Repository navigation
chore(deps): update loader dependencies non-major - #309
Open
dreadnode-renovate-bot[bot] wants to merge 1 commit into
Open
dreadnode-renovate-bot[bot] wants to merge 1 commit into
dreadnode-renovate-bot[bot] wants to merge 1 commit into
Conversation
| datasource | package | from | to | | ---------- | ------------ | ------ | ------ | | pypi | peft | 0.21.2 | 0.21.3 | | pypi | transformers | 5.22.2 | 5.23.0 |
This branch has not been deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Note
This PR body was truncated due to platform limits.
This PR contains the following updates:
| Package | Change | Age | Confidence |
|
Generated Summary:
transformerslibrary from version 5.22.2 to 5.23.0 in theautomodelloader requirements.peftlibrary from version 0.21.2 to 0.21.3 in theloraloader requirements.This summary was generated with ❤️ by rigging
| peft |
|
|
|
|
==0.21.2→==0.21.3|| transformers |
==5.22.2→==5.23.0|Release Notes
huggingface/peft (peft)
v0.21.3Compare Source
Release 0.21.3 containing:
This is mostly to fix a performance penalty with huggingface/transformers#49316 as this makes Trainer call
save_pretrainedon every rank, and with peft 0.21.2 that gathers the whole sharded model instead of just the adapter.What's Changed
Full Changelog: huggingface/peft@v0.21.2...v0.21.3
huggingface/transformers (transformers)
v5.23.0Compare Source
Release v5.23.0
New Model additions
Inkling2
Inkling 2 Small is a multimodal model from Thinking Machines Lab that accepts text and image inputs and generates text outputs, and it is intended for use in English and other languages and across multiple coding languages. Its language backbone has 48 mixture-of-experts decoder layers, each routing tokens to 8 of 384 experts, for 594B total and 17B active parameters. It keeps the Inkling decoder (learned relative attention, short causal convolutions, mixture-of-experts layers) and adds block attention residuals, gated attention with a learned sink, and dynamic key/value convolutions. Images are encoded by a Perception Encoder-based vision transformer and projected into the language model embeddings, one embedding per 40x40 pixel patch.
Links: Documentation | Blog Post
JEV
JEV is a decision head that scores single-token labels with a linear layer at the last token of the prompt, soft-capped like the logits of a language model head.
JevForDecisionis the model of the checkpoint and its head, and it reads the head files shipped next to the weights (head.safetensorsandjudge_config.jsonfor JEV). The same head is also used by GEV and reads the 255 answer codes of pplx-decider and d3, and it can be run through the experimentaldecisionpipeline or directly withdecide.Note: JEV and the decision pipeline are currently experimental. Their APIs, configuration, and behavior may change as development continues. They should be used with caution in production environments.
Links: Documentation
LFM3
LFM3 is the third generation of Liquid Foundation Models from Liquid AI. Like LFM2, it is a hybrid architecture that combines gated short convolutions with grouped-query attention, with the attention layers split between full-attention and sliding-window layers. The attention layers apply RMSNorm to the query and key heads and gate their output with a per-head sigmoid. Every layer applies a sandwich norm to the operator and feed-forward outputs before the residual addition.
Links: Documentation
Dots 3 Note Preview
Dots 3 Note Preview is a mixture-of-experts causal language model that accepts native text, image, video, and audio inputs. It uses a shared vision encoder for images and videos and a Whisper-style audio encoder, and both encoders project their outputs into the language model's hidden space before autoregressive text generation. BF16 and FP8 inference are supported.
Links: Documentation
LFM3-VL
LFM3-VL is a vision-language model family from Liquid AI, built on the LFM3 language model. It takes images and videos at their native resolution together with text. It combines a hybrid vision encoder (a linear patchifier, gated depthwise 2D convolution layers, and bidirectional grouped-query attention layers with 2D rotary embeddings) with a projector and an LFM3 language model. The projector merges each 2x2 block of patch features into one language token through a layer norm and a two-layer MLP.
Links: Documentation
IQuestCoder
IQuest-Coder-V1 is a family of code language models developed by IQuest Lab, available in 7B, 14B and 40B sizes with base, instruct and thinking variants. Its architecture is the same as Llama, so the Transformers auto classes load it with LlamaModel and LlamaForCausalLM, and only IQuestCoderConfig is specific to IQuestCoder. The looped
IQuest-Coder-V1-*-Loop-*checkpoints use a different architecture and aren't supported.Links: Documentation
InternLM3
InternLM3 is an 8B instruction-tuned language model developed by Shanghai AI Laboratory (checkpoint: internlm/internlm3-8b-instruct). Its architecture is the same as Llama, so the auto classes load it with the Llama implementation (LlamaModel and LlamaForCausalLM), and only InternLM3Config is specific to InternLM3. The config names the attention biases
qkv_biasand the output projection and MLP biasesbias. InternLM3Config loads these asattention_biasandmlp_biasand requires them to be equal, as they are in the released checkpoints.Links: Documentation
Breaking changes
ep_routerexpert-parallel style (#49615) by @IlyasMoutawwakilep_dispatch_experts(#49613) by @IlyasMoutawwakildeepgemm_megamoeEP implementation for DeepGEMM Mega MoE (#49606) by @IlyasMoutawwakil[audio]to torchaudio + soundfile + soxr (drop librosa and phonemizer) (#50610) by @eustlbExportedModel, Hub save/load, one backend registry (#50029) by @IlyasMoutawwakilkernels] Bump version (#50307) by @vasquExporters: export, share and run any model on its backend (#50029, #50032, #50309, #50311)
exporters.mp4
Every export now returns an
ExportedModel. You call it like the original model (same kwargs, same output class), push it to the Hub, and load it back on the backend's own runtime. The same API covers Dynamo, ONNX Runtime, OpenVINO and ExecuTorch.OnnxConfig(ort_fusions="cpu" | "cuda" | "webgpu" | "dml")rewrites the graph into ONNX Runtime's fused transformer ops at export time: attention,GroupQueryAttentionover the KV cache, rotary, norms andMoE. OpenVINO exports now hit OpenVINO's fused CPU kernels for attention, RoPE, MoE experts and compressed int8/int4 weights, which makes decode 2.4–3.9× faster than before.torch.compile: on EmbeddingGemma2 (8 × 128 tokens), ONNX Runtime with the TensorRT EP is 7.6× faster than eager PyTorch and 1.4× faster thantorch.compileon an A100. OpenVINO is 4.1× faster than eager and 2.2× faster thantorch.compileon a Xeon Platinum 8570 (4.9× and 2.6× with int8 weights).subfolderof the model repo. Itsexport.jsonrecords the config, input shapes, package versions and I/O names. Quantized exports (PT2E, NNCF, ONNX Runtime) save and run the same way.🚨
export()returns anExportedModel, and an ONNX program is anonnx_ir.Model.OnnxConfig's andOpenVINOConfig'soutput_pathoptions are gone in favour ofsave_pretrained. Custom backends register withregister_backend.📚 Docs: https://huggingface.co/docs/transformers/main/en/exporters
Vision
This release fixes a range of vision-model bugs: classification pipelines now respect
problem_type, and several models get corrected behavior (Step3.7 RoPE axis order, MiniMax-M3-VL packed attention isolation, Qwen VL positions with reused caches, PaddleOCR-VL interpolation alignment, Kosmos2 SDPA masking, and Emu3 float image handling). Chunked prefill now keeps image inputs on theParallelization
This release improves distributed training and inference. Expert parallelism gains a selectable
ep_implementation(all-reduce for CUDA-graph-friendly inference, all-to-all for training). The Trainer now reduces FSDP/HSDP gradients only on the last micro-batch, and the token-count normalization is fixed to use true division. Tensor-parallel fixes cover attention sinks (including Beam attention) and tied-weight comparisons on TPUep_dispatch_experts: all-reduce at inference, all-to-all in training (#49429) by @IlyasMoutawwakil in [#49429]distributed] Trainer: reduce FSDP/HSDP gradients on the last micro-batch only (#50286) by @3outeille in [#50286]Beam] Fix TP on attn (#50302) by @vasqu in [#50302]Cache
This release fixes several cache correctness issues: static-cache prefill attention in Gemma4's KV-shared layers, Qwen VL positions when a cache is reused with a new image or video, and DeepSeek-V4 compression-state reordering during beam search. It also improves cache sharing for Mistral Large 4 and adds OpenVINO export fusion for the KV cache and attention. Smaller changes cover CI and test fixes, including vLLM CI memory
Kernels
This release fixes several hub kernel issues: the MPS RMSNorm kernel repo id was corrected, the MegaBlocksMoeMLP mapping was restored for MXFP4 GPT-OSS, and a failed flash-attention kernel load is now retried instead of leaving models unregistered. It also raises the minimum
kernelsversion (fixing ROCm issues and a multi-GPU bug), bumps mamba-ssm kernels to v4GptOss] Fix the kernels autodispatch test (#50630) by @vasqu in [#50630]kernels] Bump min version (#50509) by @vasqu in [#50509]lazy_import_flash_attention(#50391) by @albertvillanova in [#50391]MegaBlocksMoeMLPhub kernel mapping for MXFP4 GPT-OSS (#50150) by @jiqing-feng in [#50150]Attention
This release fixes a range of attention bugs: vision-attention isolation in MiniMax-M3-VL, SDPA decode under
torch.compile, eager NaNs in Qwen4-Exp with left padding, an unintended causal mask in Kosmos-2's SDPA projection, and default full-attention mask selection that wrongly picked linear layers. It also retries failed attention kernel loads, excludes padding from the repetition penalty, and lets the OSDPA] Fix decode under compile (#50388) by @vasqu in [#50388]Generation
This release fixes several generation bugs: padded generation in RecurrentGemma, batched decoding in Falcon-H1, beam search cache reordering in DeepSeek-V4, and batched metadata and beam scores in
TextGenerationPipeline. It also lets an unchangedGenerationConfigbe saved without strict-validation errors, avoids copying the KV cache when decoding with grouped-query attention (GQA), and fixes an `Attributlayer_types(#50577) by @ydshieh in [#50577]Quantization
This release fixes several quantization bugs: FP8
modules_to_not_convertentries are now renamed like weight names, so modules such asf_a_projandf_b_projare no longer wrongly quantized. MXFP4 experts now dequantize to the requested dtype instead of always using bfloat16. TheMegaBlocksMoeMLPhub kernel mapping is restored for MXFP4 GPT-OSS, which fixmodules_to_not_convertentries like weight names (#50354) by @albertvillanova in [#50354]Audio
This release makes audio handling more correct and robust.
AudioClassificationPipelinenow respects the model's problem type (raw outputs for regression, sigmoid for multi-label, softmax otherwise), and VibeVoice and VibeVoice-ASR avoid GPU syncs when removing audio padding and gain asampleargument to disable VAE noise for deterministic outputs. Chat templates can also take audio decoded from video via `{"type": "audsampleargument to disable VAE noise in audio features (#50179) by @hmellor in [#50179]Bugfixes and improvements
Gemma4] Fix FA slicing (#50681) by @vasqu in [#50681]ep_dispatch_experts(#49607) by @IlyasMoutawwakil in [#49607]make run-miceasier to read and to target (#50652) by @tarekziade in [#50652]HfArgumentParseron Python 3.15 (#50651) by @qgallouedec in [#50651]Offloading] Fix tests for some models (#50631) by @vasqu in [#50631]FA] Fix FA tests for unique models (#50632) by @vasqu in [#50632]ModelRunner) (#50032) by @IlyasMoutawwakil in [#50032]OnnxConfig(ort_fusions=...)) (#50309) by @IlyasMoutawwakil in [#50309]distributed] Checkpoint the layer'sforward, not__call__, so FSDP2 hooks stay outside the checkpoint (#50490) by @3outeille in [#50490]transcriptioninput arg toapply_transcription_request(#49869)" (#50587) by @ebezzam in [#50587]ForPreTrainingmodels (argmax near-tie) (#50575) by @ydshieh in [#50575]grounding_dinofailing withoutput_mismatch(tensor values differ (6), oth (#49886) by @sergereview[bot] in [#49886]test_causal_lm_loss_honors_shift_labels(#50533) by @ydshieh in [#50533]per_layer_config(#50486) by @zucchini-nlp in [#50486]apply_chat_template(#45422) by @qgallouedec in [#45422]test_eager_matches_batched_and_grouped_inferencein fp16 / bf16 (#50500) by @ydshieh in [#50500]Configuration
📅 Schedule: (UTC)
🚦 Automerge: Enabled.
♻ 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 Mend Renovate CLI.