Hugging Face (Twitter)
RT @cgeorgiaw: 2026 will be the year of AI-for-science (and my team at @huggingface is hiring for that!)
We laid up the pins in 2025, and now we’re gonna knock them down
Post #2195
26
Channel Public Channel
HU Hugging Face
@huggingface
- Subscribers
- 303
- Photos
- 1.3K
- Videos
- 408
- Links
- 2.2K
Showing posts older than #2196 · Back to latest
Older Posts 19 shown
Post #2194
27
Hugging Face (Twitter)
RT @evalstate: couple of small updates on -- we've got improvements for the new CLI Search tools, standardised on UV for scripts and a little tweak to assist some models that need a bit of help navigation.
RT @evalstate: couple of small updates on -- we've got improvements for the new CLI Search tools, standardised on UV for scripts and a little tweak to assist some models that need a bit of help navigation.
Post #2193
23
Hugging Face (Twitter)
RT @JohannesTscharn: AR and Robotics is such an amazing combination
The video doesn’t do it justice but interacting natively in 3D space is intuitive and feels natural
This is @Spectacles driving a Reachy Mini from @huggingface and @pollenrobotics
Thanks again to @SensAIHackademy 🙂
RT @JohannesTscharn: AR and Robotics is such an amazing combination
The video doesn’t do it justice but interacting natively in 3D space is intuitive and feels natural
This is @Spectacles driving a Reachy Mini from @huggingface and @pollenrobotics
Thanks again to @SensAIHackademy 🙂
Post #2192
59
Hugging Face (Twitter)
RT @Xianbao_QIAN: Want to train your TTS model from scratch with your own cool features?
🎤 Important release for Open-Source SPEECH AI! 🌍
The incredible LEMAS team from IDEA has dropped the dataset they used to train LEMAS, which is also the largest open-source multilingual speech dataset EVER!
150K+ hours across 10 languages with word-level timestamps. They're giving away what most companies would keep locked away forever. 🙏
Two models born from this treasure:
• LEMAS-TTS: Zero-shot multilingual synthesis model
• LEMAS-Edit: Edit speech like you're editing text!!!
Check out this work:
Project home: https://lemas-project.github.io/LEMAS-Project/
Dataset & model released on @huggingface : huggingface.co/LEMAS-Project
RT @Xianbao_QIAN: Want to train your TTS model from scratch with your own cool features?
🎤 Important release for Open-Source SPEECH AI! 🌍
The incredible LEMAS team from IDEA has dropped the dataset they used to train LEMAS, which is also the largest open-source multilingual speech dataset EVER!
150K+ hours across 10 languages with word-level timestamps. They're giving away what most companies would keep locked away forever. 🙏
Two models born from this treasure:
• LEMAS-TTS: Zero-shot multilingual synthesis model
• LEMAS-Edit: Edit speech like you're editing text!!!
Check out this work:
Project home: https://lemas-project.github.io/LEMAS-Project/
Dataset & model released on @huggingface : huggingface.co/LEMAS-Project
Post #2191
29
Hugging Face (Twitter)
RT @Xianbao_QIAN: 2025 is a watershed year for open-source AI. 🚀
It started with DeepSeek R1, followed rapidly by Qwen3, Kimi, and also led to a historic milestone: @Zai_org and @MiniMax_AI became the first LLM companies globally to IPO.
The ecosystem is thriving beyond just the giants:
- Tech & Media: @BaiduResearch, @ByteDanceOSS, @xiaohongshu, @Meituan_LongCat, @ant_oss and @bilibili_en are all making remarkable contributions.
- Finance: More AI teams from quantitative trading firms like IQuest are joining the fray.
- Multimodal: The trend has expanded beyond text to models like @TencentHunyuan 3D, @SkyReels, @Alibaba_Wan, and @Alibaba_Qwen-image.
- Robotics: A huge wave of datasets and models from @XSquareRobot, @UnitreeRobotics, @GalaxeaDynamics, @agibot_research, SpiritAI, and more.
- Datasets: Non-profits like @BAAIBeijing, @OpenMMLab, and IDEA continue to pave the way with impactful open-source data.
This has been a massive year for the...
Перейти на оригинальный пост
RT @Xianbao_QIAN: 2025 is a watershed year for open-source AI. 🚀
It started with DeepSeek R1, followed rapidly by Qwen3, Kimi, and also led to a historic milestone: @Zai_org and @MiniMax_AI became the first LLM companies globally to IPO.
The ecosystem is thriving beyond just the giants:
- Tech & Media: @BaiduResearch, @ByteDanceOSS, @xiaohongshu, @Meituan_LongCat, @ant_oss and @bilibili_en are all making remarkable contributions.
- Finance: More AI teams from quantitative trading firms like IQuest are joining the fray.
- Multimodal: The trend has expanded beyond text to models like @TencentHunyuan 3D, @SkyReels, @Alibaba_Wan, and @Alibaba_Qwen-image.
- Robotics: A huge wave of datasets and models from @XSquareRobot, @UnitreeRobotics, @GalaxeaDynamics, @agibot_research, SpiritAI, and more.
- Datasets: Non-profits like @BAAIBeijing, @OpenMMLab, and IDEA continue to pave the way with impactful open-source data.
This has been a massive year for the...
Перейти на оригинальный пост
Post #2190
41
Hugging Face (Twitter)
RT @evalstate: 🚢Just Shipped -- hackable Agents and Tools for use with Hugging Face Inference Providers. Update HF Inference providers and install them with `/quickstart`. Use `/skills` to install skills from the Hugging Face Skills repo and others. Now is a great time to try Open Weights 💪 https://twitter.com/willmcgugan/status/2009976519133720904#m
RT @evalstate: 🚢Just Shipped -- hackable Agents and Tools for use with Hugging Face Inference Providers. Update HF Inference providers and install them with `/quickstart`. Use `/skills` to install skills from the Hugging Face Skills repo and others. Now is a great time to try Open Weights 💪 https://twitter.com/willmcgugan/status/2009976519133720904#m
Post #2189
38
Hugging Face (Twitter)
RT @prithivMLmods: LTX-2 Camera-Control with Dolly-in/out and Dolly-left/right LoRA demo is now available on Hugging Face, paired with ltx-2-19b-distilled-lora for fast inference.
🤗Try it now on @huggingface: https://huggingface.co/spaces/prithivMLmods/LTX-2-LoRAs-Camera-Control-Dolly
🎥Dolly-In: https://huggingface.co/Lightricks/LTX-2-19b-LoRA-Camera-Control-Dolly-In
@Lightricks
RT @prithivMLmods: LTX-2 Camera-Control with Dolly-in/out and Dolly-left/right LoRA demo is now available on Hugging Face, paired with ltx-2-19b-distilled-lora for fast inference.
🤗Try it now on @huggingface: https://huggingface.co/spaces/prithivMLmods/LTX-2-LoRAs-Camera-Control-Dolly
🎥Dolly-In: https://huggingface.co/Lightricks/LTX-2-19b-LoRA-Camera-Control-Dolly-In
@Lightricks
Post #2188
42
Hugging Face (Twitter)
RT @Xianbao_QIAN: The base model has been open sourced on @huggingface: https://huggingface.co/Spirit-AI-robotics/Spirit-v1.5
Fine-tune data for RoboChallenge task: https://huggingface.co/datasets/RoboChallenge/Table30 https://twitter.com/Xianbao_QIAN/status/2010330035450953842#m
RT @Xianbao_QIAN: The base model has been open sourced on @huggingface: https://huggingface.co/Spirit-AI-robotics/Spirit-v1.5
Fine-tune data for RoboChallenge task: https://huggingface.co/datasets/RoboChallenge/Table30 https://twitter.com/Xianbao_QIAN/status/2010330035450953842#m
Post #2187
40
Hugging Face (Twitter)
RT @Trtd6Trtd: Reachy MiniにWebニュースを読ませてる
アレクサでええやろがい、と言われたらそれまでだが、ちょこちょこ動くのが可愛いので満足
RT @Trtd6Trtd: Reachy MiniにWebニュースを読ませてる
アレクサでええやろがい、と言われたらそれまでだが、ちょこちょこ動くのが可愛いので満足
Post #2186
38
Hugging Face (Twitter)
RT @eliebakouch: Most web data in (very) low resource languages is Bible and Wikipedia. The rest? @huggingface data team ran Gemma3 27B for 3 months to translate it into english, to improve translation models and to bring cultural context from 500+ language communities into english training data. Here is the full pipeline
https://huggingface.co/datasets/HuggingFaceFW/finetranslations https://twitter.com/gui_penedo/status/2009677127671492616#m
RT @eliebakouch: Most web data in (very) low resource languages is Bible and Wikipedia. The rest? @huggingface data team ran Gemma3 27B for 3 months to translate it into english, to improve translation models and to bring cultural context from 500+ language communities into english training data. Here is the full pipeline
https://huggingface.co/datasets/HuggingFaceFW/finetranslations https://twitter.com/gui_penedo/status/2009677127671492616#m
Post #2185
31
Hugging Face (Twitter)
RT @ben_burtenshaw: TIL: you can scale rl environments as hub spaces to batch sizes of 128 on free hardware! This is ideal for most (humble) RL workloads. Plus, openenv scales from there.
this weekend I'm running a benchmarking experiment on openenv RL environments to see what scale they can reach on the main hardware types: (free/ paid) hub spaces, local docker, local python, and slurm+envoy.
the idea is to have a single env definition (openenv) which takes us from hacker scale to hyper scale, without the env getting in the way.
we're sharing this in detail on wednesday in GPU-Mode
RT @ben_burtenshaw: TIL: you can scale rl environments as hub spaces to batch sizes of 128 on free hardware! This is ideal for most (humble) RL workloads. Plus, openenv scales from there.
this weekend I'm running a benchmarking experiment on openenv RL environments to see what scale they can reach on the main hardware types: (free/ paid) hub spaces, local docker, local python, and slurm+envoy.
the idea is to have a single env definition (openenv) which takes us from hacker scale to hyper scale, without the env getting in the way.
we're sharing this in detail on wednesday in GPU-Mode
Post #2184
32
Hugging Face (Twitter)
RT @willmcgugan: Both @huggingface and @OpenHandsDev are offering free tokens for their AI models!
That's $30 free to build your next product.
Get your free tokens via Toad, the unified interface for AI in your terminal.
curl -fsSL batrachian.ai/install | sh
github.com/textualize/toad
RT @willmcgugan: Both @huggingface and @OpenHandsDev are offering free tokens for their AI models!
That's $30 free to build your next product.
Get your free tokens via Toad, the unified interface for AI in your terminal.
curl -fsSL batrachian.ai/install | sh
github.com/textualize/toad
Post #2183
28
Hugging Face (Twitter)
RT @lhoestq: We shipped it yesterday and there were already 20k+ datasets in Optimized-parquet format 🤯
Anyone can benefit from all their advantages today. Learn how it works below, it's a game changer 👇 https://twitter.com/lhoestq/status/2009631565706637768#m
RT @lhoestq: We shipped it yesterday and there were already 20k+ datasets in Optimized-parquet format 🤯
Anyone can benefit from all their advantages today. Learn how it works below, it's a game changer 👇 https://twitter.com/lhoestq/status/2009631565706637768#m
Post #2182
38
Hugging Face (Twitter)
RT @JosephJacks_: Awesome to see @liquidai trending again on @huggingface !! 🌊 📈
Congrats to @maximelabonne, @mlech26l, @ramin_m_h and colleagues. 👏🏼 👏🏼
RT @JosephJacks_: Awesome to see @liquidai trending again on @huggingface !! 🌊 📈
Congrats to @maximelabonne, @mlech26l, @ramin_m_h and colleagues. 👏🏼 👏🏼
Post #2181
38
Hugging Face (Twitter)
RT @maximelabonne: We're trending! 🥳
And congrats to @Lightricks for their amazing model
RT @maximelabonne: We're trending! 🥳
And congrats to @Lightricks for their amazing model
Post #2180
34
Hugging Face (Twitter)
RT @lhoestq: Optimized-parquet is now on @huggingface, this is a new era for datasets !
here is why it's amazing and how it works:
RT @lhoestq: Optimized-parquet is now on @huggingface, this is a new era for datasets !
here is why it's amazing and how it works:
Post #2179
27
Hugging Face (Twitter)
RT @Basilakis: I am so impressed of using Inference Endpoints from @huggingface . So easy to use, so amazing results. The stuff you can get are dope. Some many other services existing doing this, but the simplicity of repo2inference is so easy, that you wonder why to test anything else.
RT @Basilakis: I am so impressed of using Inference Endpoints from @huggingface . So easy to use, so amazing results. The stuff you can get are dope. Some many other services existing doing this, but the simplicity of repo2inference is so easy, that you wonder why to test anything else.
Post #2178
29
Hugging Face (Twitter)
RT @SOSOHAJALAB: Do you guys still cook somethings yourself?
Claudecode with plugins @huggingface doing this automatically but also, make the datasets and upload to HF itself.
How fantastic lovely simply cooking via claude code 👍
RT @SOSOHAJALAB: Do you guys still cook somethings yourself?
Claudecode with plugins @huggingface doing this automatically but also, make the datasets and upload to HF itself.
How fantastic lovely simply cooking via claude code 👍
Post #2176
27
Hugging Face (Twitter)
RT @EmileAndHisBots: One of the things we were busy with over Christmas: Reachy Mini by @pollenrobotics and @huggingface - I had my kids assemble it; it's a very well-designed robot!
RT @EmileAndHisBots: One of the things we were busy with over Christmas: Reachy Mini by @pollenrobotics and @huggingface - I had my kids assemble it; it's a very well-designed robot!