chore(deps): update loader dependencies major (major)#194
Open
dreadnode-renovate-bot[bot] wants to merge 1 commit intomainfrom
Open
chore(deps): update loader dependencies major (major)#194dreadnode-renovate-bot[bot] wants to merge 1 commit intomainfrom
dreadnode-renovate-bot[bot] wants to merge 1 commit intomainfrom
Conversation
dreadnode-renovate-bot
bot
force-pushed
the
renovate/major-loader-deps-major
branch
3 times, most recently
from
07525d6 to
3ac3e72
Compare
| datasource | package | from | to | | ---------- | ---------------------- | ------ | ----- | | docker | nvcr.io/nvidia/pytorch | 24.12 | 26.02 | | pypi | psutil | 6.1.1 | 7.2.2 | | pypi | transformers | 4.57.6 | 5.3.0 |
dreadnode-renovate-bot
bot
force-pushed
the
renovate/major-loader-deps-major
branch
from
3ac3e72 to
4daa5d1
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR contains the following updates:
24.12-py3→26.02-py3==6.1.1→==7.2.2==4.57.6→==5.3.0Release Notes
giampaolo/psutil (psutil)
v7.2.2Compare Source
=====
2026-01-28
Enhancements
Process.wait()_ now usespidfd_open()+poll()forwaiting, resulting in no busy loop and faster response times. Requires
Linux >= 5.3 and Python >= 3.9. Falls back to traditional polling if
unavailable.
Process.wait()_ now useskqueue()for waiting,resulting in no busy loop and faster response times.
Bug fixes
Fedorov)
Process.memory_full_info()andProcess.threads()._ andProcess.environ()_ may fail with ``OSError: [Errno 0] Undefined error`` (from ``sysctl(KERN_PROCARGS2)``). They now raiseAccessDenied`_ instead.v7.2.1Compare Source
=====
2025-12-29
Bug fixes
heap_info()_ does not detect small allocations(<= 1K). In order to fix that, we now flush internal jemalloc cache before
fetching the metrics.
v7.2.0Compare Source
=====
2025-12-23
Enhancements
heap_info()_ andheap_trim()_ functions, providing directaccess to the platform's native C heap allocator (glibc, mimalloc,
libmalloc). Useful to create tools to detect memory leaks.
now live under
tests/instead ofpsutil/tests.Bug fixes
include.
net_if_stats()_ due to missingPy_CLEAR.Compatibility notes
import psutil.testsno longer works (but it was never documented tobegin with).
v7.1.3Compare Source
=====
2025-11-02
Enhancements
clang-formaton all C and header files. It is now themandatory formatting style for all C sources.
raise the appropriate exception (
ZombieProcess_).sprintf/snprintf/sprintf_scalls withstr_format(). Replacestrlcat/strlcpywith safestr_copy/str_append. This unifies string handling across platforms and reducesunsafe usage of standard string functions, improving robustness.
Bug fixes
disk_usage()_ could truncate values on 32-bit platforms,potentially reporting incorrect total/free/used space for drives larger than
4GB.
Process.status()_ incorrectly returns "running" for 99%of the processes.
net_if_addrs()_.Previously, the MAC address buffer was incorrectly updated using a fixed
increment and
sprintf_s, which could overflow or misformat thestring if the MAC length or formatting changed. Also, the final '\n' was
inserted unnecessarily.
v7.1.2Compare Source
=====
2025-10-25
Enhancements
32-bit CPython is still supported, but psutil must now be built from source.
2565_: produce wheels for free-thread cPython 3.13 and 3.14 (patch by
Lysandros Nikolaou)
Bug fixes
Process.cmdline()_ andProcess.environ()_ may incorrectlyraise
NoSuchProcess_ instead ofZombieProcess_.free()inProcess.environ()_ when it failsinternally. This posed a risk of segfault.
which were (not so) sporadically spotted on CI.
Compatibility notes
v7.1.1Compare Source
=====
2025-10-19
Enhancements
Bug fixes
Process.cmdline()_ does not handle spaces properly. (patchby Ben Raz)
Compatibility notes
v7.1.0Compare Source
=====
2025-09-17
Enhancements
maintained from 2009 to 2013.
dprintCLI tool to format .yml and .md files.Bug fixes
Process.memory_maps(),Process.exe()andProcess.open_files()_ doesnot properly handle UNC paths. Paths such as
\\??\\C:\\Windows\\Tempand'\\Device\\HarddiskVolume1\\Windows\\Temp'are now converted toC:\\Windows\\Temp. (patch by Ben Peddell)special characters in their name.
Process.cwd()_ sometimes fail withFileNotFoundErrordueto a race condition.
Process.create_time(), which is used to univocallyidentify a process over time, is subject to system clock updates, and as such
can lead to
Process.is_running()returning a wrong result. A monotoniccreation time is now used instead. (patch by Jonathan Kohler)
Process.children()_ may raisePermissionError. It willnow raise
AccessDenied_ instead.boot_time()_ is off by 45 seconds (C precision issue).Process.create_time()_ doesnot reflect system clock updates.
Process.children()_ andProcess.parent()_ may not be able to return the right information.net_connections()_.boot_time()_ didn't take into account the time spentduring suspend / hibernation.
Process.memory_maps()_ may crash withIndexErroronRISCV64 due to a malformed
/proc/{PID}/smapsfile. (patch by JulienStephan)
trigger a segfault.
virtual_memory()_ "used" memory does not match recentversions of
freeCLI utility. (patch by Isaac K. Ko)psutil.sensors_battery()reports a negative amount forseconds left.
WindowsService.description()method may fail withERROR_NOT_FOUND. Now it returns an empty string instead.cpu_freq()_ segfault on ARM architectures.Compatibility notes
v7.0.0Compare Source
=====
2025-02-13
Enhancements
net_if_addrs()_ also returns thebroadcastaddressinstead of
None.2.7 is psutil 6.1.X. Install it with:
pip2 install psutil==6.1.*.Process.memory_info_ex()method. It wasdeprecated in psutil 4.0.0, released 8 years ago. Substitute is
Process.memory_full_info().Bug fixes
Process.memory_maps()for processes that use hundreds of GBs of memory.
virtual_memory()_ now relies onhost_statistics64instead of
host_statistics. This is the same approach used byvm_statCLI tool, and should grant more accurate results.
Compatibility notes
Process.memory_info_ex()method.huggingface/transformers (transformers)
v5.3.0: : EuroBERT, VibeVoice ASR, TimesFM2.5, PP-DocLayoutV2, OlmoHybrid, ModernVBert, Higgs Audio V2Compare Source
New Model additions
EuroBERT
EuroBERT is a multilingual encoder model based on a refreshed transformer architecture, akin to Llama but with bidirectional attention. It supports a mixture of European and widely spoken languages, with sequences of up to 8192 tokens.
Links: Documentation | Paper | Blog Post
VibeVoice ASR
VibeVoice ASR is an automatic speech recognition model from Microsoft that combines acoustic and semantic audio tokenizers with a causal language model for robust speech-to-text transcription. The model uses VibeVoice's acoustic and semantic tokenizers that process audio at 24kHz, paired with a Qwen2-based language decoder for generating transcriptions. It can process up to 60 minutes of continuous audio input, supports customized hotwords, performs joint ASR/diarization/timestamping, and handles over 50 languages with code-switching support.
Links: Documentation | Paper
TimesFM2.5
TimesFM 2.5 is a pretrained time-series foundation model that uses a decoder-only attention architecture with input patching for forecasting. The model is designed to provide accurate zero-shot forecasts across different domains, forecasting horizons and temporal granularities without requiring dataset-specific training. It builds on the original TimesFM architecture with enhancements including rotary attention, QK normalization, per-dimension attention scaling, and continuous quantile prediction.
Links: Documentation | Paper
PP-DocLayoutV2
PP-DocLayoutV2 is a dedicated lightweight model for layout analysis, focusing specifically on element detection, classification, and reading order prediction. The model is composed of two sequentially connected networks: an RT-DETR-based detection model that performs layout element detection and classification, followed by a pointer network that orders these layout elements. It is designed to analyze document layouts by identifying and organizing various layout components in their proper reading sequence.
Links: Documentation
OlmoHybrid
OLMo Hybrid is a hybrid architecture model from Ai2 that combines standard transformer attention layers with linear attention layers using the Gated Deltanet. This hybrid approach aims to improve efficiency while maintaining model quality by interleaving full attention layers with linear attention layers. The model uses a custom cache system that handles both KV cache for attention layers and recurrent state for linear attention layers.
Links: Documentation
ModernVBert
ModernVBert is a Vision-Language encoder that combines ModernBert with a SigLIP vision encoder. It is optimized for visual document understanding and retrieval tasks, making it suitable for processing documents that contain both text and visual elements.
Links: Documentation | Paper
ColModernVBert
ColModernVBert is a model for efficient visual document retrieval that leverages ModernVBert to construct multi-vector embeddings directly from document images, following the ColPali approach. The model enables retrieval and scoring of visual documents by processing both text queries and document images to generate embeddings that can be compared for relevance scoring.
Links: Documentation | Paper
Higgs Audio V2
Higgs Audio V2 is a powerful audio foundation model developed by Boson AI that was pretrained on over 10 million hours of audio data and diverse text data. Despite having no post-training or fine-tuning, the model excels in expressive audio generation thanks to its deep language and acoustic understanding. The model supports various audio generation tasks including single-speaker and multi-speaker smart voice, zero-shot voice cloning, and multi-speaker voice cloning.
Links: Documentation
Higgs Audio V2 Tokenizer
The Higgs Audio V2 Tokenizer is an audio tokenization model that operates at a low frame rate of 25 fps while maintaining high audio quality, effectively halving the frame rate of many baseline models. It uses unified 24 kHz training that mixes speech, music, and sound-event clips in one model to capture both semantic and acoustic details, facilitating the training of audio language models. The model enables fast inference by avoiding diffusion steps, with an encoder/decoder architecture that processes batches quickly for real-time or large-scale tasks.
Links: Documentation
Breaking changes
Tensor parallelism (TP) support for dense and MoE decoder-only models has been fixed and stabilized, requiring users to update their TP configurations and conversion mappings accordingly.
The
Ernie4.5 VL MoEmodel class and configuration names have been renamed to align with vLLM/SGLang conventions, requiring users to update any references to the old model names in their code.Ernie 4.5 VL Moe] Fix up namings to vllm/sglang convention (#44299) by @vasquSeveral pipeline tasks have been removed or updated in the V5 cleanup (including
question-answering,visual-question-answering, andimage-to-image), requiring users to migrate to the replacement pipelines or updated task names.3D position IDs for vision-language models have been unified under a common interface (sourced from
qwen2-vl), requiring users of affected VLMs (e.g., Ernie, GLM4V) to update their processors and any code that manually constructs position IDs.🚨 Tokenizer x vLLM fixes 🚨 :
Unigram tokenizers were missing the
spmprecompiled charsmap support. We ran an overall v4 vs v5 regression test and fixed what we had missed.This was done in:
Generation
Generation input preparation was significantly refactored to stop relying on
cache_positionand instead pass pre-slicedinput_ids/inputs_embedsdirectly toprepare_inputs_for_generation, simplifying the generation loop and laying groundwork for broadercache_positionremoval. Several bug fixes were also applied, including correct sampling for HiggsAudioV2, flaky cache-equality test stabilization for Idefics, and restored generation integration tests.prepare_inputs_for_generation(#44226) by @Cyrilvallez in [#44226]cache_positionto prepare inputs (#44130) by @Cyrilvallez in [#44130]Tokenization
Several tokenization bugs were fixed in this release, including resolving an
AttributeErrorinMLukeTokenizercaused by the v5 rename ofadditional_special_tokens, correcting the Fuyu tokenizer class mapping, fixingLayoutXLMtokenization test failures from the slow tokenizer removal refactor, and addingolmo_hybridto the auto-tokenizer mapping. The tokenizer documentation was also updated to reflect the new unified v5 backend architecture and reorganized for clarity.Kernels
Fixed several kernel-related issues including a security vulnerability, corrected Mamba kernel loading to handle incompatible import structures, ensured Liger Kernel is properly enabled during hyperparameter search, and expanded Flash Attention to support multiple compatible implementations.
Mamba] Fix kernel loading (#44176) by @vasqu in [#44176]Flash Attn] Enable compatible implementations (#44177) by @vasqu in [#44177]Quantization
This release adds several new quantization backends and fixes, including MLX quantization support for MPS devices, Four Over Six (4/6) NVFP4 quantization integration for NVIDIA Blackwell GPUs, and CPU support for MXFP4 models, alongside a bug fix for MXFP4 model saving using
reverse_op.Vision
Fixed backward compatibility for image processors loaded from older remote code that lack
valid_kwargsdefinitions, and resolved test failures in AMD ROCm CI by adding the missingtimmdependency to the Docker image.from_dictbackward compatibility with old remote code (#44245) by @yonigozlan in [#44245]Bugfixes and improvements
speaking_rateas an optionl forward argument (#43283) by @gau-nernst in [#43283]ProcessingKwargsImagesKwargsetc. to docs (#44269) by @yonigozlan in [#44269]has_similar_generate_outputsassertions (#44166) by @tarekziade in [#44166]TokenizersBackendfor Olmo3 to preserve custompre_tokenizer(#44294) by @mario-sanz in [#44294]Modular] Fix file type regression (#44283) by @vasqu in [#44283]Trainerclass docs (compute_loss&hyperparameter_search) (#44268) by @ethanknights in [#44268]fix] Set input_modalities on various architectures that aren't just text (#44078) by @tomaarsen in [#44078]VersionComparison.from_stringreturn type mismatch (#43709) by @tarekziade in [#43709]AnyToAnyPipeline.__call__docstring (#44229) by @alvarobartt in [#44229]test_generate_with_and_without_position_idsin GLM ORC (#44173) by @tarekziade in [#44173]Seq2SeqTrainingArgumentsdocumentation (#35258) by @qgallouedec in [#35258]__setitem__onModelOutputeven if the parameter was previouslyNone(#44080) by @tomaarsen in [#44080]simple] Fix up__repr__whitespace/brackets (#44048) by @tomaarsen in [#44048]chore] Fix incorrect forward type hint for Gemma3n (#44051) by @tomaarsen in [#44051]get_audio_features(#44040) by @zucchini-nlp in [#44040]Kosmos2ModelTesttest (#44061) by @tarekziade in [#44061]grouped_mmfallback (#44043) by @IlyasMoutawwakil in [#44043]Configuration
📅 Schedule: Branch creation - At any time (no schedule defined), Automerge - At any time (no schedule defined).
🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.
♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.
👻 Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.
This PR has been generated by Renovate Bot.