Neuropark (@NeuroPark) profile picture

Training models with @huggingface @yandexai, Check our models at https://nitter.cf/t.co/F4VaXrI5S7

Joined May 2020
What a pleasure to be able to see the content of a dataset without having to download it or navigate through raw files! 🤩
Dataset observability is key to better ML. You can now preview datasets' contents DIRECTLY on the Hub 🔥🎉 Thanks to the streaming feature of datasets, half of the community's datasets are supported out of the box. It also supports Image and Audio datasets, not just text!
1
7
Congrats Lucile and team!! Fruitful collaboration with @YandexAI @HSE_eng @mipt_eng @UofT @VectorInst @NeuroPark Today I'm gonna play youtube.com/watch?v=v8ShbLas… in an infinite loop 😎
🎉 So proud to announce that our "Distributed Deep Learning in Open Collaborations" paper has been accepted at #NeurIPS2021! 🎊 Blog post: huggingface.co/blog/collabor… Arxiv paper: arxiv.org/abs/2106.10207 Teaser: youtu.be/v8ShbLasRF8 via @YouTube #researchpaper #MachineLearning
1
19
A huge thank you to all my co-authors: @michael_diskin, Alexey Bukhtiyarov, @m_ryabinin, @qlhoest, Anton Sinitsin, Dmitry Popov, Dmitry Pyrkin, Maxim Kashirin, sasha_borzunov, @avillanovamoral, @deniskamazur, Ilia Kobelev, @YJernite, @Thom_Wolf, and Gennady Pekhimenko
1
1
9
And this project is itself the result of a fruitful collaboration with @YandexAI, @HSE_eng, @mipt_eng, @UofT, @VectorInst and @NeuroPark. ➡️ nitter.cf/m_ryabinin/status/1406…
In our latest work, we propose DeDLOC — a method for efficient collaborative training. This approach allowed us to pretrain sahajBERT (a Bengali-language ALBERT) together with the help of volunteers from the community! (1/10) arxiv.org/abs/2106.10207 github.com/yandex-research/D…
1
1
8
🔐🌐This work proposes DeDLOC: a method to collaboratively train large neural networks with diverse and scattered computing resources Among the experiments conducted, we trained SahajBERT a Bengali language model with 40 volunteers that is competitive with SOTA models 😱
1
1
8
🎉 So proud to announce that our "Distributed Deep Learning in Open Collaborations" paper has been accepted at #NeurIPS2021! 🎊 Blog post: huggingface.co/blog/collabor… Arxiv paper: arxiv.org/abs/2106.10207 Teaser: youtu.be/v8ShbLasRF8 via @YouTube #researchpaper #MachineLearning
4
51
3
265
🤓 The blog reports the work of @michael_diskin, Alexey Bukhtiyarov, @m_ryabinin, @LucileSaulnier, @qlhoest, Anton Sinitsin, Dmitry Popov, Dmitry Pyrkin, Maxim Kashirin, @sasha_borzunov, @avillanovamoral, @deniskamazur, Ilia Kobelev, @YJernie, @Thom_Wolf, and Gennady Pekhimenko.
1
1
4
Who said that training large DL models was only possible with HPC clusters?🧐 In hf.co/blog/collaborative-tra…, we explain how we collaboratively trained a model with 40 volunteers.😱 ⏰You only have 5min to understand the method? Watch our video: youtu.be/zdVsg5zsGdc via @YouTube
4
33
5
180
🧐What if we could use existing computing resources from different entities to train large models? We have shown that is possible by collaboratively training a language model with 40 volunteers🔥🔥 Read our new blog post on our 🤗 website to know more! 🔍 hf.co/blog/collaborative-tra…
9
66
13
229
Finally, we would like to thank all the volunteers of the @Neuropark community 🐣 who accompanied us in this unique experience realized in collaboration with @YandexAI. ✨
1
2
19
New feature alert❤️‍🔥Host your @huggingface + @streamlit apps directly on🤗spaces huggingface.co/docs/hub/spac… 16GB RAM and 8 CPU cores 🔥
4
65
1
231
We've had 3 @huggingface local team reunions in the past week. Can you guess which is which?
6
3
1
50
Neuropark retweeted
🤗🖼️HuggingPics - Easily train and deploy image classifiers for anything using pictures from the web in less than 5 minutes. Automagically share beautiful @huggingface hub model repos to show off your results. Check it out: github.com/nateraw/huggingpi…
4
63
13
243
In our latest work, we propose DeDLOC — a method for efficient collaborative training. This approach allowed us to pretrain sahajBERT (a Bengali-language ALBERT) together with the help of volunteers from the community! (1/10) arxiv.org/abs/2106.10207 github.com/yandex-research/D…
3
39
10
132
Trained using the Hivemind library Code: learning-at-home.github.io/ Paper: "Towards Crowdsourced Training of Large Neural Networks using Decentralized Mixture-of-Experts" arxiv.org/abs/2002.04013
3
12
Collaborative community training of a Bengali language model using Kaggle/Colab GPUs A really interesting project, well done @NeuroPark , @yandexcom , @m_ryabinin and @huggingface for the initiative 🙌 See their W&B dashboard here: wandb.ai/learning-at-home/Ma…
My blog post describing distributed training for the ALBERT Language Model sahajBERT for Bengali organized by @NeuroPark, @yandexcom and @huggingface using learning-at-home/hivemind downloadable soon from Hugging Face -- sujitpal.blogspot.com/2021/0…
1
5
25