Who contributed to Hugging Face on August 22, 2026?

12 developers shipped this update, including sayakpaul, akshan-main, dg845, darthjaja6, hq-fang, pkooij, moon-bot-app[bot], and moon-bot-app, and 4 more.

What were the notable Hugging Face updates?

perf(fastwam): add compiled cached action inference, fix(molmoact2): align continuous training with official recipe, and feat(pi05): add optional training-time RTC.

@huggingface

Transformers, Datasets, and the open AI-model layer

github ↗

Pick a date

Topics: Python AI / ML Full archive →

The Wire · Showcase

LEROBOT CUTS FASTWAM OVERHEAD WITH COMPILED INFERENCE

By RepoJournal · Filed 06:59 UTC on August 22, 2026 · About Hugging Face

12 people shipped this

darthjaja6 @darthjaja6 1 cited

hq-fang @hq-fang 1 cited

pkooij @pkooij 1 cited

moon-bot-app[bot] @moon-bot-app[bot] 1 cited

Pierrci @Pierrci 1 cited

sergiopaniego @sergiopaniego 1 cited

albertvillanova @albertvillanova 1 cited

sayakpaul @sayakpaul

akshan-main @akshan-main

dg845 @dg845

moon-bot-app @moon-bot-app

HuggingFaceInfra @HuggingFaceInfra

LeRobot's FastWAM policy gets an opt-in compiled path that slashes Python and CUDA kernel-launch overhead on cached action inference.

LeRobot shipped an opt-in compiled inference path for FastWAM that cuts Python and CUDA kernel-launch overhead on cached action inference, a win for anyone running repeated denoising steps on the same video features [1]. The same repo tightened MolmoAct2 continuous training to match the official recipe, adopting a BF16 VLM / FP32 action-expert precision policy and the official 30K scheduler clock [2]. On the policy side, Pi05 now supports training-time Real-Time Chunking behind a flag that defaults to 0, so existing pipelines stay untouched [3]. Outside LeRobot, a stale Pi install command on model pages has been fixed: huggingface.js now points to the new @earendil-works/pi-coding-agent package, which sits at 0.84.2 versus the frozen 0.73.1 release [4]. Hub docs clarify that PRO subscription credits are now general-purpose compute credits, applied automatically before pay-as-you-go, usable across Inference Endpoints, Spaces, and Jobs [5]. TRL fixed a CLI parsing bug where --scale_rewards group failed because SDPOConfig typed it as bool; it now matches GRPOConfig's string type [6]. The async trainer tests in TRL also upgraded to a Flash Attention-compatible model so head_size is a multiple of 8, un-xfailing a set of tests that were expected to fail [7].

Action items

→ Test FastWAM with the new compiled action inference path to measure latency gains huggingface/lerobot [plan]
→ Update any stale Pi install commands to @earendil-works/pi-coding-agent huggingface/huggingface.js [immediate]
→ Review MolmoAct2 fine-tuning scripts to match the aligned official recipe huggingface/lerobot [monitor]

References

[1] perf(fastwam): add compiled cached action inference ↗ huggingface/lerobot
[2] fix(molmoact2): align continuous training with official recipe ↗ huggingface/lerobot
[3] feat(pi05): add optional training-time RTC ↗ huggingface/lerobot
[4] Update Pi local-app snippet to the new package name and repo URL (#2390) huggingface/huggingface.js
[5] included sub credits are now compute credits (#2719) huggingface/hub-docs
[6] Fix SDPO scale_rewards CLI parsing and warn on ignored feedback ↗ huggingface/trl
[7] Use a Flash Attention compatible model in the async trainer tests ↗ huggingface/trl

Quick answers

What shipped in Hugging Face on August 22, 2026?: LeRobot's FastWAM policy gets an opt-in compiled path that slashes Python and CUDA kernel-launch overhead on cached action inference. In total, 18 commits and 18 pull requests landed.
Who contributed to Hugging Face on August 22, 2026?: 12 developers shipped this update, including sayakpaul, akshan-main, dg845, darthjaja6, hq-fang, pkooij, moon-bot-app[bot], and moon-bot-app, and 4 more.
What were the notable Hugging Face updates?: perf(fastwam): add compiled cached action inference, fix(molmoact2): align continuous training with official recipe, and feat(pi05): add optional training-time RTC.

Transformers Brings Biology to v5 Main

Hugging Face's flagship repo hauled two protein-folding models onto the v5 mainline, and a wav2vec2 attention fix is ready to unblock your SDPA workflows.

python 33 shipped

@huggingface Aug 20

TRL DROPS ASYNC DISTILLATION, DIFFUSERS GO TENSOR PARALLEL

HuggingFace's training and inference stacks moved in lockstep overnight.

python 22 shipped

@huggingface Aug 19

TWO NEW TOOLCHAINS AND A PATCH THAT MATTERS

Hugging Face's kernels repo brings Intel's next-gen GPU into the fold, while Transformers quietly fixes a cross-attention bug that could have been eating your encoder context.

python 84 shipped

@huggingface Aug 18

Harbor-HF Walls Off Sandboxes While Serge Patrols Its Gates

Overnight, Hugging Face's internal infrastructure hardened its edge: Harbor-HF locks down sandbox capabilities, Serge's verify workflow stops dispatching with empty inputs, and dataset-viewer workers finally log their own startups.

python 76 shipped

Elsewhere on the wire

AI Agents Aug 22

Cline hubs get drain and replay as Langchain plumb gates

Cline just shipped rollover-grade hub commands while Langchain exposes gateway internals, and OpenHands quietly broke its own monitoring contract.

ai-agents 42 shipped

Local LLMs Aug 22

SGLang ships v0.5.18 with 710 PRs and DeepSeek V4 FP4 auto-routing

The biggest release of the week just landed: SGLang v0.5.18 adds new models, fixes a silent disagg bug, and defaults DeepSeek V4 FP4 checkpoints to the right MoE runner.