Did you know there are other dialog agents like ChatGPT? 🤔
And what if I told you the secret sauce is IFT, RLHF, CoT, and SFT 🤯
We explain each of these terms and why they are relevant to ChatGPT by comparing with 4 other dialog agents.
Check our blog: huggingface.co/blog/dialog-ag
Follow
Click to Follow anton_lozhkov
Anton Lozhkov
@anton_lozhkov
Joined January 2015
Anton Lozhkov’s Tweets
📢🌀 Released my "Image Mixer" model!
Mix up concepts in multiple images and words to generate novel pictures!
Try it on spaces here: huggingface.co/spaces/lambdal
32
197
974
Show this thread
Hey, we're opening an internship for spring/summer to build pretraining datasets for the net generation of LLMs. If you're as excited about that as I am, you can apply at apply.workable.com/huggingface/j/ :)
6
61
240
Show this thread
For the last months , I worked on transformers and... graphs!
So here is a small blog, if you wonder what one could use graphs for, or how to machine learn on them 🔎
(Spoiler: they are everywhere 🧬🚗✍️)
26
462
2,229
Show this thread
diffusers v0.10.0 is here packed with lots of goodies 🎁
🪄 Depth2Image model
🤩StableDiffusion 2.1 support
🦺 Safe Tensors integration for super fast & secure loading.
✨ Your favourite samplers from k-diffusion now integrated with v-prediction support.
🖌️ Paint-by-example
6
48
270
Show this thread
You thought the first Whisper was impressive? Try this one! 📈
Quote Tweet
Whisper large-v2 from @OpenAI, a state of the art model for speech recognition 
Same data, more training => transcribe speech with 5-10% improved accuracy 
Available now in 
Transformers
huggingface.co/openai/whisper
Same data, more training => transcribe speech with 5-10% improved accuracy
Available now in Transformers
huggingface.co/openai/whisperShow this thread
6
The new Stable Diffusion 2.1 models are now available in diffusers 🎉
Make sure to install from main to use the 768 model.
Quote Tweet
read image description
ALT
We’re happy to release Stable Diffusion, Version 2.1!
With new ways of prompting, 2.0 provided fantastic results, and 2.1 supports the new prompting style, but also brings back many of the old prompts too!
Link → stability.ai/blog/stabledif
Show this thread
24
59
Christmas came early for the team! 😍
2x4090 from our Santas and 🎅🎅
We're gonna debug so many models with this monster
3
2
34
🧨Diffusers supports Stable Diffusion 2 !
Run 's Stable Diffusion 2 with zero changes to your code using your familiar diffusers API.
Everything is supported: attention optimizations, fp16, img2image, swappable schedulers, and more🤗
5
94
460
Show this thread
🧨Diffusers 0.8.0 is out with so many new models!
-🎁Versatile Diffusion all-in-one model
-🌍Multilingual AltDiffusion
-✨Safe Latent Diffusion
-🔁CycleDiffusion
-🔥New SOTA scheduler: DPM
And much more!
Check out the release notes for more info: github.com/huggingface/di
60
255
Safetensors 0.2.4 out.
pip install -U safetensors
Load your models 150x faster on CPU (PyTorch)
And 2x faster on GPU.
And safely.
12
77
460
Show this thread
A comparison between DDIM, PLMS(a.k.a. PNDM) and DPM-Solver:
2
9
46
Show this thread
Quote Tweet
Happy to announce that our recent work "DPM-Solver" (Neurips 2022 Oral) and "DPM-Solver++" have been supported by the widely-used diffusion library @diffuserslib! An online demo for DPM-Solver with Stable-Diffusion: huggingface.co/spaces/LuCheng. Many thanks to @huggingface teams!
Show this thread
11
My version 0.7.2 was released last week - maybe you heard about how I run #StableDiffusion up to 2x faster in all hardware? 🏎️💨
What you may not be aware of is that now it is even easier to swap schedulers! Here's how to do it with the new Euler ones contributed by 🤩
1
19
109
Show this thread
Second question: what hardware to you want supported the most?
- Apple M1/M233.3%
- Nvidia GPUs (CUDA)54.8%
- Pro AMD GPUs (ROCm)2.4%
- Consumer AMD GPUs (RDNA)9.5%
42 votesFinal results
1
1
Show this thread
Hey #StableDiffusion crowd! Getting ambitious with 🧨 Diffusers platform support here, and need your feedback!
First question: which OS do you (want to) use the latest diffusion models on?
- Linux45.5%
- Mac31.2%
- Windows23.4%
77 votesFinal results
2
1
2
Show this thread
This was collaboratively generated in real-time with Stable Diffusion! 🤯🤯🤯
huggingface.co/spaces/hugging
9
54
309
The carbon footprint paper is live!! 🎉
Check it out to see how we calculated BLOOM's carbon footprint, covering all steps from the manufacturing of equipment 💻 to deployment! 🚀
10
65
187
Show this thread
They grow up so fast, our library is already using twitter 🥺
Quote Tweet
Me, the
diffusers library, am now on Twitter
This is an informal place for us to chat about the latest diffusion models advancements, news regarding the library, give support and celebrate our amazing community 
github.com/huggingface/di
diffusers library, am now on Twitter
This is an informal place for us to chat about the latest diffusion models advancements, news regarding the library, give support and celebrate our amazing community
github.com/huggingface/di6
Diffusers 0.7 + PyTorch 1.13 bring first-class support for Stable Diffusion on Apple Silicon
🐛Optimizations applied, bugs splattered*
🧠Memory efficient inference via attention slicing
♾️Continuous integration tests for M1/M2 added to diffusers
* No bugs were actually harmed
4
27
154
Show this thread
Second question: what hardware to you want supported the most?
- Apple M1/M233.3%
- Nvidia GPUs (CUDA)54.8%
- Pro AMD GPUs (ROCm)2.4%
- Consumer AMD GPUs (RDNA)9.5%
42 votesFinal results
1
1
Show this thread
Hey #StableDiffusion crowd! Getting ambitious with 🧨 Diffusers platform support here, and need your feedback!
First question: which OS do you (want to) use the latest diffusion models on?
- Linux45.5%
- Mac31.2%
- Windows23.4%
77 votesFinal results
2
1
2
Show this thread
The Diffusers releases are getting so packed with features that we've ran out of space for the release notes 😅 v0.7.0 is out!
github.com/huggingface/di
1
8
We've joined forces with to launch a new course on Diffusion Models 🧨!
Come learn the nuts and bolts of generative ML and how to train models that can create images, audio, and more 🔥
Register here 👉: eepurl.com/icSzXv
10
78
254
Attention, everyone interested in NeRFs! 🚜 Impromptu meetup tomorrow at the Hugging Face Paris office with himself! (will be recorded as well)
Quote Tweet
making it simple and casual: you can just stop by our Hugging Face's Paris office tomorrow either at 5 p.m. for a small presentation/discussion on NeRF or at 5:45 for informal drinks/meetup at the office. No RSVP needed, just show up at 5pm or 5:45pm here: g.page/124-Rue-Reaumu
Show this thread
1
2
Quote Tweet
We've been working on a multiplayer experiment for #StableDiffusion powered by @huggingface GPU Space and multiplayer API by @liveblocks, inspired by @lkwq007 SD infinity project
coming soon -
Show this thread
2:18
13.8K views
1
6
We just crossed 1,000,000 downloads of Stable Diffusion on the hub! Congrats to Robin Rombach, Patrick Esser, Compvis, , everyone who participated and the whole open-source community. Such a milestone!
huggingface.co/CompVis/stable
10
59
487
The PyLadies open-source contribution workshop is starting! Very excited to start contriuting to github.com/scikit-learn/s. Thank you for hosting! - cc
2
16
68
After just 2 weeks of the new course, our students are already making research advances in Stable Diffusion.
developed a novel yet simple modification to classifier-free guidance that gives better results (previous approach on left, new approach on right)
11
105
888
Show this thread
Spent some time this weekend fixing a nasty bug in the stable diffusion videos codebase. Interpolation quality should be much better now :)
Here’s a video for you 🦙
0:10
3.3K views
4
9
78
After some, uh, developments yesterday:
- Stable Diffusion v1-5 is out by
- Fine-tuned image decoder (VAE) out by
Magic of open source🧙 collaboration continues no matter what, here's the Best Available Stable Diffusion™ notebook: colab.research.google.com/drive/174iLnLZ
10
117
644
Show this thread
Stable Diffusion Inpainting is now available in the latest version of the diffusers library (v0.6.0) 🎉
Get started here: github.com/huggingface/di
27
115
Show this thread
We are life now!
Join us!
For registered members, please join the
Zoom: cuhk.zoom.us/j/98375654143 or YouTube: youtu.be/ICfBzXaT5xM
6
9
🤗 Diffusers 0.5.1 was released yesterday with JAX support.
We've experimentally used Diffusers + JAX + TPU in our main demo Space for the past 2 weeks. Almost 50 million images were served in 14 days (more than 3 M per day) 🤯
huggingface.co/spaces/stabili
A couple of details 👇
read image description
ALT
3
30
132
Show this thread
Superfast #stablediffusion with JAX on TPUs & GPUs
🧨diffusers 0.5.0 now supports JAX!
What does that mean? Many cool things, including that you can generate 8 images in ~8s on Colab using TPU 🤯colab.research.google.com/github/hugging
Learn how to use JAX with diffusers
7
124
590
Show this thread
The interest in the Stable Diffusion Videos project has been mind blowing 🤯
I figured it would be nice to break down the process and explain how it works under the hood (using diffusers) this new blog post. 🚀
9
135
578
Show this thread
🧨 diffusers 0.5.0 now supports JAX for super fast #stablediffusion inference on TPUs
You can generate 8 images in ~8s on Colab Free using TPU 🚀
colab.research.google.com/github/hugging
2
91
368
Show this thread