XetHub is joining Hugging Face!
โข
81
Julien Chaumond PRO
julien-c
AI & ML interests
<3 ML/AI for everyone, building products to propel communities fwd
Recent Activity
liked
a model
4 days ago
Lightricks/LTX-Video
liked
a dataset
5 days ago
CohereForAI/Global-MMLU
liked
a model
5 days ago
fal/AuraFlow
Articles
Organizations
julien-c's activity
replied to
burtenshaw's
post
6 days ago
works for orgs too!
reacted to
burtenshaw's
post with ๐คโค๏ธ
6 days ago
Post
2534
People are flexing their end of year stats, so I made this app to show hub stats in a tidy design!
Thanks @Ameeeee and @jfcalvo for the feature from Argilla!
burtenshaw/recap
Thanks @Ameeeee and @jfcalvo for the feature from Argilla!
burtenshaw/recap
-
- 1 reply
yes
@DouglasMourray
imo the Local apps menu should solve part of what you're looking for:
https://www.reddit.com/r/LocalLLaMA/comments/1cwjc4n/hugging_face_adds_an_option_to_directly_launch/
Keen on hearing more feedback!
reacted to
Kseniase's
post with ๐ฅ
9 days ago
Post
2644
TL;DR: The Story of Attention's Development by
@karpathy
Origin: First proposed in 2014 by @Dzmitry Bahdanau, @KyunghyunCho , and Yoshua Bengio in Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473) . Inspired by cognitive processes and later renamed from "RNNSearch."
Key Idea: A data-dependent weighted average for pooling and communication, enabling flexible and powerful neural network connections.
Breakthrough: Bahdanau's "soft search" mechanism (softmax + weighted averaging) solved encoder-decoder bottlenecks in machine translation.
Transformer Revolution: Attention Is All You Need (1706.03762) (2017) by @ashishvaswanigoogle et al. simplified architectures by stacking attention layers, introducing multi-headed attention and positional encodings.
Legacy: Attention replaced RNNs, driving modern AI systems like ChatGPT. It emerged independently but was influenced by contemporaneous work like Alex Gravesโs Neural Turing Machines (1410.5401) and Jason Westonโs Memory Networks (1410.3916) .
Attention to history: Jรผrgen Schmidhuber claims his 1992 Fast Weight Programmers anticipated modern attention mechanisms. While conceptually similar, the term โattentionโ was absent, and thereโs no evidence it influenced Bahdanau, Cho, and Bengioโs 2014 work. Paying attention (!) to history might have brought us to genAI earlier โ but credit for the breakthrough still goes to Montreal.
Referenced Papers:
Attention Origin: Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473)
Transformers: Attention Is All You Need (1706.03762)
Alex Graves' Work: Neural Turing Machines (1410.5401), Generating Sequences With Recurrent Neural Networks (1308.0850)
Jason Weston @spermwhale 's Memory Networks (1410.3916)
Sequence to Sequence Learning with Neural Networks (1409.3215) by Ilya Sutskever ( @ilyasut ), Oriol Vinyals, Quoc V. Le
Who else deserves recognition in this groundbreaking narrative of innovation? Letโs ensure every contributor gets the credit they deserve. Leave a comment below ๐๐ป๐ค
Origin: First proposed in 2014 by @Dzmitry Bahdanau, @KyunghyunCho , and Yoshua Bengio in Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473) . Inspired by cognitive processes and later renamed from "RNNSearch."
Key Idea: A data-dependent weighted average for pooling and communication, enabling flexible and powerful neural network connections.
Breakthrough: Bahdanau's "soft search" mechanism (softmax + weighted averaging) solved encoder-decoder bottlenecks in machine translation.
Transformer Revolution: Attention Is All You Need (1706.03762) (2017) by @ashishvaswanigoogle et al. simplified architectures by stacking attention layers, introducing multi-headed attention and positional encodings.
Legacy: Attention replaced RNNs, driving modern AI systems like ChatGPT. It emerged independently but was influenced by contemporaneous work like Alex Gravesโs Neural Turing Machines (1410.5401) and Jason Westonโs Memory Networks (1410.3916) .
Attention to history: Jรผrgen Schmidhuber claims his 1992 Fast Weight Programmers anticipated modern attention mechanisms. While conceptually similar, the term โattentionโ was absent, and thereโs no evidence it influenced Bahdanau, Cho, and Bengioโs 2014 work. Paying attention (!) to history might have brought us to genAI earlier โ but credit for the breakthrough still goes to Montreal.
Referenced Papers:
Attention Origin: Neural Machine Translation by Jointly Learning to Align and Translate (1409.0473)
Transformers: Attention Is All You Need (1706.03762)
Alex Graves' Work: Neural Turing Machines (1410.5401), Generating Sequences With Recurrent Neural Networks (1308.0850)
Jason Weston @spermwhale 's Memory Networks (1410.3916)
Sequence to Sequence Learning with Neural Networks (1409.3215) by Ilya Sutskever ( @ilyasut ), Oriol Vinyals, Quoc V. Le
Who else deserves recognition in this groundbreaking narrative of innovation? Letโs ensure every contributor gets the credit they deserve. Leave a comment below ๐๐ป๐ค
- 5 replies
replied to
Duskfallcrew's
post
13 days ago
Public storage- y'all ... HF are you nuts?
i can neither confirm nor deny
reacted to
FranckAbgrall's
post with ๐
13 days ago
Post
1971
Hey!
โจ If you're using HF access tokens, we just released an overview of the permissions for fine-grained tokens by hovering over the badge on token settings page (org and user)
It will show the highest permission you've set for each entity ๐
โจ If you're using HF access tokens, we just released an overview of the permissions for fine-grained tokens by hovering over the badge on token settings page (org and user)
It will show the highest permission you've set for each entity ๐
reacted to
their
post with ๐๐ค
14 days ago
Post
7575
After some heated discussion ๐ฅ, we clarify our intent re. storage limits on the Hub
TL;DR:
- public storage is free, and (unless blatant abuse) unlimited. We do ask that you consider upgrading to PRO and/or Enterprise Hub if possible
- private storage is paid above a significant free tier (1TB if you have a paid account, 100GB otherwise)
docs: https://huggingface.co/docs/hub/storage-limits
We optimize our infrastructure continuously to scale our storage for the coming years of growth in Machine learning, to the benefit of the community ๐ฅ
cc: @reach-vb @pierric @victor and the HF team
TL;DR:
- public storage is free, and (unless blatant abuse) unlimited. We do ask that you consider upgrading to PRO and/or Enterprise Hub if possible
- private storage is paid above a significant free tier (1TB if you have a paid account, 100GB otherwise)
docs: https://huggingface.co/docs/hub/storage-limits
We optimize our infrastructure continuously to scale our storage for the coming years of growth in Machine learning, to the benefit of the community ๐ฅ
cc: @reach-vb @pierric @victor and the HF team
- 28 replies
reacted to
their
post with ๐๐คโค๏ธ๐ฅ
15 days ago
Post
7575
After some heated discussion ๐ฅ, we clarify our intent re. storage limits on the Hub
TL;DR:
- public storage is free, and (unless blatant abuse) unlimited. We do ask that you consider upgrading to PRO and/or Enterprise Hub if possible
- private storage is paid above a significant free tier (1TB if you have a paid account, 100GB otherwise)
docs: https://huggingface.co/docs/hub/storage-limits
We optimize our infrastructure continuously to scale our storage for the coming years of growth in Machine learning, to the benefit of the community ๐ฅ
cc: @reach-vb @pierric @victor and the HF team
TL;DR:
- public storage is free, and (unless blatant abuse) unlimited. We do ask that you consider upgrading to PRO and/or Enterprise Hub if possible
- private storage is paid above a significant free tier (1TB if you have a paid account, 100GB otherwise)
docs: https://huggingface.co/docs/hub/storage-limits
We optimize our infrastructure continuously to scale our storage for the coming years of growth in Machine learning, to the benefit of the community ๐ฅ
cc: @reach-vb @pierric @victor and the HF team
-
-
-
- 28 replies
replied to
Duskfallcrew's
post
15 days ago
reacted to
burtenshaw's
post with ๐ฅ
15 days ago
Post
2385
Quick update from week 1 of smol course. The community is taking the driving seat and using the material for their own projects. If you want to do the same, join in!
- we have ongoing translation projects in Korean, Vietnamese, Portuguese, and Spanish
- 3 chapters are ready for students. On topics like, instruction tuning, preference alignment, and parameter efficient fine tuning
- 3 chapters are in progress on evaluation, vision language models, and synthetic data.
- around 780 people have forked the repo to use it for learning, teaching, sharing.
โญ๏ธ Next step is to support people that want to use the course for teaching, content creation, internal knowledge sharing, or anything. If you're into this. Drop an issue or PR
REPO: https://buff.ly/3ZCMKX2
discord channel: https://buff.ly/4f9F8jA
- we have ongoing translation projects in Korean, Vietnamese, Portuguese, and Spanish
- 3 chapters are ready for students. On topics like, instruction tuning, preference alignment, and parameter efficient fine tuning
- 3 chapters are in progress on evaluation, vision language models, and synthetic data.
- around 780 people have forked the repo to use it for learning, teaching, sharing.
โญ๏ธ Next step is to support people that want to use the course for teaching, content creation, internal knowledge sharing, or anything. If you're into this. Drop an issue or PR
REPO: https://buff.ly/3ZCMKX2
discord channel: https://buff.ly/4f9F8jA
reacted to
bartowski's
post with ๐
15 days ago
Post
7395
Looks like Q4_0_N_M file types are going away
Before you panic, there's a new "preferred" method which is online (I prefer the term on-the-fly) repacking, so if you download Q4_0 and your setup can benefit from repacking the weights into interleaved rows (what Q4_0_4_4 was doing), it will do that automatically and give you similar performance (minor losses I think due to using intrinsics instead of assembly, but intrinsics are more maintainable)
You can see the reference PR here:
https://github.com/ggerganov/llama.cpp/pull/10446
So if you update your llama.cpp past that point, you won't be able to run Q4_0_4_4 (unless they add backwards compatibility back), but Q4_0 should be the same speeds (though it may currently be bugged on some platforms)
As such, I'll stop making those newer model formats soon, probably end of this week unless something changes, but you should be safe to download and Q4_0 quants and use those !
Also IQ4_NL supports repacking though not in as many shapes yet, but should get a respectable speed up on ARM chips, PR for that can be found here: https://github.com/ggerganov/llama.cpp/pull/10541
Remember, these are not meant for Apple silicon since those use the GPU and don't benefit from the repacking of weights
Before you panic, there's a new "preferred" method which is online (I prefer the term on-the-fly) repacking, so if you download Q4_0 and your setup can benefit from repacking the weights into interleaved rows (what Q4_0_4_4 was doing), it will do that automatically and give you similar performance (minor losses I think due to using intrinsics instead of assembly, but intrinsics are more maintainable)
You can see the reference PR here:
https://github.com/ggerganov/llama.cpp/pull/10446
So if you update your llama.cpp past that point, you won't be able to run Q4_0_4_4 (unless they add backwards compatibility back), but Q4_0 should be the same speeds (though it may currently be bugged on some platforms)
As such, I'll stop making those newer model formats soon, probably end of this week unless something changes, but you should be safe to download and Q4_0 quants and use those !
Also IQ4_NL supports repacking though not in as many shapes yet, but should get a respectable speed up on ARM chips, PR for that can be found here: https://github.com/ggerganov/llama.cpp/pull/10541
Remember, these are not meant for Apple silicon since those use the GPU and don't benefit from the repacking of weights
-
- 12 replies
posted
an
update
15 days ago
Post
7575
After some heated discussion ๐ฅ, we clarify our intent re. storage limits on the Hub
TL;DR:
- public storage is free, and (unless blatant abuse) unlimited. We do ask that you consider upgrading to PRO and/or Enterprise Hub if possible
- private storage is paid above a significant free tier (1TB if you have a paid account, 100GB otherwise)
docs: https://huggingface.co/docs/hub/storage-limits
We optimize our infrastructure continuously to scale our storage for the coming years of growth in Machine learning, to the benefit of the community ๐ฅ
cc: @reach-vb @pierric @victor and the HF team
TL;DR:
- public storage is free, and (unless blatant abuse) unlimited. We do ask that you consider upgrading to PRO and/or Enterprise Hub if possible
- private storage is paid above a significant free tier (1TB if you have a paid account, 100GB otherwise)
docs: https://huggingface.co/docs/hub/storage-limits
We optimize our infrastructure continuously to scale our storage for the coming years of growth in Machine learning, to the benefit of the community ๐ฅ
cc: @reach-vb @pierric @victor and the HF team
-
-
-
- 28 replies
reacted to
cfahlgren1's
post with ๐๐
19 days ago
Post
2989
We just dropped an LLM inside the SQL Console ๐คฏ
The amazing, new Qwen/Qwen2.5-Coder-32B-Instruct model can now write SQL for any Hugging Face dataset โจ
It's 2025, you shouldn't be hand writing SQL! This is a big step in making it where anyone can do in depth analysis on a dataset. Let us know what you think ๐ค
The amazing, new Qwen/Qwen2.5-Coder-32B-Instruct model can now write SQL for any Hugging Face dataset โจ
It's 2025, you shouldn't be hand writing SQL! This is a big step in making it where anyone can do in depth analysis on a dataset. Let us know what you think ๐ค