benchmark-test (Testing Benchmarks on the Hub)

lewtun

posted an update 1 day ago

Post

2177

Introducing OpenR1-Math-220k!

open-r1/OpenR1-Math-220k

The community has been busy distilling DeepSeek-R1 from inference providers, but we decided to have a go at doing it ourselves from scratch 💪

What’s new compared to existing reasoning datasets?

♾ Based on AI-MO/NuminaMath-1.5: we focus on math reasoning traces and generate answers for problems in NuminaMath 1.5, an improved version of the popular NuminaMath-CoT dataset.

🐳 800k R1 reasoning traces: We generate two answers for 400k problems using DeepSeek R1. The filtered dataset contains 220k problems with correct reasoning traces.

📀 512 H100s running locally: Instead of relying on an API, we leverage vLLM and SGLang to run generations locally on our science cluster, generating 180k reasoning traces per day.

⏳ Automated filtering: We apply Math Verify to only retain problems with at least one correct answer. We also leverage Llama3.3-70B-Instruct as a judge to retrieve more correct examples (e.g for cases with malformed answers that can’t be verified with a rules-based parser)

📊 We match the performance of DeepSeek-Distill-Qwen-7B by finetuning Qwen-7B-Math-Instruct on our dataset.

🔎 Read our blog post for all the nitty gritty details: https://huggingface.co/blog/open-r1/update-2

lewtun

authored a paper 5 days ago

SmolLM2: When Smol Goes Big -- Data-Centric Training of a Small Language Model

Paper • 2502.02737 • Published 7 days ago • 153

lewtun

posted an update 17 days ago

Post

10040

We are reproducing the full DeepSeek R1 data and training pipeline so everybody can use their recipe. Instead of doing it in secret we can do it together in the open!

🧪 Step 1: replicate the R1-Distill models by distilling a high-quality reasoning corpus from DeepSeek-R1.

🧠 Step 2: replicate the pure RL pipeline that DeepSeek used to create R1-Zero. This will involve curating new, large-scale datasets for math, reasoning, and code.

🔥 Step 3: show we can go from base model -> SFT -> RL via multi-stage training.

Follow along: https://github.com/huggingface/open-r1

5 replies

·

lewtun

posted an update about 1 month ago

Post

3831

I was initially pretty sceptical about Meta's Coconut paper [1] because the largest perf gains were reported on toy linguistic problems. However, these results on machine translation are pretty impressive!

https://x.com/casper_hansen_/status/1875872309996855343

Together with the recent PRIME method [2] for scaling RL, reasoning for open models is looking pretty exciting for 2025!

[1] Training Large Language Models to Reason in a Continuous Latent Space (2412.06769)
[2] https://huggingface.co/blog/ganqu/prime

lewtun

posted an update about 1 month ago

Post

2280

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!

1 reply

·

lewtun

posted an update about 2 months ago

Post

6853

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

posted an update 10 months ago

Post

5081

Introducing Zephyr 141B-A35B 🪁:

HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1

Yesterday, Mistral released their latest base model (via magnet link of course 😅) and the community quickly converted it to transformers format and pushed it to the Hub: mistral-community/Mixtral-8x22B-v0.1

Early evals of this model looked extremely strong, so we teamed up with Argilla and KAIST AI to cook up a Zephyr recipe with a few new alignment techniques that came out recently:

🧑‍🍳 Align the base model with Odds Ratio Preference Optimisation (ORPO). This novel algorithm developed by @JW17 and @nlee-208 and @j6mes and does not require an SFT step to achieve high performance and is thus much more computationally efficient than methods like DPO and PPO.

🦫 Use a brand new dataset of 7k high-quality, multi-turn preferences that has been developed by our friends at Argilla. To create this dataset, they took the excellent Capybara SFT dataset from @LDJnr LDJnr/Capybara and converted it into a preference dataset by augmenting the final turn with responses from new LLMs that were then ranked by GPT-4.

What we find especially neat about this approach is that training on 7k samples only takes ~1.3h on 4 H100 nodes, yet produces a model that is very strong on chat benchmarks like IFEval and BBH.

Kudos to @alvarobartt @JW17 and @nlee-208 for this very nice and fast-paced collab!

For more details on the paper and dataset, checkout our collection: HuggingFaceH4/zephyr-orpo-6617eba2c5c0e2cc3c151524

lewtun

posted an update 11 months ago

Post

Can we align code generation models to be good at chat without compromising their base capabilities 🤔?

This was the question the H4 team asked itself when BigCode released StarCoder2 a bit over a week ago. We knew that code models like deepseek-ai/deepseek-coder-6.7b-instruct and m-a-p/OpenCodeInterpreter-DS-33B get impressive scores on code benchmarks like HumanEval, but they tend to score poorly on chat benchmarks like MT Bench and IFEval. We also knew that the Zephyr recipe we applied to Mistral 7B produced a strong chat model, so we wondered -- could be tweaked to produce a strong coding assistant?

It turns out the answer is yes and I'm happy to share StarChat2, a DPO fine-tune of StarCoder2 15B that scores highly on both HumanEval and MT Bench / IFEval 🌟!

The most interesting lesson for me was that you get better models by blending in more code/math data than chat during the SFT step - in terms of tokens, we found a ratio of 3:1 worked best.

Anyway, here's a demo of the model, along with all the code and datasets we used to train it:

* Demo: HuggingFaceH4/starchat2-playground
* Collection: HuggingFaceH4/starchat2-15b-65f068417b330fafad751fce
* Recipe: https://github.com/huggingface/alignment-handbook

Hope it's useful to others!