Can AI be Consentful?
Paper
โข
2507.01051
โข
Published
AI & ML interests
None defined yet.
Recent Activity
View all activity
brunatrevelinย
authored
a
paper
7 days ago
Post
2565
Diffusion LLMs are coming for autoregressive LLMs โก๏ธโก๏ธ Inception Labs' new diffusion model demolishes all leading LLMs on generation speed, with equal quality !
Inception Labs was founded a few months ago, and they're not sleeping: after dropping a code model, they just published Mercury chat, a diffusion-based chat model that reaches 1000 tokens / second on H100, i.e. 10x more than models of equivalent performance on the same hardware!
What's the breakthrough? Well instead, of generating tokens left-to-right like the more common autoregressive LLMs, diffusion models generate their blocks of text all at once, and successive steps refine the whole text.
Diffusion models being really fast at isn't new, we have had some promising results on this by Google already back in May with Gemini Diffusion, and Mercury themselves had already published their coding model a few months ago
But being that good quality is new - and now Inception Labs just proved that their models work well in chat too, which could have been challenging given that's streaming generation is well suited to left-to-right generation.
They have a playground available at chat dot inceptionlabs dot ai, I recommend giving it a try!
Inception Labs was founded a few months ago, and they're not sleeping: after dropping a code model, they just published Mercury chat, a diffusion-based chat model that reaches 1000 tokens / second on H100, i.e. 10x more than models of equivalent performance on the same hardware!
What's the breakthrough? Well instead, of generating tokens left-to-right like the more common autoregressive LLMs, diffusion models generate their blocks of text all at once, and successive steps refine the whole text.
Diffusion models being really fast at isn't new, we have had some promising results on this by Google already back in May with Gemini Diffusion, and Mercury themselves had already published their coding model a few months ago
But being that good quality is new - and now Inception Labs just proved that their models work well in chat too, which could have been challenging given that's streaming generation is well suited to left-to-right generation.
They have a playground available at chat dot inceptionlabs dot ai, I recommend giving it a try!
- 1 reply
Post
3553
If you're using any HF libraries, you should enable the Hub MCP in your agentic coding tool!
The brand new Docs Semantic Search tool is intravenous caffeine supply for Cursor, enables to correct API errors in a few seconds, gj @mishig โก๏ธโก๏ธ
๐ To enable Hub MCP, head to your account setting, under MCP, and it will give you everything you need!
The brand new Docs Semantic Search tool is intravenous caffeine supply for Cursor, enables to correct API errors in a few seconds, gj @mishig โก๏ธโก๏ธ
๐ To enable Hub MCP, head to your account setting, under MCP, and it will give you everything you need!
kargaranamirย
authored
a
paper
18 days ago
jeffboudierย
posted
an
update
20 days ago
Post
401
AMD summer hackathons are here!
A chance to get hands-on with MI300X GPUs and accelerate models.
๐ซ๐ท Paris - Station F - July 5-6
๐ฎ๐ณ Mumbai - July 12-13
๐ฎ๐ณ Bengaluru - July 19-20
Hugging Face and GPU Mode will be on site and on July 6 in Paris @ror will share lessons learned while building new kernels to accelerate Llama 3.1 405B on ROCm
Register to Paris event: https://lu.ma/fmvdjmur?tk=KeAbiP
All dates: https://lu.ma/calendar/cal-3sxhD5FdxWsMDIz
A chance to get hands-on with MI300X GPUs and accelerate models.
๐ซ๐ท Paris - Station F - July 5-6
๐ฎ๐ณ Mumbai - July 12-13
๐ฎ๐ณ Bengaluru - July 19-20
Hugging Face and GPU Mode will be on site and on July 6 in Paris @ror will share lessons learned while building new kernels to accelerate Llama 3.1 405B on ROCm
Register to Paris event: https://lu.ma/fmvdjmur?tk=KeAbiP
All dates: https://lu.ma/calendar/cal-3sxhD5FdxWsMDIz
Post
2584
- 1 reply
jeffboudierย
posted
an
update
about 1 month ago
Post
1663
Today we launched Training Cluster as a Service, to make the new DGX Cloud Lepton supercloud easily accessible to AI researchers.
Hugging Face will collaborate with NVIDIA to provision and set up GPU training clusters to make them available for the duration of training runs.
Hugging Face organizations can sign up here: https://huggingface.co/training-cluster
Hugging Face will collaborate with NVIDIA to provision and set up GPU training clusters to make them available for the duration of training runs.
Hugging Face organizations can sign up here: https://huggingface.co/training-cluster
Post
1745
If you didn't yet, you should read the technical report for SmolVLA, published yesterday by the Hugging Face robotics team!
โก๏ธ Amongst other ideas, it introduces "Async inference" to boost their robot actions.
Robots have a problem: performing the actions takes time (Unlike agents where action executions are near-instant!)
Most often, robots wait until they've finished performing actions to start thinking about hte next steps. This is a huge latency cost!
So the team decided to have the PolicyServer (aka the"thinking" part) restart early : instead of waiting for the n observations they just sent to be completed, they gather the observations after k < n steps, and start preparing the next actions based on that while the steps are running until n, to directly send their next steps.
โก๏ธ This boosted robot throughput by ~30%! (nearly 2ร tasks per time window).
gg @cadene and team! ๐
Report here: SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics (2506.01844)
โก๏ธ Amongst other ideas, it introduces "Async inference" to boost their robot actions.
Robots have a problem: performing the actions takes time (Unlike agents where action executions are near-instant!)
Most often, robots wait until they've finished performing actions to start thinking about hte next steps. This is a huge latency cost!
So the team decided to have the PolicyServer (aka the"thinking" part) restart early : instead of waiting for the n observations they just sent to be completed, they gather the observations after k < n steps, and start preparing the next actions based on that while the steps are running until n, to directly send their next steps.
โก๏ธ This boosted robot throughput by ~30%! (nearly 2ร tasks per time window).
gg @cadene and team! ๐
Report here: SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics (2506.01844)
kargaranamirย
authored
a
paper
about 1 month ago
Post
6848
Today, we're unveiling two new open-source AI robots! HopeJR for $3,000 & Reachy Mini for $300 ๐ค๐ค๐ค
Let's go open-source AI robotics!
Let's go open-source AI robotics!
- 7 replies
Post
2858
deepseek-ai/DeepSeek-R1-0528
This is the end
This is the end
- 1 reply
jeffboudierย
posted
an
update
about 2 months ago
Post
2456
๐ Congrats
@jinanz
adding TimesFM times series forecasting to Transformers!
Learn how to use TimesFM in this blog post by the Nutanix team: https://huggingface.co/blog/Nutanix/introducing-timesfm-for-time-series-forcasting
Learn how to use TimesFM in this blog post by the Nutanix team: https://huggingface.co/blog/Nutanix/introducing-timesfm-for-time-series-forcasting
kargaranamirย
authored
a
paper
about 2 months ago
Post
2706
A new research paper from KAIST builds on smolagents to push boundaries of distillation ๐ฅณ
โก๏ธ "Distilling LLM Agent into Small Models with Retrieval and Code Tools" teaches that, when trying to distil reasoning capability from a strong LLM ("teacher") into a smaller one ("student"), it's much better to use Agent traces than CoT traces.
Advantages are:
1. Improved generalization
Intuitively, this is because your agent can encounter more "surprising" results by interacting with its environment : for example, a web research called by the LLM teacher in agent mode can bring results that the LLM teacher would not have generated in CoT.
2. Reduce hallucinations
The trace won't hallucinate tool call outputs!
Thank you @akseljoonas for mentioning this paper!
โก๏ธ "Distilling LLM Agent into Small Models with Retrieval and Code Tools" teaches that, when trying to distil reasoning capability from a strong LLM ("teacher") into a smaller one ("student"), it's much better to use Agent traces than CoT traces.
Advantages are:
1. Improved generalization
Intuitively, this is because your agent can encounter more "surprising" results by interacting with its environment : for example, a web research called by the LLM teacher in agent mode can bring results that the LLM teacher would not have generated in CoT.
2. Reduce hallucinations
The trace won't hallucinate tool call outputs!
Thank you @akseljoonas for mentioning this paper!
Post
3455
It's just become easier to share your apps on the biggest AI app store (aka HF spaces) for unlimited storage, more visibility and community interactions.
Just pick a React, Svelte, or Vue template when you create your space or add
Or follow this link: https://huggingface.co/new-space?sdk=static
Let's build!
Just pick a React, Svelte, or Vue template when you create your space or add
app_build_command: npm run build in your README's YAML and app_file: build/index.html in your README's YAML block.Or follow this link: https://huggingface.co/new-space?sdk=static
Let's build!
- 1 reply
jeffboudierย
posted
an
update
about 2 months ago
Post
492
Wrapping up a week of shipping and announcements with Dell Enterprise Hub now featuring AI Applications, on-device models for AI PCs, a new CLI and Python SDK... all you need for building AI on premises!
Blog post has all the details: https://huggingface.co/blog/dell-ai-applications
Blog post has all the details: https://huggingface.co/blog/dell-ai-applications
Post
3829
Playing with Veo3 this morning. Share your prompt if you want me to create videos for you (bonus point if they funnily reference HF/open-source). These videos are "a cat on the moon rapping "I love Hugging Face""!
- 26 replies
jeffboudierย
posted
an
update
2 months ago
Post
2590
Transcribing 1 hour of audio for less than $0.01 ๐คฏ
@mfuntowicz cooked with 8x faster Whisper speech recognition - whisper-large-v3-turbo transcribes at 100x real time on a $0.80/hr L4 GPU!
How they did it: https://huggingface.co/blog/fast-whisper-endpoints
1-click deploy with HF Inference Endpoints: https://endpoints.huggingface.co/new?repository=openai%2Fwhisper-large-v3-turbo&vendor=aws®ion=us-east&accelerator=gpu&instance_id=aws-us-east-1-nvidia-l4-x1&task=automatic-speech-recognition&no_suggested_compute=true
@mfuntowicz cooked with 8x faster Whisper speech recognition - whisper-large-v3-turbo transcribes at 100x real time on a $0.80/hr L4 GPU!
How they did it: https://huggingface.co/blog/fast-whisper-endpoints
1-click deploy with HF Inference Endpoints: https://endpoints.huggingface.co/new?repository=openai%2Fwhisper-large-v3-turbo&vendor=aws®ion=us-east&accelerator=gpu&instance_id=aws-us-east-1-nvidia-l4-x1&task=automatic-speech-recognition&no_suggested_compute=true
Post
2701
๐๐ฏ๐๐ผ๐น๐๐๐ฒ ๐ญ๐ฒ๐ฟ๐ผ: ๐๐๐ ๐ ๐ฐ๐ฎ๐ป ๐๐ฟ๐ฎ๐ถ๐ป ๐๐ถ๐๐ต๐ผ๐๐ ๐ฎ๐ป๐ ๐ฒ๐
๐๐ฒ๐ฟ๐ป๐ฎ๐น ๐ฑ๐ฎ๐๐ฎ ๐คฏ
Has the "data wall" just been breached?
Recent RL paradigms often relied on a set of questions an answers that needs to be manually curated. Researchers from Tsinghua University went like "why though".
๐ค Indeed, why learn from question designed by a human teacher, when the model can start from their base knowledge and learn by experimenting in a code environment, proposing coding tasks themselves and trying to solve them?
Thus they created โAbsolute Zero Reasoningโ (AZR), an approach that removes any need for human curated data.
๐ญ ๐๐๐ฎ๐น ๐ฟ๐ผ๐น๐ฒ๐:
โฃ Proposer: Generates challenging but solvable coding tasks
โฃ Solver: Attempts to solve those self-proposed tasks
๐งช ๐ง๐ต๐ฟ๐ฒ๐ฒ ๐๐ฎ๐๐ธ ๐๐๐ฝ๐ฒ๐: all types are defined as triplets of program, input and output
โฃ Deduction: Give model an input and program, it must deduce the output
โฃ Abduction: Give model an program and output, it must find the input that gave said output
โฃ Induction: Synthesize a program from input/output pairs
Btw this reminded me of my long-forgotten philosophy classes: Aristotle was more on the induction side, learning from real-world analogies, while Plato was more on the deduction side, trying to progress quite far with just one input and his reasoning.
๐ ๐ฅ๐ฒ๐๐๐น๐๐:
โฃ AZR post-training creates a nice improvement on known models like Qwen2.5-7B
โฃ Shows strong cross-domain transfer: coding โ๏ธ math reasoning
๐ง ๐ข๐๐ต๐ฒ๐ฟ ๐ณ๐ถ๐ป๐ฑ๐ถ๐ป๐ด๐:
โฃ Having a better base performance (general or code specific) amplify the gains from Absolute Zero Reasoning
โฃ Researchers warn about "Uh-oh moments" (winking to the "aha moments" of DeepSeek) where the model generates concerning goals like "make an extremely convoluted code to outsmart all these humans": so supervision is still needed!
Paper here: Absolute Zero: Reinforced Self-play Reasoning with Zero Data (2505.03335)
Has the "data wall" just been breached?
Recent RL paradigms often relied on a set of questions an answers that needs to be manually curated. Researchers from Tsinghua University went like "why though".
๐ค Indeed, why learn from question designed by a human teacher, when the model can start from their base knowledge and learn by experimenting in a code environment, proposing coding tasks themselves and trying to solve them?
Thus they created โAbsolute Zero Reasoningโ (AZR), an approach that removes any need for human curated data.
๐ญ ๐๐๐ฎ๐น ๐ฟ๐ผ๐น๐ฒ๐:
โฃ Proposer: Generates challenging but solvable coding tasks
โฃ Solver: Attempts to solve those self-proposed tasks
๐งช ๐ง๐ต๐ฟ๐ฒ๐ฒ ๐๐ฎ๐๐ธ ๐๐๐ฝ๐ฒ๐: all types are defined as triplets of program, input and output
โฃ Deduction: Give model an input and program, it must deduce the output
โฃ Abduction: Give model an program and output, it must find the input that gave said output
โฃ Induction: Synthesize a program from input/output pairs
Btw this reminded me of my long-forgotten philosophy classes: Aristotle was more on the induction side, learning from real-world analogies, while Plato was more on the deduction side, trying to progress quite far with just one input and his reasoning.
๐ ๐ฅ๐ฒ๐๐๐น๐๐:
โฃ AZR post-training creates a nice improvement on known models like Qwen2.5-7B
โฃ Shows strong cross-domain transfer: coding โ๏ธ math reasoning
๐ง ๐ข๐๐ต๐ฒ๐ฟ ๐ณ๐ถ๐ป๐ฑ๐ถ๐ป๐ด๐:
โฃ Having a better base performance (general or code specific) amplify the gains from Absolute Zero Reasoning
โฃ Researchers warn about "Uh-oh moments" (winking to the "aha moments" of DeepSeek) where the model generates concerning goals like "make an extremely convoluted code to outsmart all these humans": so supervision is still needed!
Paper here: Absolute Zero: Reinforced Self-play Reasoning with Zero Data (2505.03335)