this has this in it:

> You shall not use the Materials for any commercial purpose without obtaining a separate commercial license from us.

It probably will be much cheaper to use than other image models, but it seems that will be up to the whims of Qwen/Alibaba rather than just being the cost of putting it in a cloud provider.

https://github.com/QwenLM/Qwen-Image-2.1/blob/main/LICENSE

jfoster 25 minutes ago | parent

A lot of the previous Qwen models seem to have used Apache licenses, among others:

https://en.wikipedia.org/wiki/Qwen#List_of_models

Unfortunately, it looks like this model is using a much more restrictive license:

https://github.com/QwenLM/Qwen-Image-2.1/blob/main/LICENSE

gregoriol 22 minutes ago | parent

Was going to post about this: the last image models with Apache 2.0 license seem to be from 2025, recent Qwen models are "non-commercial use".

trentor 23 minutes ago | parent

They finally fixed their VAE. It really held back their models over the last 2 years.

BlackGlory 17 minutes ago | parent

They tried hard, but still couldn't remove the piss filter from the stolen dataset.

reedf1 10 minutes ago | parent

Context?

JimDabell 6 minutes ago | parent

People say ChatGPT generates images with a yellow tint. The person you are replying to is suggesting that these images have a yellow tint and therefore this model is distilled from ChatGPT.

gunalx 8 minutes ago | parent

Its happy to see a new open image model from qwen. But the license is a let down. And it dosent even beat their closed qwen3 image wich is already a bit old.