transformer-deploy VS stable-diffusion-webui

For 2), I am aware of a few options. Triton inference server is an obvious one as is the ‘transformer-deploy’ version from LDS. My only reservation here is that they require the model compilation or are architecture specific. I am aware of others like Bento, Ray serving and TorchServe. Ideally I would have something that allows any (PyTorch model) to be used without the extra compilation effort (or at least optionally) and has some convenience things like ease of use, easy to deploy, easy to host multiple models and can perform some dynamic batching. Anyway, I am really interested to hear people's experience here as I know there are now quite a few options! Any help is appreciated! Disclaimer - I have no affiliation or are connected in any way with the libraries or companies listed here. These are just the ones I know of. Thanks in advance.

We work for Lefebvre Sarrut, a leading European legal publisher. Several of our products include transformer models in latency sensitive scenarios (search, content recommendation). So far, ONNX Runtime and TensorRT served us well, and we learned interesting patterns along the way that we shared with the community through an open-source library called transformer-deploy. However, recent changes in our environment made our needs evolve:

here you will find a notebook for T5 on GPU with some tricks to make it fast: https://github.com/ELS-RD/transformer-deploy/blob/main/demo/generative-model/t5.ipynb

Check the notebook https://github.com/ELS-RD/transformer-deploy/blob/main/demo/TorchDynamo/benchmark.ipynb for detailed results, but what we will keep in mind:

notebook: https://github.com/ELS-RD/transformer-deploy/blob/main/demo/generative-model/t5.ipynb (Onnx Runtime only)

Regarding CPU inference, quantization is very easy, and supported by Transformer-deploy , however performance on transformer are very low outside corner cases (like no batch, very short sequence and distilled model), and last Intel generation CPU based instance like C6 or M6 on AWS are quite expensive compared to a cheap GPU like Nvidia T4, to say it otherwise, on transformer, until you are ok with slow inference and takes a small instance (for a PoC for instance), CPU inference is probably not a good idea.

The end to end tutorial: https://github.com/ELS-RD/transformer-deploy/blob/main/demo/quantization_end_to_end.ipynb

Want to try it 👉 https://github.com/ELS-RD/transformer-deploy

* LLaVA model: I'll add more documentation. You are right Llava could not generate images. For image generation I don't have immediate plans, but checkout these projects for local image generation.

- https://diffusionbee.com/

- https://github.com/comfyanonymous/ComfyUI

- https://github.com/AUTOMATIC1111/stable-diffusion-webui

I would love to be able to have a native stable diffusion experience, my rx 580 takes 30s to generate a single image. But it does work after following https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki...

I got this up and running on my windows machine in short order and I don't even know what stable diffusion is.

But again, it would be nice to have first class support to locally participate in the fun.

In Auto1111, that just uses Image.blend. :)

https://github.com/AUTOMATIC1111/stable-diffusion-webui/blob...

I used SDXL in automatic1111 webui for both images. Now that I think about it, the procedure I described was how I made this one, but the one that looks like an illustration was done in two steps. I used the canny ControlNet as I said for the outer part of the logo to preserve the shape of the fonts, but I had to turn it off for the boot to give SDXL leeway to add detail and make it look more like a boot.

That said, please do learn coding and don't get discouraged when somebody says to learn PyTorch or recommends using a Jupiter notebook with no further information on how to translate the skill into images. I would highly recommend some short term goals. Get your feet wet by taking apart the UIs. The comfy API documentation is here and the A1111 API documentation is here. There is a difference in completeness, welcome to programming. Writing nodes or plugins is also a good way to jump into this world. Custom wildcard logic might be very attractive to you if you aren't the type that want to deal with a nested file structure to simulate logic.

Possibly Unprompted? https://github.com/AUTOMATIC1111/stable-diffusion-webui/discussions/8094

version: v1.6.1  •  python: 3.10.13  •  torch: 2.0.1+cu118  •  xformers: 0.0.20  •  gradio: 3.41.2  •  checkpoint: 6e8d4871f8