Who contributed to Hugging Face on July 5, 2026?

4 developers shipped this update, including albertvillanova, gary149, imstevenpmwork, and HuggingFaceInfra.

What were the notable Hugging Face updates?

Add entropy regularization to GRPO, Promote KTO to stable API, and Remove KTO shims from stable.

@huggingface

Transformers, Datasets, and the open AI-model layer

github ↗

Pick a date

Topics: Python AI / ML Full archive →

The Wire · Showcase

GRPO GETS ENTROPY CONTROL, KTO GRADUATES TO STABLE

By RepoJournal · Filed 06:02 UTC on July 5, 2026 · About Hugging Face

4 people shipped this

albertvillanova @albertvillanova 3 cited

gary149 @gary149 1 cited

imstevenpmwork @imstevenpmwork 1 cited

HuggingFaceInfra @HuggingFaceInfra 1 cited

TRL shipped entropy regularization for GRPO while promoting KTO from experimental to production-ready, two moves that tighten the reinforcement learning toolkit.

The GRPO trainer now supports entropy regularization with both static and adaptive control [1], a feature that prevents policy collapse and encourages exploration during training. This closes a known gap in the reinforcement learning pipeline. In parallel, KTO has been promoted from the experimental API to stable [2], and the old shim wrappers have been removed [3] to clean up the codebase. Users importing from `trl.experimental.kto` will see deprecation warnings directing them to the main `trl` namespace. Over at chat-ui, the design team made two precision contrast moves: dimmed prose body text in dark mode [4] and capped emphasis styling on headings and links [5] so text hierarchy reads cleaner without glowing white. The lerobot project completed an automated dependency refresh [6] across CPU and GPU test suites. Documentation stayed in sync with an automated update to the Inference Providers docs [7].

Action items

→ Review GRPO entropy regularization config if you're training policies with exploration concerns huggingface/trl [plan]
→ Update KTO imports from trl.experimental.kto to trl before deprecation window closes huggingface/trl [plan]
→ Monitor chat-ui rendering if you've customized dark mode prose styling huggingface/chat-ui [monitor]

References

[1] Add entropy regularization to GRPO ↗ huggingface/trl
[2] Promote KTO to stable API ↗ huggingface/trl
[3] Remove KTO shims from stable ↗ huggingface/trl
[4] Dim assistant prose body text in dark mode ↗ huggingface/chat-ui
[5] Dim emphasized prose text in dark mode huggingface/chat-ui
[6] chore(dependencies): update uv.lock ↗ huggingface/lerobot
[7] [Bot] Update Inference Providers documentation ↗ huggingface/hub-docs

Quick answers

What shipped in Hugging Face on July 5, 2026?: TRL shipped entropy regularization for GRPO while promoting KTO from experimental to production-ready, two moves that tighten the reinforcement learning toolkit. In total, 6 commits and 7 pull requests landed.
Who contributed to Hugging Face on July 5, 2026?: 4 developers shipped this update, including albertvillanova, gary149, imstevenpmwork, and HuggingFaceInfra.
What were the notable Hugging Face updates?: Add entropy regularization to GRPO, Promote KTO to stable API, and Remove KTO shims from stable.

TRANSFORMERS-CI SHIPS INTELLIGENT TEST TRIAGE, CUTS LLM NOISE 100X

Hugging Face's serge verify loop now distills 100KB failure tracebacks down to the actual assertion, runs every test 5 times to catch flakiness, and dispatches each failure category with a custom instruction before spinning GPU or consuming a single LLM token.

python 45 shipped 2-min read

@huggingface 2 days ago

TRL CI BURNS, BITSANDBYTES PINNED BELOW 0.50.0

Hugging Face's TRL training library is firefighting two separate CI failures overnight, forcing temporary pins on bitsandbytes and xfailing critical tests against a broken transformers dev build.

python 7 shipped 1-min read

@huggingface 3 days ago

TRANSFORMERS OVERHAULS LINEAR ATTENTION WHILE DEPRECATING LEGACY RESPONSE SCHEMA

The transformers library is retiring its fragile response_schema prototype in favor of streaming-compatible parsing, while simultaneously refactoring every linear attention model to use standardized convolution patterns.

python 70 shipped 2-min read

@huggingface 4 days ago

TRANSFORMERS SHIPS FSDP DISTRIBUTED TRAINING STACK, HUB LIBRARY PLUGS REDOS HOLE

Hugging Face landed distributed training orchestration in transformers while plugging a regex vulnerability that could stall untrusted card parsing for minutes.

+10

python 61 shipped 1-min read

Elsewhere on the wire

AI Agents about 21 hours ago

CLINE SHIPS FIVE CRITICAL AGENT FIXES AFTER END-TO-END TESTING BLITZ

Saoudrizwan just landed a cascade of fixes that prevent the agent loop from running wild, stop file reads from resolving to the wrong directory, and plug the compaction pipeline.

ai-agents 16 shipped 2-min read

Local LLMs about 21 hours ago

LLAMA.CPP SHIPS MINIMAX-M3 SPARSE ATTENTION, VLLM UNLOCKS HETEROGENEOUS PREFILL-DECODE

The inference stack gets a new model architecture and a fundamental shift in how prefill and decode can run on different hardware configurations.

+10

llms 125 shipped 1-min read

@CachyOS about 21 hours ago

GLIBC 2.43 REBUILD SHIPS, INSTALLER JUMPS TO C++26

CachyOS pushed glibc 2.43 through the pipeline while the CLI installer upgraded its C++ standard to C++26, signaling a major refresh across core infrastructure.

infra 97 shipped 1-min read

Elixir & Phoenix about 21 hours ago

ECTO FIXES SUBQUERY PLANNING IN JOINS, ELIXIR PATCHES IEX PARSING

Ecto's query planner now handles subqueries correctly in interpolated join filters, while Elixir fixes a syntax error that broke 'not in' as a continuation operator in IEx.