Hugging Face's logo
Activity Feed

AI & ML interests

None defined yet.

Recent Activity
xianbao  submitted a paper about 2 months ago
The Curse and Blessing of Mean Bias in FP4-Quantized LLM Training
sayakpaul  authored a paper about 2 months ago
Fine-Grained Perturbation Guidance via Attention Head Selection
sayakpaul  authored a paper about 2 months ago
From Statics to Dynamics: Physics-Aware Image Editing with Latent Transition Priors
View all activity
qgallouedec 
posted an update 4 days ago
view post
Post
7760

TRL v1.3 ships day-one training support for Qwen 3.6 🚀

The new Qwen 3.6 family (Qwen/Qwen3.6-27B, Qwen/Qwen3.6-35B-A3B) reuses the Qwen3.5-MoE architecture but ships a slightly different chat template, so we updated the stack end-to-end: new training template with {% generation %} markers, tool-call response schema routing, tiny test models for the VLM matrix.

SFT with assistant-only loss works out of the box:

from trl import SFTConfig, SFTTrainer

trainer = SFTTrainer(
    model="Qwen/Qwen3.6-27B",
    args=SFTConfig(assistant_only_loss=True),
    train_dataset=dataset,
)
trainer.train()


So does GRPO tool-calling — just hand tools=[...] to GRPOTrainer.

v1.3 also brings a new experimental TPO trainer (Triple Preference Optimization), speculative decoding in trl vllm-serve (Qwen3 MTP / Eagle3 drafts), 12 more KTO ↔ DPO alignment PRs (KTO promotion to stable is now in reach), three more {% generation %} chat templates (Gemma/Gemma 2, Phi-3, GLM-4-MoE), and a chunky SFT entropy bug fix.

Full release notes: https://github.com/huggingface/trl/releases/tag/v1.3.0
qgallouedec 
posted an update 14 days ago
view post
Post
1880
TRL v1.2 introduces the SSDTrainer 🚀

Simple Self-Distillation (SSD) from Apple's paper "Embarrassingly Simple Self-Distillation Improves Code Generation" is now available as an experimental trainer in TRL.

The recipe is as minimal as the name suggests: sample completions from the model itself at a training-time temperature, then fine-tune on those raw, unverified samples with plain cross-entropy. No reward model. No verifier. No teacher model. No reinforcement learning. Just prompts and the model.

from trl.experimental.ssd import SSDConfig, SSDTrainer

trainer = SSDTrainer(
    model="Qwen/Qwen3-4B-Instruct",
    args=SSDConfig(temperature=0.6, top_k=20, top_p=0.95),
    train_dataset=dataset,
)
trainer.train()


v1.2 also ships expanded tool-calling support (LLaMA 3.1 / 3.2, DeepSeek-V3), another round of KTO ↔ DPO alignment getting us closer to promoting KTO to stable, a big GRPO simplification for overlong tool results, deprecation of use_transformers_paged, and key fixes for VLM response parsing.

Full release notes: https://github.com/huggingface/trl/releases/tag/v1.2.0
victor 
posted an update 17 days ago
view post
Post
5222
Want to share my enthusiasm for zai-org/GLM-5.1 here too 🔥

I think we have it: our open source Claude Code = GLM-5.1 + Pi (https://pi.dev/) - Built a Three.js racing game to eval and it's extremely impressive. Thoughts:

- One-shot car physics with real drift mechanics (this is hard)

- My fav part: Awesome at self iterating (with no vision!) created 20+ Bun.WebView debugging tools to drive the car programmatically and read game state. Proved a winding bug with vector math without ever seeing the screen

- 531-line racing AI in a single write: 4 personalities, curvature map, racing lines, tactical drifting. Built telemetry tools to compare player vs AI speed curves and data-tuned parameters

- All assets from scratch: 3D models, procedural textures, sky shader, engine sounds, spatial AI audio!

- Can do hard math: proved road normals pointed DOWN via vector cross products, computed track curvature normalized by arc length to tune AI cornering speed

You are going to hear about this model a lot in the next months - open source let's go - and thanks z-ai🚀🚀
  • 4 replies
·
tomaarsen 
posted an update 21 days ago
view post
Post
752
🌐 I've just published Sentence Transformers v5.4 to make the project fully multimodal for embeddings and reranking. The release also includes a modular CrossEncoder, and automatic Flash Attention 2 input flattening. Details:

You can now use SentenceTransformer and CrossEncoder with text, images, audio, and video, with the same familiar API. That means you can compute embeddings for an image and a text query using model.encode(), compare them with model.similarity(), and it just works. Models like Qwen3-VL-Embedding-2B and jinaai/jina-reranker-m0 are supported out of the box.

Beyond multimodal, I also fully modularized the CrossEncoder class. It's now a torch.nn.Sequential of composable modules, just like SentenceTransformer has been. This unlocked support for generative rerankers (CausalLM-based models like mxbai-rerank-v2 and the Qwen3 rerankers) via a new LogitScore module, which wasn't possible before without custom code.

Also, Flash Attention 2 now automatically skips padding for text-only inputs. If your batch has a mix of short and long texts, this gives you a nice speedup and lower VRAM usage for free.

I wrote a blog post walking through the multimodal features with practical examples. Check it out if you want to get started, or just point your Agent to the URL: https://huggingface.co/blog/multimodal-sentence-transformers

This release has set up the groundwork for more easily introducing late-interaction models (both text-only and multimodal) into Sentence Transformers in the next major release. I'm looking forward to it!
qgallouedec 
posted an update about 1 month ago
view post
Post
2372
TRL v1.0 is out!

Hugging Face's TRL library is downloaded 3 million times a month. Over 130k models trained with it are public on the Hub, and major projects like @unsloth and @axolotl-ai-co build directly on top of it. v1.0 is the moment we acknowledged that responsibility explicitly, with a real stability contract.

The field hasn't settled. Building stable software in a domain that keeps invalidating its own assumptions is the actual problem we're solving. The answer is a design that can absorb the next shift without breaking what people rely on.

What's in v1.0:
Deep Hugging Face integration, low infrastructure burden
What's next: asynchronous GRPO, better scaling support, and making training legible enough that agents can inspect and steer it.

pip install --upgrade trl


Read more: hf.co/blog/trl-v1
xianbao 
submitted a paper to Daily Papers about 2 months ago
sayakpaul 
authored 2 papers about 2 months ago
albertvillanova 
posted an update 2 months ago
view post
Post
2517
🚀 TRL v0.29.0 introduces trl-training: an agent-native training skill.

This makes the TRL CLI a structured, agent-readable capability, allowing AI agents to reliably execute training workflows such as:
- Supervised Fine-Tuning (SFT)
- Direct Preference Optimization (DPO)
- Group Relative Policy Optimization (GRPO)

We’re excited to see what the community builds on top of this.

If you’re working on AI agents, alignment research, or scalable RL training infrastructure: give TRL v0.29.0 a try! 🤗

The future of ML tooling is agent-native.
🔗 https://github.com/huggingface/trl/releases/tag/v0.29.0
sayakpaul 
submitted a paper to Daily Papers 2 months ago
victor 
submitted a paper to Daily Papers 2 months ago
sayakpaul 
authored a paper 2 months ago
qgallouedec 
posted an update 2 months ago
view post
Post
3032
@CohereLabs just released 🌿 Tiny Aya: a fully open-source 3B parameter model that speaks 70+ languages 🌍! But there’s a catch:

Tiny Aya is just a language model. It doesn’t support tool calling, the key capability that turns frontier models into powerful *agents*.
So the real question is:

How hard is it to turn Tiny Aya into an agent?

Turns out… it’s simple, thanks to Hugging Face TRL.
We’re sharing a hands-on example showing how to train Tiny Aya to turn it into a tool-calling agent using TRL, unlocking what could become the first *massively multilingual open agent*.

Small model. Global reach. Agent capabilities.

👉 https://github.com/huggingface/trl/blob/main/examples/notebooks/sft_tool_calling.ipynb
  • 1 reply
·
lewtun 
submitted 2 papers to Daily Papers 3 months ago
albertvillanova 
posted an update 3 months ago
view post
Post
1894
5 years already working in democratizing AI 🤗
Grateful to be part of such an awesome team making it happen every day.
cfahlgren1 
submitted a paper to Daily Papers 3 months ago
victor 
posted an update 3 months ago
view post
Post
2753
Interesting article: use Claude Code to help open models write CUDA kernels (for eg) by turning CC traces into Skills. They made a library out of it 👀

https://huggingface.co/blog/upskill
yjernite 
authored a paper 3 months ago
pcuenq 
posted an update 4 months ago
view post
Post
4671
👉 What happened in AI in 2025? 👈

We prepared the 2025 version of the HF AI Timeline Grid, highlighting open vs API-based model releases, and allowing you to browse and filter by access, modality, and release type!

Play with it here:
2025-ai-timeline/2025-ai-timeline

Here's my personal quarterly TL;DR:

1️⃣ Q1 — Learning to Reason
Deepseek not only releases a top-notch reasoning model, but shows how to train them and compete with closed frontier models. OpenAI debuts Deep Research.

Significant milestones: DeepSeek R1 & R1-Zero, Qwen 2.5 VL, OpenAI Deep Research, Gemini 2.5 Pro (experimental)

2️⃣ Q2 — Multimodality and Coding
More LLMs embrace multimodality by default, and there's a surge in coding agents. Strong vision, audio, and generative models emerge.

Significant milestones: Llama 4, Qwen 3, Imagen 4, OpenAI Codex, Google Jules, Claude 4

3️⃣ Q3 — "Gold" rush, OpenAI opens up, the community goes bananas
Flagship models get gold in Math olympiads and hard benchmarks. OpenAI releases strong open source models and Google releases the much anticipated nano-banana for image generation and editing. Agentic workflows become commonplace.

Significant milestones: Gemini and OpenAI IMO Gold, gpt-oss, Gemini 2.5 Flash Image, Grok 4, Claude Sonnet 4.5

4️⃣ Q4 — Mistral returns, leaderboard hill-climbing
Mistral is back with updated model families. All labs release impressive models to wrap up the year!

Significant milestones: Claude Opus 4.5, DeepSeek Math V2, FLUX 2, GPT 5.1, Kimi K2 Thinking, Nano Banana Pro, GLM 4.7, Gemini 3, Mistral 3, MiniMax M2.1 🤯

Credits
🙏 NHLOCAL for the source data https://github.com/NHLOCAL/AiTimeline

🫡 @reach-vb for the original idea, design and recipe

🙌 @ariG23498 and yours truly for compiling and verifying the 2025 edition

🥳 Here's to 2026, wishing it becomes the best year ever for open releases and on-device-first use-cases! 🥂
  • 3 replies
·