Hugging Face

Enterprise
company
Verified
Activity Feed

AI & ML interests

The AI community building the future.

Recent Activity

sayakpaulΒ  updated a dataset about 14 hours ago
huggingface/diffusers-metadata
evijitΒ  updated a dataset about 17 hours ago
huggingface/policy-docs
nielsrΒ  updated a Space about 21 hours ago
huggingface/ai-deadlines
View all activity

Articles

huggingface's activity

m-ricΒ 
posted an update about 14 hours ago

[FEEDBACK] Follow

14
#14 opened over 1 year ago by
victor
burtenshawΒ 
posted an update 2 days ago
view post
Post
1482
The open LLM leaderboard is completed, retired, dead, β€˜ascended to a higher plane’. And in its shadow we have an amazing range of leaderboards built and maintained by the community.

In this post, I just want to list some of those great leaderboards that you should bookmark for staying up to date:

- Chatbot Arena LLM Leaderboard is the first port of call for checking out the best model. It’s not the fastest because humans will need to use the models to get scores, but it’s worth the wait. lmarena-ai/chatbot-arena-leaderboard

- OpenVLM Leaderboard is great for getting scores on vision language models opencompass/open_vlm_leaderboard

- Ai2 are doing a great job on RewardBench and I hope they keep it up because reward models are the unsexy workhorse of the field. allenai/reward-bench

- The GAIA leaderboard is great for evaluating agent applications. gaia-benchmark/leaderboard

🀩 This seems like such a sustainable way of building for the long term, where rather than leaning on a single company to evaluate all LLMs, we share the load.
  • 1 reply
Β·
burtenshawΒ 
posted an update 3 days ago
view post
Post
1592
Still speed running Gemma 3 to think. Today I focused on setting up gpu poor hardware to run GRPO.

This is a plain TRL and PEFT notebook which works on mac silicone or colab T4. This uses the 1b variant of Gemma 3 and a reasoning version of GSM8K dataset.

πŸ§‘β€πŸ³ There’s more still in the oven like releasing models, an Unsloth version, and deeper tutorials, but hopefully this should bootstrap your projects.

Here’s a link to the 1b notebook: https://colab.research.google.com/drive/1mwCy5GQb9xJFSuwt2L_We3eKkVbx2qSt?usp=sharing
  • 1 reply
Β·
burtenshawΒ 
posted an update 3 days ago
view post
Post
1532
everybody and their dog is fine-tuning Gemma 3 today, so I thought I'd do a longer post on the tips and sharp edges I find. let's go!

1. has to be install everything form main and nightly. this is what I'm working with to get unsloth and TRL running

git+https://github.com/huggingface/transformers@main
git+https://github.com/huggingface/trl.git@main
bitsandbytes
peft


plus this with --no-deps

git+https://github.com/unslothai/unsloth-zoo.git@nightly
git+https://github.com/unslothai/unsloth.git@nightly


2. will brown's code to turn GSM8k into a reasoning dataset is a nice toy experiment https://gist.github.com/willccbb/4676755236bb08cab5f4e54a0475d6fb

3. with a learning rate of 5e-6 rewards and loss stayed flat for the first 100 or so steps.

4. so far none of my runs have undermined the outputs after 1 epoch. therefore, I'm mainly experimenting with bigger LoRA adapters.

from trl import GRPOConfig

training_args = GRPOConfig(
    learning_rate = 5e-6,
    adam_beta1 = 0.9,
    adam_beta2 = 0.99,
    weight_decay = 0.1,
    warmup_ratio = 0.1,
    lr_scheduler_type = "cosine",
    optim = "adamw_8bit",
    logging_steps = 1,
    per_device_train_batch_size = 2,
    gradient_accumulation_steps = 1,
    num_generations = 2,
    max_prompt_length = 256,
    max_completion_length = 1024 - 256,
    num_train_epochs = 1,
    max_steps = 250,
    save_steps = 250,
    max_grad_norm = 0.1,
    report_to = "none",
)


5. vision fine-tuning isn't available in TRL's GRPOTrainer, so stick to text datasets. but no need to load the model differently in transformers or Unsloth

from transformers import AutoModelForImageTextToText

model = AutoModelForImageTextToText.from_pretrained("google/gemma-3-4b-it)


if you want an introduction to GRPO, check out the reasoning course, it walks you through the algorithm, theory, and implementation in a smooth way.

https://huggingface.co/reasoning-course
  • 2 replies
Β·
jsulzΒ 
posted an update 3 days ago
view post
Post
1220
It's finally here ❀️

Build faster than ever with lightning fast upload and download speeds starting today on the Hub ⚑

Xet storage is rolling out access across the Hub - join the waitlist here https://huggingface.co/join/xet

You can apply for yourself, or your entire organization. Head over to your account settings for more information or join anywhere you see the Xet logo on a repository you know.

Have questions? Join the conversation below πŸ‘‡ or open a discussion on the Xet team page xet-team/README
Β·
burtenshawΒ 
posted an update 4 days ago
view post
Post
1690
Here’s a notebook to make Gemma reason with GRPO & TRL. I made this whilst prepping the next unit of the reasoning course:

In this notebooks I combine together google’s model with some community tooling

- First, I load the model from the Hugging Face hub with transformers’s latest release for Gemma 3
- I use PEFT and bitsandbytes to get it running on Colab
- Then, I took Will Browns processing and reward functions to make reasoning chains from GSM8k
- Finally, I used TRL’s GRPOTrainer to train the model

Next step is to bring Unsloth AI in, then ship it in the reasoning course. Links to notebook below.

https://colab.research.google.com/drive/1Vkl69ytCS3bvOtV9_stRETMthlQXR4wX?usp=sharing
Β·
eliebakΒ 
posted an update 4 days ago
view post
Post
1375
Google just dropped an exciting technical report for the brand-new Gemma3 model! πŸš€ Here are my personal notes highlighting the most intriguing architectural innovations, design choices, and insights from this release:

1) Architecture choices:
> No more softcaping, replace by QK-Norm
> Both Pre AND Post Norm
> Wider MLP than Qwen2.5, ~ same depth
> SWA with 5:1 and 1024 (very small and cool ablation on the paper!)
> No MLA to save KV cache, SWA do the job!

2) Long context
> Only increase the rope in the global layer (to 1M)
> Confirmation that it's harder to do long context for smol models, no 128k for the 1B
> Pretrained with 32k context? seems very high
> No yarn nor llama3 like rope extension

3) Distillation
> Only keep te first 256 logits for the teacher
> Ablation on the teacher gap (tl;dr you need some "patience" to see that using a small teacher is better)
> On policy distillation yeahh (by
@agarwl_
et al), not sure if the teacher gap behave the same here, curious if someone have more info?

4) Others
> Checkpoint with QAT, that's very cool
> RL using improve version of BOND, WARM/WARP good excuse to look at
@ramealexandre
papers
> Only use Zero3, no TP/PP if i understand correctly ?
> Training budget relatively similar than gemma2
  • 1 reply
Β·
clefourrierΒ 
posted an update 4 days ago
view post
Post
1583
Gemma3 family is out! Reading the tech report, and this section was really interesting to me from a methods/scientific fairness pov.

Instead of doing over-hyped comparisons, they clearly state that **results are reported in a setup which is advantageous to their models**.
(Which everybody does, but people usually don't say)

For a tech report, it makes a lot of sense to report model performance when used optimally!
On leaderboards on the other hand, comparison will be apples to apples, but in a potentially unoptimal way for a given model family (like some user interact sub-optimally with models)

Also contains a cool section (6) on training data memorization rate too! Important to see if your model will output the training data it has seen as such: always an issue for privacy/copyright/... but also very much for evaluation!

Because if your model knows its evals by heart, you're not testing for generalization.
freddyaboultonΒ 
posted an update 4 days ago
view post
Post
1687
Privacy matters when talking to AI! πŸ”‡

We've just added a microphone mute button to FastRTC in our latest update (v0.0.14). Now you control exactly what your LLM hears.

Plus lots more features in this release! Check them out:
https://github.com/freddyaboulton/fastrtc/releases/tag/0.0.14
lewtunΒ 
posted an update 4 days ago
view post
Post
1883
Introducing OlympicCoder: a series of open reasoning models that can solve olympiad-level programming problems πŸ§‘β€πŸ’»

- 7B open-r1/OlympicCoder-7B
- 32B open-r1/OlympicCoder-32B

We find that OlympicCoder models outperform Claude 3.7 Sonnet, as well as others over 100x larger πŸ’ͺ

Together with the models, we are releasing:

πŸ“ŠCodeForces-CoTs: new dataset of code problems from the most popular competitive coding platform, with R1 traces in C++ and Python open-r1/codeforces-cots

πŸ† IOI'2024: a new benchmark of VERY hard programming problems where even frontier models struggle to match human performance open-r1/ioi

For links to the models and datasets, check out our latest progress report from Open R1: https://huggingface.co/blog/open-r1/update-3
  • 1 reply
Β·