94
README
🌍
ZeroGPU Explorers
community
AI & ML interests
None defined yet.
Recent Activity
View all activity
zero-gpu-explorers's activity
Update README.md
1
#152 opened 14 days ago
by
fdaudens
fdaudens
updated
a
Space
14 days ago
Update README.md
1
#152 opened 14 days ago
by
fdaudens
IliaLarchenko
posted
an
update
15 days ago
Post
2042
I am presenting Decoder-Only Transformer (DOT) Policy a simple Behavioral Control policy that outperforms SOTA models on two simple benchmark tasks:
✅ PushT (pushing an object to a goal) – 84% success on keypoints, 74% on images (previous best: 75% / 69%)
✅ ALOHA Insert (precise bimanual insertion) – 30% success (previous best: ~21%)
The best part? DOT is much smaller (sometimes 100 times less parameters) than previous SOTA models, trains faster, and avoids complexity:
🚫 No generative models (Diffusion, VAE, GANs)
🚫 No discretization/tokenization of actions
🚫 No reinforcement learning or multi-stage training
✅ Just learns from human demos, plain and simple
This is still early — more complex real-life tasks need testing, and no guarantees it will actually work well there, but I think it's interesting to share. Sometimes, simpler approaches can be just as effective (or even better) than complex ones.
🔗 Open-source code and detailed description: https://github.com/IliaLarchenko/dot_policy
Trained models on Hugging Face:
IliaLarchenko/dot_pusht_keypoints
IliaLarchenko/dot_pusht_images
IliaLarchenko/dot_bimanual_insert
✅ PushT (pushing an object to a goal) – 84% success on keypoints, 74% on images (previous best: 75% / 69%)
✅ ALOHA Insert (precise bimanual insertion) – 30% success (previous best: ~21%)
The best part? DOT is much smaller (sometimes 100 times less parameters) than previous SOTA models, trains faster, and avoids complexity:
🚫 No generative models (Diffusion, VAE, GANs)
🚫 No discretization/tokenization of actions
🚫 No reinforcement learning or multi-stage training
✅ Just learns from human demos, plain and simple
This is still early — more complex real-life tasks need testing, and no guarantees it will actually work well there, but I think it's interesting to share. Sometimes, simpler approaches can be just as effective (or even better) than complex ones.
🔗 Open-source code and detailed description: https://github.com/IliaLarchenko/dot_policy
Trained models on Hugging Face:
IliaLarchenko/dot_pusht_keypoints
IliaLarchenko/dot_pusht_images
IliaLarchenko/dot_bimanual_insert
Post
2145
🙋🏻♂️hey there folks ,
Goedel's Theorem Prover is now being demo'ed on huggingface : Tonic/Math
give it a try !
Goedel's Theorem Prover is now being demo'ed on huggingface : Tonic/Math
give it a try !
Post
2873
Simple Paper Review #5
I briefly reviewed the paper "SFT Memorizes, RL Generalizes," which compares SFT and RL in post-training of LLM/VLM from HKU, UC Berkeley, Google DeepMind, and New York University
The conclusion suggests SFT excels in memorization, while RL is better for generalization. However, since LLM/VLM should benefit humans beyond just generalization, a mix of SFT and RL is advisable. Typically, some SFT is followed by RL to understand prompt formats and enhance generalization through trial and error.
The study focused on one model, Llama-3.2-Vision-11B, using environments like General Points for arithmetic reasoning and V-IRL for spatial reasoning. Training data was used for both SFT and RL, with evaluations on in-distribution and out-of-distribution data to assess memorization and generalization.
I want to apply RL extensively, but it requires building a similar simulation environment. For domain-specific models, significant investment in creating a "playground" for the model is crucial, as the effort will directly influence the outcomes.
https://arxiv.org/abs/2501.17161
I briefly reviewed the paper "SFT Memorizes, RL Generalizes," which compares SFT and RL in post-training of LLM/VLM from HKU, UC Berkeley, Google DeepMind, and New York University
The conclusion suggests SFT excels in memorization, while RL is better for generalization. However, since LLM/VLM should benefit humans beyond just generalization, a mix of SFT and RL is advisable. Typically, some SFT is followed by RL to understand prompt formats and enhance generalization through trial and error.
The study focused on one model, Llama-3.2-Vision-11B, using environments like General Points for arithmetic reasoning and V-IRL for spatial reasoning. Training data was used for both SFT and RL, with evaluations on in-distribution and out-of-distribution data to assess memorization and generalization.
I want to apply RL extensively, but it requires building a similar simulation environment. For domain-specific models, significant investment in creating a "playground" for the model is crucial, as the effort will directly influence the outcomes.
https://arxiv.org/abs/2501.17161
Post
4278
A brief summary of the o3-mini
The OpenAI o3-mini model is a significant improvement over the o1-mini, reaching o1 performance levels. While generally good, its performance isn't universally better than previous models (o1, o1-prev.) or GPT-4o across all benchmarks. This means workflows should be re-evaluated with each model upgrade.
The o3-mini has "low," "medium," and "high" versions, with "low" being the base model used for benchmarking. It's speculated that the higher versions simply involve more processing. A fair comparison with other models like Gemini 2.0 Thinking or DeepSeek-R1 would likely need to use the "low" version and a similar "think more" mechanism.
The system card is recommended reading due to its comprehensive benchmark data.
https://openai.com/index/openai-o3-mini/
The OpenAI o3-mini model is a significant improvement over the o1-mini, reaching o1 performance levels. While generally good, its performance isn't universally better than previous models (o1, o1-prev.) or GPT-4o across all benchmarks. This means workflows should be re-evaluated with each model upgrade.
The o3-mini has "low," "medium," and "high" versions, with "low" being the base model used for benchmarking. It's speculated that the higher versions simply involve more processing. A fair comparison with other models like Gemini 2.0 Thinking or DeepSeek-R1 would likely need to use the "low" version and a similar "think more" mechanism.
The system card is recommended reading due to its comprehensive benchmark data.
https://openai.com/index/openai-o3-mini/
mkshing
authored
2
papers
22 days ago
Post
3854
I have just released a new blogpost about kv caching and its role in inference speedup 🚀
🔗 https://huggingface.co/blog/not-lain/kv-caching/
some takeaways :
🔗 https://huggingface.co/blog/not-lain/kv-caching/
some takeaways :
- 4 replies
Post
2008
Simple summary on DeepSeek AI's Janus-Pro: A fresh take on multimodal AI!
It builds on its predecessor, Janus, by tweaking the training methodology rather than the model architecture. The result? Improved performance in understanding and generating multimodal data.
Janus-Pro uses a three-stage training strategy, similar to Janus, but with key modifications:
✦ Stage 1 & 2: Focus on separate training for specific objectives, rather than mixing data.
✦ Stage 3: Fine-tuning with a careful balance of multimodal data.
Benchmarks show Janus-Pro holds its own against specialized models like TokenFlow XL and MetaMorph, and other multimodal models like SD3 Medium and DALL-E 3.
The main limitation? Low image resolution (384x384). However, this seems like a strategic choice to focus on establishing a solid "recipe" for multimodal models. Future work will likely leverage this recipe and increased computing power to achieve higher resolutions.
It builds on its predecessor, Janus, by tweaking the training methodology rather than the model architecture. The result? Improved performance in understanding and generating multimodal data.
Janus-Pro uses a three-stage training strategy, similar to Janus, but with key modifications:
✦ Stage 1 & 2: Focus on separate training for specific objectives, rather than mixing data.
✦ Stage 3: Fine-tuning with a careful balance of multimodal data.
Benchmarks show Janus-Pro holds its own against specialized models like TokenFlow XL and MetaMorph, and other multimodal models like SD3 Medium and DALL-E 3.
The main limitation? Low image resolution (384x384). However, this seems like a strategic choice to focus on establishing a solid "recipe" for multimodal models. Future work will likely leverage this recipe and increased computing power to achieve higher resolutions.
Post
2906
🙋🏻♂️ Hey there folks ,
our team made a game during the @mistral-game-jam and we're trying to win the community award !
try our game out and drop us a ❤️ like basically to vote for us !
Mistral-AI-Game-Jam/TextToSurvive
hope you like it !
our team made a game during the @mistral-game-jam and we're trying to win the community award !
try our game out and drop us a ❤️ like basically to vote for us !
Mistral-AI-Game-Jam/TextToSurvive
hope you like it !
Post
1725
New look for AI powered paper reviews from the list by Hugging Face Daily Papers ( managed by the
@akhaliq
)
Bookmark the webpage along, check comprehensive reviews by Google DeepMind Gemini 1.5, and listen to audio podcast made by the same tech used in NotebookLM.
Link: https://deep-diver.github.io/ai-paper-reviewer/
This is not an official service by Hugging Face. It is just a service developed by an individual developer using his own money :)
Bookmark the webpage along, check comprehensive reviews by Google DeepMind Gemini 1.5, and listen to audio podcast made by the same tech used in NotebookLM.
Link: https://deep-diver.github.io/ai-paper-reviewer/
This is not an official service by Hugging Face. It is just a service developed by an individual developer using his own money :)
Post
2015
Simple summarization of Evolving Deeper LLM Thinking (Google DeepMind)
The process starts by posing a question.
1) The LLM generates initial responses.
2) These generated responses are evaluated according to specific criteria (program-based checker).
3) The LLM critiques the evaluated results.
4) The LLM refines the responses based on the evaluation, critique, and original responses.
The refined response is then fed back into step 2). If it meets the criteria, the process ends. Otherwise, the algorithm generates more responses based on the refined ones (with some being discarded, some remaining, and some responses potentially being merged).
Through this process, it demonstrated excellent performance in complex scheduling problems (travel planning, meeting scheduling, etc.). It's a viable method for finding highly effective solutions in specific scenarios.
However, there are two major drawbacks:
🤔 An excessive number of API calls are required. (While the cost might not be very high, it leads to significant latency.)
🤔 The evaluator is program-based. (This limits its use as a general method. It could potentially be modified/implemented using LLM as Judge, but that would introduce additional API costs for evaluation.)
https://arxiv.org/abs/2501.09891
The process starts by posing a question.
1) The LLM generates initial responses.
2) These generated responses are evaluated according to specific criteria (program-based checker).
3) The LLM critiques the evaluated results.
4) The LLM refines the responses based on the evaluation, critique, and original responses.
The refined response is then fed back into step 2). If it meets the criteria, the process ends. Otherwise, the algorithm generates more responses based on the refined ones (with some being discarded, some remaining, and some responses potentially being merged).
Through this process, it demonstrated excellent performance in complex scheduling problems (travel planning, meeting scheduling, etc.). It's a viable method for finding highly effective solutions in specific scenarios.
However, there are two major drawbacks:
🤔 An excessive number of API calls are required. (While the cost might not be very high, it leads to significant latency.)
🤔 The evaluator is program-based. (This limits its use as a general method. It could potentially be modified/implemented using LLM as Judge, but that would introduce additional API costs for evaluation.)
https://arxiv.org/abs/2501.09891
Post
2037
Simple Summarization on DeepSeek-R1 from DeepSeek AI
The RL stage is very important.
↳ However, it is difficult to create a truly helpful AI for people solely through RL.
↳ So, we applied a learning pipeline consisting of four stages: providing a good starting point, reasoning RL, SFT, and safety RL, and achieved performance comparable to o1.
↳ Simply fine-tuning other open models with the data generated by R1-Zero (distillation) resulted in performance comparable to o1-mini.
Of course, this is just a brief overview and may not be of much help. All models are accessible on Hugging Face, and the paper can be read through the GitHub repository.
Model: https://huggingface.co/deepseek-ai
Paper: https://github.com/deepseek-ai/DeepSeek-R1
The RL stage is very important.
↳ However, it is difficult to create a truly helpful AI for people solely through RL.
↳ So, we applied a learning pipeline consisting of four stages: providing a good starting point, reasoning RL, SFT, and safety RL, and achieved performance comparable to o1.
↳ Simply fine-tuning other open models with the data generated by R1-Zero (distillation) resulted in performance comparable to o1-mini.
Of course, this is just a brief overview and may not be of much help. All models are accessible on Hugging Face, and the paper can be read through the GitHub repository.
Model: https://huggingface.co/deepseek-ai
Paper: https://github.com/deepseek-ai/DeepSeek-R1
- 1 reply
Post
556
Computational Model for Symbolic Representations: An Interaction Framework for Human-AI Collaboration
Hey everyone. I need your help to see if this concept, scientific logic, and testing with prompts can invalidate or validate it. My goal isn’t to make any bold statements or claims about AI, I just really want to know if I’ve stumbled upon something that can be useful in AI interactions. Here’s my proposal in a nutshell:
The Computational Model for Symbolic Representations Framework introduces a method for enhancing human-AI collaboration by assigning user-defined symbolic representations (glyphs) to guide interactions with computational models. This interaction and syntax is called Glyph Code-Prompting. Glyphs function as conceptual tags or anchors, representing abstract ideas, storytelling elements, or domains of focus (e.g., pacing, character development, thematic resonance). Users can steer the AI’s focus within specific conceptual domains by using these symbols, creating a shared framework for dynamic collaboration. Glyphs do not alter the underlying
The Core Point: Glyphs, acting as collaboratively defined symbols linking related concepts, add a layer of multidimensional semantic richness to user-AI interactions by serving as contextual anchors that guide the AI's focus. This enhances the AI's ability to generate more nuanced and contextually appropriate responses. For instance, a symbol like ! can carry multidimensional semantic meaning and connections, demonstrating the practical value of glyphs in conveying complex intentions efficiently.
Link to my full initial overview and sharing: https://huggingface.co/blog/Severian/computational-model-for-symbolic-representations
Try out the HF Assistant Version: https://hf.co/chat/assistant/678cfe9655026c306f0a4dab
Hey everyone. I need your help to see if this concept, scientific logic, and testing with prompts can invalidate or validate it. My goal isn’t to make any bold statements or claims about AI, I just really want to know if I’ve stumbled upon something that can be useful in AI interactions. Here’s my proposal in a nutshell:
The Computational Model for Symbolic Representations Framework introduces a method for enhancing human-AI collaboration by assigning user-defined symbolic representations (glyphs) to guide interactions with computational models. This interaction and syntax is called Glyph Code-Prompting. Glyphs function as conceptual tags or anchors, representing abstract ideas, storytelling elements, or domains of focus (e.g., pacing, character development, thematic resonance). Users can steer the AI’s focus within specific conceptual domains by using these symbols, creating a shared framework for dynamic collaboration. Glyphs do not alter the underlying
The Core Point: Glyphs, acting as collaboratively defined symbols linking related concepts, add a layer of multidimensional semantic richness to user-AI interactions by serving as contextual anchors that guide the AI's focus. This enhances the AI's ability to generate more nuanced and contextually appropriate responses. For instance, a symbol like ! can carry multidimensional semantic meaning and connections, demonstrating the practical value of glyphs in conveying complex intentions efficiently.
Link to my full initial overview and sharing: https://huggingface.co/blog/Severian/computational-model-for-symbolic-representations
Try out the HF Assistant Version: https://hf.co/chat/assistant/678cfe9655026c306f0a4dab
Post
1616
we now have more than 2000 public AI models using ModelHubMixin🤗
Post
1864
🙋🏻♂️ Hey there folks ,
Facebook AI just released JASCO models that make music stems .
you can try it out here : Tonic/audiocraft
hope you like it
Facebook AI just released JASCO models that make music stems .
you can try it out here : Tonic/audiocraft
hope you like it
Post
4908
🆕 LLM Course 2025 edition!
I updated the LLM Scientist roadmap and added a ton of new information and references. It covers training, datasets, evaluation, quantization, and new trends like test-time compute scaling.
The LLM Course has been incredibly popular (41.3k stars!) and I've been touched to receive many, many messages about how it helped people in their careers.
I know how difficult this stuff can be, so I'm super proud of the impact it had. I want to keep updating it in 2025, especially with the LLM Engineer roadmap.
Thanks everyone, hope you'll enjoy it!
💻 LLM Course: https://huggingface.co/blog/mlabonne/llm-course
I updated the LLM Scientist roadmap and added a ton of new information and references. It covers training, datasets, evaluation, quantization, and new trends like test-time compute scaling.
The LLM Course has been incredibly popular (41.3k stars!) and I've been touched to receive many, many messages about how it helped people in their careers.
I know how difficult this stuff can be, so I'm super proud of the impact it had. I want to keep updating it in 2025, especially with the LLM Engineer roadmap.
Thanks everyone, hope you'll enjoy it!
💻 LLM Course: https://huggingface.co/blog/mlabonne/llm-course
Zero GPU not picking up app.py file
1
#148 opened about 1 month ago
by
Tonic