Who contributed to Hugging Face on July 10, 2026?

3 developers shipped this update, including Rocketknight1, albertvillanova, and yzhautouskay.

What were the notable Hugging Face updates?

Remove executorch from all-latest-gpu image + add torch smoke test (#47196), Make executorch exporter tests always use xnnpack backend (#47201), and Cleanup pipelines, stop materializing generators.

@huggingface

Transformers, Datasets, and the open AI-model layer

github ↗

Pick a date

Topics: Python AI / ML Full archive →

The Wire · Showcase

TRANSFORMERS FIXES GPU IMAGE MELTDOWN, TRL GRADUATES KTO TO STABLE

By RepoJournal · Filed 06:59 UTC on July 10, 2026 · About Hugging Face

3 people shipped this

albertvillanova @albertvillanova 2 cited

Rocketknight1 @Rocketknight1 1 cited

yzhautouskay @yzhautouskay 1 cited

The nightly all-latest-gpu image was broken by a dependency conflict. It's fixed. And KTO just left experimental.

The executorch dependency was uninstalling the pinned cu126 torch in the GPU image, pulling a mismatched CUDA/NCCL stack that crashed at import [1]. That's now removed from the GPU build, and executorch export tests have been reclassified as CPU-only to prevent future cross-stack breakage [2]. In parallel, transformers cleaned up pipeline generators to stop materializing them entirely, which was causing OOMs on complex inputs [3]. Over in TRL, version 1.8.0 promotes KTOTrainer from experimental to stable [4], and trainers now accept DatasetDict and IterableDatasetDict directly as eval_dataset, which means you can pass load_dataset output without reshaping [5]. On the diffusers side, Cosmos3 now has full modular pipeline support with dedicated encoding, denoising, and decoding blocks [6]. In smolagents, transformers 5.13.0 is now pinned out due to a compatibility issue that was breaking CI [7].

Action items

→ If you test nightly GPU builds: pull the latest transformers immediately, the image is usable again huggingface/transformers [immediate]
→ Upgrade to TRL 1.8.0 if you're using KTO, update your imports from experimental to top-level huggingface/trl [plan]
→ Pin transformers != 5.13.0 in any smolagents-dependent projects huggingface/smolagents [plan]
→ Monitor Cosmos3 modular pipeline adoption if you're building video generation workflows huggingface/diffusers [monitor]

References

[1] Remove executorch from all-latest-gpu image + add torch smoke test (#47196) huggingface/transformers
[2] Make executorch exporter tests always use xnnpack backend (#47201) huggingface/transformers
[3] Cleanup pipelines, stop materializing generators ↗ huggingface/transformers
[4] v1.8.0 ↗ huggingface/trl
[5] Support DatasetDict and IterableDatasetDict as eval_dataset in trainers ↗ huggingface/trl
[6] Cosmos3 ModularPipeline ↗ huggingface/diffusers
[7] Fix CI AttributeError: 'str' object has no attribute '__module__' ↗ huggingface/smolagents

Quick answers

What shipped in Hugging Face on July 10, 2026?: The nightly all-latest-gpu image was broken by a dependency conflict. It's fixed. And KTO just left experimental. In total, 24 commits, 25 pull requests, and 1 releases landed.
Who contributed to Hugging Face on July 10, 2026?: 3 developers shipped this update, including Rocketknight1, albertvillanova, and yzhautouskay.
What were the notable Hugging Face updates?: Remove executorch from all-latest-gpu image + add torch smoke test (#47196), Make executorch exporter tests always use xnnpack backend (#47201), and Cleanup pipelines, stop materializing generators.

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.