Hugging Face's logo

Hugging Face H4

Enterprise
company
https://github.com/huggingface/alignment-handbook
Activity Feed

AI & ML interests

Aligning LLMs to be helpful, honest, harmless, and huggy (H4)

Recent Activity

lewtun  updated a collection about 8 hours ago
Scaling Test-Time Compute with Open Models
lewtun  updated a Space about 8 hours ago
HuggingFaceH4/blogpost-scaling-test-time-compute
lewtun  new activity 1 day ago
HuggingFaceH4/blogpost-scaling-test-time-compute:Link to the community tab in the blogpost
View all activity

HuggingFaceH4's activity

lewtun 
updated a collection about 8 hours ago
lewtun 
updated a Space about 8 hours ago
merve 
posted an update about 18 hours ago
view post
Post
679
supercharge your LLM apps with smolagents 🔥

however cool your LLM is, without being agentic it can only go so far

enter smolagents: a new agent library by Hugging Face to make the LLM write code, do analysis and automate boring stuff!

Here's our blog for you to get started https://huggingface.co/blog/smolagents
lewtun 
in HuggingFaceH4/blogpost-scaling-test-time-compute 1 day ago

Link to the community tab in the blogpost

1
#3 opened 16 days ago by
clem
lewtun 
posted an update 2 days ago
view post
Post
1704
This paper ( HuatuoGPT-o1, Towards Medical Complex Reasoning with LLMs (2412.18925)) has a really interesting recipe for inducing o1-like behaviour in Llama models:

* Iteratively sample CoTs from the model, using a mix of different search strategies. This gives you something like Stream of Search via prompting.
* Verify correctness of each CoT using GPT-4o (needed because exact match doesn't work well in medicine where there are lots of aliases)
* Use GPT-4o to reformat the concatenated CoTs into a single stream that includes smooth transitions like "hmm, wait" etc that one sees in o1
* Use the resulting data for SFT & RL
* Use sparse rewards from GPT-4o to guide RL training. They find RL gives an average ~3 point boost across medical benchmarks and SFT on this data already gives a strong improvement.

Applying this strategy to other domains could be quite promising, provided the training data can be formulated with verifiable problems!
merve 
posted an update 8 days ago
regisss 
posted an update 13 days ago
freddyaboulton 
posted an update 14 days ago
merve 
posted an update 14 days ago
view post
Post
2702
Aya by Cohere For AI can now see! 👀

C4AI community has built Maya 8B, a new open-source multilingual VLM built on SigLIP and Aya 8B 🌱 works on 8 languages! 🗣️

The authors extend Llava dataset using Aya's translation capabilities with 558k examples!
ry it here kkr5155/maya_demo

Dataset maya-multimodal/pretrain

Model maya-multimodal/maya 👏
kudos @nahidalam and team
  • 1 reply
·
freddyaboulton 
posted an update 15 days ago
merve 
posted an update 15 days ago
view post
Post
3115
Apollo is a new family of open-source video language models by Meta, where 3B model outperforms most 7B models and 7B outperforms most 30B models 🧶

✨ the models come in 1.5B https://huggingface.co/Apollo-LMMs/Apollo-1_5B-t32, 3B https://huggingface.co/Apollo-LMMs/Apollo-3B-t32 and 7B https://huggingface.co/Apollo-LMMs/Apollo-7B-t32 with A2.0 license, based on Qwen1.5 & Qwen2
✨ the authors also release a benchmark dataset https://huggingface.co/spaces/Apollo-LMMs/ApolloBench

The paper has a lot of experiments (they trained 84 models!) about what makes the video LMs work ⏯️

Try the demo for best setup here https://huggingface.co/spaces/Apollo-LMMs/Apollo-3B
they evaluate sampling strategies, scaling laws for models and datasets, video representation and more!
> The authors find out that whatever design decision was applied to small models also scale properly when the model and dataset are scaled 📈 scaling dataset has diminishing returns for smaller models
> They evaluate frame sampling strategies, and find that FPS sampling is better than uniform sampling, and they find 8-32 tokens per frame optimal
> They also compare image encoders, they try a variation of models from shape optimized SigLIP to DINOv2
they find google/siglip-so400m-patch14-384 to be most powerful 🔥
> they also compare freezing different parts of models, training all stages with some frozen parts give the best yield

They eventually release three models, where Apollo-3B outperforms most 7B models and Apollo 7B outperforms 30B models 🔥
·
lewtun 
posted an update 16 days ago
view post
Post
6581
We outperform Llama 70B with Llama 3B on hard math by scaling test-time compute 🔥

How? By combining step-wise reward models with tree search algorithms :)

We show that smol models can match or exceed the performance of their much larger siblings when given enough "time to think"

We're open sourcing the full recipe and sharing a detailed blog post.

In our blog post we cover:

📈 Compute-optimal scaling: How we implemented DeepMind's recipe to boost the mathematical capabilities of open models at test-time.

🎄 Diverse Verifier Tree Search (DVTS): An unpublished extension we developed to the verifier-guided tree search technique. This simple yet effective method improves diversity and delivers better performance, particularly at large test-time compute budgets.

🧭 Search and Learn: A lightweight toolkit for implementing search strategies with LLMs and built for speed with vLLM

Here's the links:

- Blog post: HuggingFaceH4/blogpost-scaling-test-time-compute

- Code: https://github.com/huggingface/search-and-learn

Enjoy!
  • 2 replies
·
lewtun 
in HuggingFaceH4/blogpost-scaling-test-time-compute 16 days ago

Use HF papers

1
#2 opened 16 days ago by
qgallouedec

Update app/src/index.html

1
#1 opened 16 days ago by
qgallouedec
qgallouedec 
in HuggingFaceH4/blogpost-scaling-test-time-compute 16 days ago

Use HF papers

1
#2 opened 16 days ago by
qgallouedec
qgallouedec 
updated a Space 16 days ago
qgallouedec 
in HuggingFaceH4/blogpost-scaling-test-time-compute 16 days ago

Update app/src/index.html

1
#1 opened 16 days ago by
qgallouedec
lewtun 
updated a dataset 16 days ago
lewtun 
updated a collection 19 days ago