Release v5.17.0

New Model additions

HYV4

Hy4-Preview is a 780B-parameter mixture-of-experts language model that activates 49B parameters per token. Each MoE layer holds 256 routed experts plus one always-active shared expert and routes every token to 8 of them. The context window is 1M tokens.

The architecture combines four features:

Multi-head Latent Attention (MLA) compresses keys and values into a low-rank latent (kv_lora_rank) that kv_b_proj expands back to one key/value per query head.

DeepSeek Sparse Attention (DSA) selects index_topk keys per query with a lightweight indexer.

Following IndexShare, only the layers marked "full" in indexer_types run an indexer; "shared" layers reuse the previous full layer's selection.

Gated MLA with learnable attention sinks, where each head owns a sink logit that participates in the softmax and contributes no value, as in GPT-OSS.

Independent Hyper-Connections (iHC) replace the plain residual path with hc_mult parallel residual streams that are collapsed before, and redistributed after, every sublayer.

The implementation does not execute the multi-token prediction (MTP) layers. Released checkpoints keep those weights so that other runtimes can use them for speculative decoding; they are ignored at load time.

Links: Documentation

Add h4 (#48473) by @ArthurZucker in #48473

VibeVoice

VibeVoice is a novel framework for synthesizing high-fidelity, long-form speech with multiple speakers by employing a next-token diffusion approach within a Large Language Model (LLM) structure. It's designed to capture the authentic conversational "vibe" and is particularly suited for generating audio content like podcasts and multi-participant audiobooks.

Links: Documentation

Implement VibeVoice (#40546) by @pengzhiliang in #40546

NeoMME

NeoMME is a family of efficient 260M and 800M parameter multimodal-native multilingual foundation encoders from H Company. It processes multilingual text tokens and raw image patches in a single bidirectional Transformer encoder, without a separately pretrained vision tower or causal language model.

NeoMME-Retriever is a model fine-tuned from the NeoMME backbone for visual document retrieval with joint late-interaction and dense objectives. It takes text queries and documents (text or page screenshots) and produces multi-vector embeddings for MeanMaxSim scoring (late-interaction) and mean-pooled embeddings for cosine similarity (dense).

Links: Documentation

Add NeoMME and NeoMME-Retriever (#47992) by @tonywu71 in #47992

Fun-ASR-Nano

Fun-ASR-Nano is an 800M-parameter end-to-end speech recognition model developed by Alibaba DAMO Academy's FunAudioLLM team. It achieves state-of-the-art performance on Chinese, English, and Japanese ASR benchmarks while being significantly smaller than comparable models.

Key features are

Chinese, English, and Japanese, including 7 Chinese dialects and 26 regional accents

Hotword customization for domain-specific vocabulary

Native punctuation output (no separate punctuation model needed)

Links: Documentation

Add Fun-ASR-Nano model (#46180) by @LauraGPT in #46180

KimiLinear

Kimi Linear is a hybrid linear attention architecture from Moonshot AI, introduced in Kimi Linear: An Expressive, Efficient Attention Architecture.

At its core is Kimi Delta Attention (KDA), a refinement of Gated DeltaNet that gives each key channel its own forget gate, so the recurrent state decays per channel instead of per head. KDA is used in most layers; every fourth layer keeps a full-attention block that reuses DeepSeek-V3's Multi-head Latent Attention (MLA), and the feed-forward blocks are DeepSeek-V3-style MoE with a shared expert.

Links: Documentation

Kimi linear (#48250) by @remi-or in #48250

Canary

Canary-1B-v2, a fast, robust multilingual model for Automatic Speech Recognition (ASR) and Speech-to-Text Translation (AST):

Canary reuses the Fast Conformer encoder from Parakeet (loaded through [ParakeetEncoder] / [ParakeetEncoderConfig]) and pairs it with a Transformer decoder that uses fixed sinusoidal positional embeddings, cross-attention to the encoder outputs and tied input/output embeddings. The task is selected through a decoder prompt prefix built by [CanaryProcessor] of the form <|startofcontext|> <|startoftranscript|> <|emo:undefined|> <source_lang> <target_lang> <pnc|nopnc> <|noitn|> <|notimestamp|> <|nodiarize|>, where source_lang == target_lang selects transcription and otherwise selects translation.

Links: Documentation

model: Add NVIDIA Canary-1B-v2 to Transformers (#46825) by @harshaljanjani in #46825

NeuCodec

The NeuCodec model was proposed in Finite Scalar Quantization Enables Redundant and Transmission-Robust Neural Audio Compression at Low Bit-rates.

NeuCodec is a neural audio codec extending on XCodec2. It takes advantage of the following features:

Finite Scalar Quantization (FSQ) quantisation resulting in a single codebook, making it ideal for downstream modeling with Speech Language Models.

Trained with CC data such that there are no Non-Commercial data restrictions.

At 50 tokens/sec and 16 bits per token, the overall bit-rate is 0.8kbps.

The codec takes in 16kHz input and outputs 24kHz using an upsampling decoder.

The FSQ encoding scheme allows for bit-level error resistance suitable for unreliable and noisy channels.

Links: Documentation

Add support for NeuCodec (#47143) by @harryjulian in #47143

Breaking changes

Vision rotary embeddings (2D/3D) have been standardized into a unified RoPE frequency computation module, so users with custom vision models relying on attention-layer-level or model-specific RoPE grid interleaving logic must migrate to the new centralized modeling_rope_utils.py implementation.

🚨 Vision (2d/3d) rotary embeddings (#48105) by @zucchini-nlp

Generation

Generation improvements include a performance optimization that avoids unnecessary accelerator synchronization on every decode step (reducing per-step overhead), and a fix to prevent unconditional downloading of remote hub files during generation. Several correctness fixes were also applied, including enforcing auto-compile cache checks for encoder-decoder models, standardizing past_key_values naming in AfMoE, and resolving flaky export and integration test failures.

[Generate] Avoid unconditionally downloading remote hub file (#48620) by @vasqu in [#48620]

[generate] stop synchronizing the accelerator on every decode step (#47975) by @SunMarc in [#47975]

[AfMoE] Standardize past_key_values argument naming across forward and generate (#48430) by @shenhuaqingshi in [#48430]

Fix MTP generation test regex gate for escaped layer ignore keys (#48003) (#48262) by @Noxtimo in [#48262]

fix(generation): Enforce the auto-compile cache check for encoder-decoder models (#48364) by @harshaljanjani in [#48364]

[serge] Fix 4 integration tests for model generation failing with output_mismatch (list output differs (4)) (#48133) by @sergereview[bot] in [#48133]

[VibeVoice] Skip generate export tests (flaky) (#48396) by @ydshieh in [#48396]

Cache

Fixed several cache-related bugs, including a quantized cache issue in VibeVoice, incorrect rejection of non-static cache implementations in VoxtralRealtime, missing auto-compile cache checks for encoder-decoder models, and a silent failure when paged attention is called without a cache. Documentation was also updated to clarify ContinuousBatchingConfig usage and sliding window model limitations.

vibevoice: fix bug for quant cache (#48487) by @kaixuanliu in [#48487]

Fix VoxtralRealtime rejecting non-static cache implementations (#48082) by @jiqing-feng in [#48082]

Raise when a paged attention forward is called with no cache (#48297) by @qgallouedec in [#48297]

[docs] Pass ContinuousBatchingConfig and sliding window models (#48381) by @stevhliu in [#48381]

Retry get_daily_ci_runs on stale GitHub API cache (#48374) by @ydshieh in [#48374]

Kernels

Kernel support was improved with fixes for nested FLA kernel imports when only fla-core is installed, a warning when hub-kernel functions silently fall back to slower pure-PyTorch reference implementations, and the ability to register standalone functions (e.g., RoPE) in KernelConfig with optional non-inheritance of default mappings. Additional fixes include corrected repository paths for ESMFold2 kernels and updated documentation for KernelConfig customization.

Support nested FLA kernel imports for fla-core (#48221) by @DimensionSTP in [#48221]

Warn once when a hub-kernel function falls back to its reference PyTorch path (#48185) by @qgallouedec in [#48185]

[docs] Kernel updates (#48465) by @stevhliu in [#48465]

[Kernels] Enable functions into kernels registry and allow non inheritance (#48443) by @vasqu in [#48443]

Fix kernel commit and repo paths for ESMFold2 (#48186) by @Rocketknight1 in [#48186]

Quantization

Fixed several quantization bugs, including a quant cache issue in VibeVoice, incorrect FP8 embedding handling for Qwen models, missing FP8 tensor parallelism layer overrides, and unnecessary MXFP4 weight dequantization on XPU devices.

fix qwen4exp-fp8 ple embedding (#48368) by @JJJYmmm in [#48368]

Keep MXFP4 weights quantized on XPU when use_kernels is set (#47923) by @jiqing-feng in [#47923]

Fix missing FP8 TP layer overrides (#48343) by @changwangss in [#48343]

Bugfixes and improvements

MRoPE continued (#48594) by @zucchini-nlp in [#48594]

[fix] Update stale expected strings in HunYuanVL integration tests (#48646) by @ydshieh in [#48646]

[Quantizaiton]support 5/6/7 bits in AutoRound (#48481) by @wenhuach21 in [#48481]

[fix] Update stale golden values and fix expected_logits shape in FlavaForPreTraining integration tests (#48639) by @ydshieh in [#48639]

Fix YOLOS device mismatch with device_map="auto" (#46886) by @swankystark in [#46886]

Honor shift_labels in decoder-only LLM/VLM losses (#48493) by @qgallouedec in [#48493]

[KimiLinear] Fix test_cpu_offload: set num_local_experts=4 in model tester (#48624) by @ydshieh in [#48624]

[docs] Per-layer config (#48601) by @stevhliu in [#48601]

Fix generate_flags parsing in transformers chat (#48597) by @SunMarc in [#48597]

[fix] Fix how we read package versions - triggered by torch 2.14+ (#48615) by @ydshieh in [#48615]

[Docker] Upgrade CPU torch to <=2.14.0, torchcodec to <=0.16.0 (#48614) by @ydshieh in [#48614]

Fix AttributeError in gradient_checkpointing_enable(offload=True) (#48590) by @tarekziade in [#48590]

Another day fixing CI (#48591) by @zucchini-nlp in [#48591]

esmfold2: keep distogram_head in fp32 as well (#48488) by @kaixuanliu in [#48488]

Add supports_context_parallel to PreTrainedModel (#48442) by @qgallouedec in [#48442]

[docs] mlinter reference (#48460) by @stevhliu in [#48460]

[docs] Add a LiteRT page under community integrations (#48540) by @john-rocky in [#48540]

[GLM 5.3 Flash] Fix NaN gradients in chunked KDA (#48455) by @imvladikon in [#48455]

[docs] Fix [[autodoc]] directives in ALBERT model documentation (#48593) by @samyuktahegde in [#48593]

[Fix] Fix A10 expectations for a test (#48454) by @remi-or in [#48454]

Add PR comment CI for AMD (MI300) (#48065) by @ydshieh in [#48065]

docs: fix docstring parameter names that do not match signatures (#48575) by @simpleqt in [#48575]

docs: remove phantom parameters from docstrings (#48576) by @simpleqt in [#48576]

extend some case to xpu as well (#48502) by @sywangyi in [#48502]

[serge] Fix 2 integration tests for model glm4_moe failing with OOM (other (2)) (#48551) by @sergereview[bot] in [#48551]

[serge] Fix 2 integration tests for model nemotron failing with import_or_config (other (2)) (#48582) by @sergereview[bot] in [#48582]

Compress the agent conventions file and document two modular pitfalls (#48586) by @tarekziade in [#48586]

Guard against a None video processor class when the backend is unavailable (#48557) by @caiotheodoro in [#48557]

[serge] Fix 2 integration tests regressed by commit 83d46aa (PR #47625) (#48580) by @sergereview[bot] in [#48580]

[nit] use requires_backends (#47576) by @eustlb in [#47576]

[serge] Fix 2 integration tests for model kosmos2 failing with import_or_config (other (2)) (#48552) by @sergereview[bot] in [#48552]

Fix failing tests for cohere_compass (#48005) by @kaixuanliu in [#48005]

[Fix] Use dedicated helpers for DeepGEMM and SonicMoE tests (#48523) by @remi-or in [#48523]

CI: Point test fixtures at hf-internal-testing copies we already host (#48521) by @tarekziade in [#48521]

fix failed test cases for glm5_next (#48497) by @kaixuanliu in [#48497]

Pass kwargs to the Mamba2 mixer in Nemotron-H, Falcon-H1 and Mamba2 (#48490) by @kfastino in [#48490]

Retire test_multi_gpu_data_parallel_forward (#48508) by @tarekziade in [#48508]

Processing tests [part 2] (#47922) by @zucchini-nlp in [#47922]

QA: Add noisy comment checker (#48484) by @tarekziade in [#48484]

Allow nested rope params for tiny models (#48435) by @zucchini-nlp in [#48435]

Fix sliding-window mask layer_idx in Gemma3/Gemma4 create_masks_for_vision_model (#48482) by @jiqing-feng in [#48482]

add xpu expectations for hunyuan_vl model tests (#48504) by @kaixuanliu in [#48504]

[Qwen 3.5 Moe] Fix decorators (#48436) by @vasqu in [#48436]

Infinite loop in dependency search (#48393) by @zucchini-nlp in [#48393]

[serge] Fix 2 integration tests for model fsmt failing with output_mismatch (tensor values differ (2)) (#48496) by @sergereview[bot] in [#48496]

Fix some tests by removing the deprecation cycle (#48503) by @Cyrilvallez in [#48503]

Remove deprecation (#48500) by @Cyrilvallez in [#48500]

Fix Pix2StructTextAttention init using hidden_size instead of d_kv (#47558) by @ in [#47558]

Fix pre patch release utility (#48499) by @Cyrilvallez in [#48499]

Update dev version (#48498) by @Cyrilvallez in [#48498]

Fix Inkling inputs_embeds and add more tests (#47827) by @Cyrilvallez in [#47827]

Simplify and fix qwen4 tests (#48340) by @Cyrilvallez in [#48340]

[docs] Partial checkpointing and group_by_length (#48463) by @stevhliu in [#48463]

doc: fix syntax error and typos in VibeVoice documentation (#48489) by @VimalN2005 in [#48489]

[Qwen4 Exp] Use partial to avoid skipping mask more easily (#48456) by @vasqu in [#48456]

Add support for NeuCodec (#47143) by @harryjulian in [#47143]

[serge] Fix 1 integration tests regressed by commit bd95093 (PR #47493) (#48426) by @sergereview[bot] in [#48426]

Fix rotary embedding regression (#48477) by @Cyrilvallez in [#48477]

Remove deprecated mask functions (#48476) by @Cyrilvallez in [#48476]

[MTP] Save memory by only capturing the last layer's hidden_states (#48475) by @Cyrilvallez in [#48475]

Allow capturing only necessary hidden_states with capture_outputs (#48081) by @sywangyi in [#48081]

Support per-layer MTP configuration (#48264) by @eladsegal in [#48264]

[docs] Fix code snippets (#47772) by @stevhliu in [#47772]

[Fix] Sparse TikToken tokenizers silently fail (#48446) by @remi-or in [#48446]

No inherit decorator for NeoMME (#48457) by @zucchini-nlp in [#48457]

Batch Rebalance Data Sampler (#47340) by @delock in [#47340]

[serge] Fix 2 integration tests for model cwm failing with import_or_config (other (2)) (#48414) by @sergereview[bot] in [#48414]

fix: Add DEIMv2 attribution (#48448) by @harshaljanjani in [#48448]

[fix] inkling: mps + cuda mel spec extraction (#47432) by @eustlb in [#47432]

fix: decode() batch path respects self.clean_up_tokenization_spaces (#47793) by @lorenzozanee in [#47793]

Grounding dino fp16 dtype [backlog] (#48438) by @molbap in [#48438]

Init the process group with a load-scaled timeout for sharded loading (#48228) by @qgallouedec in [#48228]

Raise a clear error when a token is both forced and suppressed (#47511) by @qgallouedec in [#47511]

Clarify device placement in pipelines (#47367) by @LysandreJik in [#47367]

Add offload to gradient checkpointing (#48444) by @qgallouedec in [#48444]

[MiniCPMV4_6] Update test_small_model_vision_generation_batch expected output (value drift) (#48406) by @ydshieh in [#48406]

[serge] Fix 2 integration tests for model hyperclovax failing with other (other (2)) (#48440) by @sergereview[bot] in [#48440]

fix(models): Drop the position-indexed token type lookup in RoPE encoders (#48407) by @harshaljanjani in [#48407]

Document image_hidden_states/pixel_values mutual exclusivity for SmolVLM/Idefics2/Idefics3 (#47714) by @verma8076 in [#47714]

Re-order a bit for easier navigation (#48434) by @zucchini-nlp in [#48434]

Deprecated stuff gone (#48367) by @zucchini-nlp in [#48367]

[serge] Fix 6 integration tests for model seamless_m4t_v2 failing with other (other (6)) (#48425) by @sergereview[bot] in [#48425]

[Docs]: Update GLM 5.3 (#48401) by @Dovis01 in [#48401]

Avoid print to stdout that fails the job check_failed_tests job (#48391) by @ydshieh in [#48391]

Fix incorrect tuple return annotations on forward methods returning a Tensor (#48359) by @Gronoxx in [#48359]

Fix interval merge invariant in _find_disjoint (#47860) by @sharmax-vikas in [#47860]

skip mtp slow tests for now (#48328) (#48329) by @tarekziade in [#48329]

Update Tailscale action version in workflow (#48394) by @glegendre01 in [#48394]

fix some failure in xpu (#48252) by @sywangyi in [#48252]

[Improvement] Make gated delta rule more explicit (#47625) by @remi-or in [#47625]

[CB] Fix wrong device scoping (#48370) by @remi-or in [#48370]

Bump transformers-mlinter to 0.1.5 and clear the new findings (#48259) by @tarekziade in [#48259]

fix: flash-attn fallback failing on torch2.13 (#48388) by @NanoCode012 in [#48388]

[LongcatFlash] Fix test_longcat_generation_cpu: use device_map="cpu" to avoid MoE disk offload issue (#48377) by @ydshieh in [#48377]

[Qwen3VLMoe] Update test_small_model_integration_test_batch expected output (value drift) (#48376) by @ydshieh in [#48376]

[ONNX] Skip affected models on torch 2.13 (two dynamo regressions) (#48191) by @ydshieh in [#48191]

Fix safe_open mmap memory exhaustion on Windows by using pread backend (#48341) by @eryk-roch in [#48341]

Fix Zamba2 construction for num_mem_blocks > 1 checkpoints (#48325) by @john-rocky in [#48325]

[Docs] Change 5.3 Flash pos in toc (#48366) by @Dovis01 in [#48366]

[conftest] Use get_cpu_ram_total_gib for psutil patch (cgroup-aware) (#48290) by @ydshieh in [#48290]

Fix flaky test_training_gradient_checkpointing for BigBirdPegasus (fp noise filter) (#48332) by @ydshieh in [#48332]

Quiet continuous batching at default verbosity (#48314) by @qgallouedec in [#48314]

Wait for the first request in the async continuous batching bootstrap (#48304) by @qgallouedec in [#48304]

Ignore a stale best checkpoint recorded in a resumed trainer state (#48319) by @VaggelisGian in [#48319]

[docs] Fix links and remove TokenizerFast (#47748) by @stevhliu in [#47748]

Fix incorrect token classification prefix for ESMC (#48348) by @Rocketknight1 in [#48348]

Create the continuous batching CPU group with local synchronization (#48302) by @qgallouedec in [#48302]

Resolve continuous batching config against the text config for composite models (#48299) by @qgallouedec in [#48299]

[CI] Unblock fast CI for now (failing tests) (#48344) by @vasqu in [#48344]

[qwen4_exp] disable torch/onnx export tests due to data-dependent control flow (#48345) by @ydshieh in [#48345]

[debug] Trace previous CI run selection in get_previous_daily_ci.py (#48338) by @ydshieh in [#48338]

Normalize HunYuanVL's legacy field aliases via attribute_map (#48261) by @hmellor in [#48261]

Significant community contributions

The following contributors have made significant changes to the library over the last release:

@ydshieh

[fix] Update stale expected strings in HunYuanVL integration tests (#48646)

[fix] Update stale golden values and fix expected_logits shape in FlavaForPreTraining integration tests (#48639)

[tests] Fix integration test golden values broken by fast image processor default (PR #41388) (#48637)

[KimiLinear] Fix test_cpu_offload: set num_local_experts=4 in model tester (#48624)

Fix GPU memory teardown in CLI serve tests (#48618)

[fix] Fix how we read package versions - triggered by torch 2.14+ (#48615)

[Docker] Upgrade CPU torch to <=2.14.0, torchcodec to <=0.16.0 (#48614)

Add PR comment CI for AMD (MI300) (#48065)

[MiniCPMV4_6] Update test_small_model_vision_generation_batch expected output (value drift) (#48406)

Avoid print to stdout that fails the job check_failed_tests job (#48391)

[VibeVoice] Skip generate export tests (flaky) (#48396)

[LongcatFlash] Fix test_longcat_generation_cpu: use device_map="cpu" to avoid MoE disk offload issue (#48377)

[Qwen3VLMoe] Update test_small_model_integration_test_batch expected output (value drift) (#48376)

[ONNX] Skip affected models on torch 2.13 (two dynamo regressions) (#48191)

Retry get_daily_ci_runs on stale GitHub API cache (#48374)

[conftest] Use get_cpu_ram_total_gib for psutil patch (cgroup-aware) (#48290)

Fix flaky test_training_gradient_checkpointing for BigBirdPegasus (fp noise filter) (#48332)

[qwen4_exp] disable torch/onnx export tests due to data-dependent control flow (#48345)

[debug] Trace previous CI run selection in get_previous_daily_ci.py (#48338)

@LauraGPT

Add Fun-ASR-Nano model (#46180)

@tarekziade

Fix AttributeError in gradient_checkpointing_enable(offload=True) (#48590)

Compress the agent conventions file and document two modular pitfalls (#48586)

CI: Point test fixtures at hf-internal-testing copies we already host (#48521)

Retire test_multi_gpu_data_parallel_forward (#48508)

QA: Add noisy comment checker (#48484)

skip mtp slow tests for now (#48328) (#48329)

Bump transformers-mlinter to 0.1.5 and clear the new findings (#48259)

@remi-or

[Fix] Fix A10 expectations for a test (#48454)

Kimi linear (#48250)

[Fix] Use dedicated helpers for DeepGEMM and SonicMoE tests (#48523)

[Fix] Sparse TikToken tokenizers silently fail (#48446)

[Improvement] Make gated delta rule more explicit (#47625)

[CB] Fix wrong device scoping (#48370)

[CB] Fail faster (#48334)

@ArthurZucker

Add h4 (#48473)

@harryjulian

Add support for NeuCodec (#47143)

@delock

Batch Rebalance Data Sampler (#47340)

@harshaljanjani

fix: Add DEIMv2 attribution (#48448)

model: Add NVIDIA Canary-1B-v2 to Transformers (#46825)

fix(generation): Enforce the auto-compile cache check for encoder-decoder models (#48364)

fix(models): Drop the position-indexed token type lookup in RoPE encoders (#48407)

@tonywu71

Add NeoMME and NeoMME-Retriever (#47992)

@Dovis01

[Docs]: Update GLM 5.3 (#48401)

[Docs] Change 5.3 Flash pos in toc (#48366)

[Glm 5.3 Flash] GLM 5.3 Flash Support (#48342)

@pengzhiliang

Implement VibeVoice (#40546)

Original source

Release v5.16.1

This is a special release as we include GLM! (and a few small fixes)

GLM-5.3-Flash

GLM-5.3-Flash, the first natively multimodal model in the GLM-5 series. With 320B total parameters and just 18B active parameters, it outperforms GLM-5.2 across benchmarks and real-world workloads at one-tenth the price, while approaching Claude Opus 4.8 on coding and agentic benchmarks.

GLM-5.3-Flash starts from a newly trained base model, with its architecture and training recipe redesigned around capability and efficiency. For the first time in the GLM series, we introduce a hybrid architecture combining sparse and linear attention, sharply reducing long-context serving costs while preserving precise long-context capabilities. The model also adopts Manifold-Constrained Hyper-Connections (mHC) to further improve scaling efficiency. Together with our latest 30T-token multimodal pre-training corpus, these changes enable GLM-5.3-Flash to deliver more intelligence with less compute.

Links: Documentation

[Glm 5.3 Flash] GLM 5.3 Flash Support (#48342) by @Dovis01 in #48342

Small patch fixes

Mainly BC behavior for TP and pinning a hf kernel for security reasons 🤗

Restore BC for the tensor-parallel API (#48300) by @ArthurZucker

Fix kernel commit and repo paths for ESMFold2 (#48186) by @Rocketknight1

Full Changelog: v5.16.0...v5.16.1

Original source

Release v5.16.0

New Model additions

Qwen4-Exp

Qwen4-Exp builds on Qwen3.5's hybrid text and multimodal architecture with three key components: GatedResidual (GR), Qwen Sparse Attention (QSA), and Per-Layer Embedding (PLE).

GR is a Qwen-developed residual architecture that combines Hyper-Connection with GatedNorm. It mixes multiple residual streams with fine-grained elementwise gating before each attention and Mixture-of-Experts (MoE) block, then controls how much of the block output is injected back into each stream.

QSA uses multiple query heads to score compressed key blocks, selects the most relevant contiguous token blocks, and keeps the incomplete trailing block uncompressed. This block-level selection reduces indexing overhead and improves memory locality for long sequences. Combined with Gated DeltaNet, QSA makes Qwen4-Exp the first hybrid architecture to integrate linear and sparse attention, substantially improving inference efficiency for long-context workloads.

PLE enriches selected decoder layers with layer-specific lexical features derived from hashed token n-grams and a dilated depthwise convolution.

Links: Documentation

Add Qwen4Exp model (#48337) by @Cyrilvallez in #48337

GraniteSpeech5

Granite Speech 5.0 Turbo CTC is a lightweight (~470M parameters) conformer encoder for automatic speech recognition, trained with Connectionist Temporal Classification (CTC) on BPE targets. It is a fast, encoder-only member of the Granite Speech family: transcription requires a single forward pass followed by greedy CTC decoding, with no autoregressive decoder.

Architecturally, it extends the Granite Speech conformer CTC encoder with:

• Frame stacking + block-wise time subsampling: the feature extractor stacks pairs of log-mel(+delta) frames (2x), and the first two conformer blocks each subsample time by 2 through a stride-2 depthwise convolution (with a mean-pooled residual), for a total 8x time reduction at 10 ms mel hop.
• Block attention with Shaw's relative positional embeddings: attention is computed over fixed-size blocks (the sequence is right-padded to a whole number of blocks, with padded frames masked out), using separate bias-free query/key/value projections.
• Self-conditioned CTC: the CTC posteriors of the middle layer are projected and fed back into the hidden states, and the CTC head is shared between this mid-layer self-conditioning and the final prediction.

Links: Documentation

Add Granite Speech 5.0 - (#48288) by @eustlb in #48288

Step3p7

Step-3.7-Flash was proposed in Step 3.7 Flash by StepFun. It is a 198B-parameter sparse Mixture-of-Experts vision-language model, pairing a 196B-parameter MoE language backbone with a 1.8B-parameter vision encoder for native image understanding.

StepFun hasn't published a technical report for Step-3.7-Flash, so the details below are drawn from the released checkpoint's configuration rather than a paper.

• Sparse MoE decoder: all but the first 3 decoder layers route through a MoE block of 288 routed experts (top-8 per token) plus a single shared expert. The router scores experts with a sigmoid and a learned per-expert bias instead of an auxiliary load-balancing loss, the same strategy as DeepSeek-V3.
• Gated attention: each attention layer adds an extra projection whose sigmoid output gates the attention output per head, before the output projection — the same Gated Attention mechanism used in Qwen3-Next. A subset of layers use fewer heads and a sliding window instead of full attention.
• Multi-token prediction: some checkpoints ship extra decoder layers trained for multi-token prediction, which [~GenerationMixin.generate] can use for speculative decoding via use_mtp=True.
• Vision encoder: a SigLIP-style ViT with 2-D rotary position embeddings and a learned per-layer scale on the attention and MLP branches. Its output is downsampled 4x by two stride-2 convolutions before a linear projector maps it into the text model's hidden size.
• Dynamic image tiling: instead of a fixed tile grid, the image processor picks its tiling window from each image's own aspect ratio, producing one downscaled global view plus zero or more local high-resolution crops per image.

Links: Documentation

[new model] step 3.7 (#46658) by @itazap in #46658

CohereCompass

CohereCompass is the base architecture for small, specialized (vision-)language models trained by Cohere.

Links: Documentation

Add CohereCompass modeling (#47878) by @calpt in #47878

ESMC and ESMFold2

ESMC and ESMFold2 are new state-of-the-art protein language and folding models from BioHub. ESMC is trained with a masked language modeling objective, and it can be easily transferred to sequence and token classification tasks for proteins. Checkpoints exist in various sizes, from 300M parameters up to 6B parameters. It works as a drop-in replacement for older ESM-2 and ESM-3 models, with significantly higher accuracy.

ESMFold2 is a state-of-the-art protein folding model which produces high accuracy predictions. It uses an iterated diffusion approach that is significantly different from the original ESMFold, offering huge improvements in accuracy for more complex structures.

Links: Documentation ESMC, Documentation ESMFold2

Port ESMC and ESMFold2 to Transformers (#46419) by @Rocketknight1 in #46419

Breaking changes

The legacy tensor-parallel implementation has been replaced with a DTensor-native backend, so users relying on the previous TP API for inference or training must migrate to the new DTensor-based interface.

🚨 TP dtensor API inference + training (#47579) by @3outeille

attn_implementation="sdpa" dispatch is now properly supported for wav2vec2_conformer, wav2vec2-bert, and SeamlessM4T/v2 models, which may change initialization behavior for users who previously worked around this limitation.

🚨[wav2vec2] Support attn_implementation=sdpa dispatch (#46196) by @YangKai0616

FuyuProcessor no longer returns the image_patch_indices output, so any code that depends on this field must be updated to remove references to it.

🚨 Leftover processors (#47924) by @zucchini-nlp

Cache

Several cache-related bugs were fixed in this release, including an off-by-one error in the sliding window cache, Whisper speculative decoding cache corruption, CpmAnt use-cache failures, Qwen2.5-Omni/Qwen3-Omni-MoE generation with compilable caches, and compressed-tensors loading for KV-cache-only quantized models. Documentation was also added for cache token removal using negative values, and per-layer cache configuration support (allowing models to use different cache settings per layer) was introduced.

Cpmant fix use cache (#48013) by @jiqing-feng in [#48013]

[docs] Cache crop (#47950) by @stevhliu in [#47950]

Revert "Support per-layer cache configuration and attention-mask selection" (#48175) by @Cyrilvallez in [#48175]

Support per-layer cache configuration and attention-mask selection (#47901) by @eladsegal in [#47901]

Fix Qwen2.5-Omni / Qwen3-Omni-MoE generation with a compilable cache (#47872) by @jiqing-feng in [#47872]

Fix sliding window cache index off-by-one on wraparound (#47708) by @hameedibrh in [#47708]

[Whisper] Fix speculative decoding: UnboundLocalError, cache corruption, and speed regression (#48000) by @ydshieh in [#48000]

Fix compressed-tensors loading for KV-cache-only quantized models (#47904) by @kylesayrs in [#47904]

Generation

This release fixes several generation bugs across multiple models, including Whisper speculative decoding issues (UnboundLocalError, cache corruption, speed regression, and left-padded batch position IDs), broken image generation in Emu3, garbage output in OLMo/GPTNeoX, and Qwen2.5-Omni/Qwen3-Omni-MoE generation with compilable caches. Additionally, logit distributions for candidate generators using sampling are now aligned by returning logits after applying logit processors.

[Whisper] Fix speculative decoding: preserve cleared suppress tokens through super().generate() (#48108) by @ydshieh in [#48108]

[Whisper] Fix decoder position IDs for left-padded batches in longform generation (#48028) by @ydshieh in [#48028]

[serge] Fix 2 integration tests for model generation failing with import_or_config (other (2)) (#48061) by @sergereview[bot] in [#48061]

Align logit distributions for CandidateGenerators using sampling (#48007) by @Cyrilvallez in [#48007]

[GPTNeoX] Fix post_processor not overridden when loading from pretrained (OLMo garbage generation) (#47988) by @ydshieh in [#47988]

[emu3] 🦮 Black Labrador is back! Fix image generation broken since #37033 (#47948) by @ydshieh in [#47948]

Attention

Several attention-related bug fixes were made in this release, including correcting a SigLIP2 documentation typo, fixing Flash/SDPA attention dispatch tests for xcodec2 and ROCm RDNA GPUs, resolving a GPT2 cross-attention mask being silently discarded, and enabling SDPA support declaration in TimmWrapper. Per-layer cache configuration and attention-mask selection support was also introduced, allowing models with heterogeneous layer configurations to use distinct sliding_window, attention_chunk_size, and number_of_conv_states values per layer.

doc: Fix typo in SigLIP2 Flash Attention code example (#48197) by @VimalN2005 in [#48197]

[xcodec2] Fix flex attention and flash dispatch tests (#48244) by @jiqing-feng in [#48244]

Fix ROCm SDPA-flash skip guard that crashes on RDNA GPUs (#47965) by @Abdennacer-Badaoui in [#47965]

[GPT2] Fix encoder_attention_mask being silently discarded in cross-attention (#47946) by @DavidJohnQuinlan in [#47946]

Declare sdpa support in TimmWrapper (#47939) by @jiqing-feng in [#47939]

Quantization

Quantization improvements include adding NVFP4 quantization support via HF kernels (enabling on-the-fly BF16 weight quantization with ~50% memory reduction), and fixing several bugs: reverting a regression in is_quantization_compressed that caused incorrect module layouts for packed-format checkpoints, fixing CLIP weight initialization failures with quantized checkpoints, and restoring KV-cache quantization setup for KV-cache-only quantized models.

Revert "[Quantization]: Refactor is_quantization_compressed for format-based detection" (#48072) by @subin9 in [#48072]

feat: add nvfp4 quantization (#47883) by @drbh in [#47883]

[DeepSeekV2] Fix integration tests OOM: use device_map=auto instead of 8-bit quantization (#47991) by @ydshieh in [#47991]

Fix CLIP _init_weights when a child module carries quantized weights (#47921) by @Bluear7878 in [#47921]

Parallelization

Introduced a naive pipeline parallel inference engine supporting tied/untied weight embeddings with seamless generate() integration, while restoring backward compatibility for the tensor-parallel API with a deprecation cycle for tp_plan in from_pretrained(). Additionally fixed a model parallel bug in the BLT model affecting beam search.

Restore BC for the tensor-parallel API (#48300) by @ArthurZucker in [#48300]

fix bug for blt model parallel bug (#48327) by @kaixuanliu in [#48327]

Pipeline parallel naive inference (#47289) by @3outeille in [#47289]

Kernels

Kernel support was improved with documentation updates highlighting supported models, a fix for export crashes on kernel-decorated functions by adding a is_torchdynamo_exporting guard, and the default Flash Attention 2 hub kernel version was bumped to v3 to resolve compatibility issues with newer PyTorch versions.

[docs] Kernel supported models (#48258) by @stevhliu in [#48258]

[Fix] Export crashes on kernel-decorated function (#47808) by @remi-or in [#47808]

Bump default flash-attn2 hub kernel version to v3 (#47863) by @jiqing-feng in [#47863]

Bugfixes and improvements

Fix video-llama modular conversion (#48336) by @zucchini-nlp in [#48336]

CI: gate the hunyuan-moe slow test (#48330) by @tarekziade in [#48330]

Add a regression test for force_accelerate_hooks signature preservation (#48260) by @wtdcode in [#48260]

Add an opt-in per-frame pixel cap (cap_pixels_per_frame) to the Qwen3-VL video processor (#48071) by @dkrisman in [#48071]

Fix scores type in stopping criteria docstrings (#47676) by @qgallouedec in [#47676]

Docstring check didn't match some file - fix it (#48121) by @zucchini-nlp in [#48121]

Add shared ImageProcessingTester (#47745) by @guarin in [#47745]

Fix AutoTokenizer returning TokenizersBackend for DeepSeek-R1-Distill-Qwen models (#48211) by @ydshieh in [#48211]

[docs] Fix failing doctests (#47687) by @stevhliu in [#47687]

Fix build_2d_sinusoidal_position_embedding on MPS (#47897) by @guarin in [#47897]

[Gemma4] Investigate flaky test_generation_beyond_sliding_window_1_eager (#48236) by @ydshieh in [#48236]

Let gradient checkpointing skip layers with every_n_layers (#48200) by @qgallouedec in [#48200]

Disable daily nightly CI (#48292) by @remi-or in [#48292]

gs (#48288) by @eustlb in [#48288]

CI: fix muse OOMs (#48284) by @tarekziade in [#48284]

Fix BayesianDetectorModel.from_pretrained() by calling post_init() (#48254) by @woojinpaik in [#48254]

[Fix] Avoid duplicating tests in CI (#48287) by @remi-or in [#48287]

[GDN] Fix recurrent FLA fallback (#48266) by @vasqu in [#48266]

Fix tie_word_embeddings not lifted from text_config for some VLM configs (BC regression) (#45857) by @qgallouedec in [#45857]

replace xpu-smi subprocess call in benchmark_v2 (#48083) by @kaixuanliu in [#48083]

ignore mlinter ci file (#48267) by @tarekziade in [#48267]

Fix dtype mismatch in grouped_mm_fallback for LoRA training on Mamba+… (#47933) by @adh-aakriti in [#47933]

Compute MoE load-balancing loss per layer to avoid giant one-hot materialization −99.7% @ 128k (#48131) by @qgallouedec in [#48131]

deterministic layer_types buffer registration in multiple models (#48162) by @mowoe in [#48162]

Fix nemotron_h save_pretrained emitting singular backbone.embedding.weight (#48075) by @yuekaizhang in [#48075]

force_accelerate_hooks should not hide the signature it wraps (#48156) by @SunMarc in [#48156]

fix(data_collator): align TokenClassification numpy_call with torch_call (#48212) by @ in [#48212]

Fix a typo in a use of a local variable field_ in a test (#48184) by @AleksMat in [#48184]

[serge] Fix 2 integration tests regressed by commit 16780c8 (PR #47622) (#48134) by @sergereview[bot] in [#48134]

[Gemma4] Fix stale expected values in integration tests (#48233) by @ydshieh in [#48233]

[TableTransformer, PI0] Fix stale expected values and OOM in integration tests (#48198) by @ydshieh in [#48198]

Post two CI badges on a PR: CPU PR CI and GPU run-slow (#48190) by @tarekziade in [#48190]

Fix stale expected values in integration tests (cuda sm_86 / Aug04 regressions) (#48171) by @ydshieh in [#48171]

Assign a reviewer even when a codeowner has left, and route models by modality (#48085) by @tarekziade in [#48085]

[VITS] Un-skip test_model_forward (#46375) by @blipbyte in [#46375]

Apply context parallelism to the evaluation path (#48167) by @qgallouedec in [#48167]

Fix gpt_oss runs on GPU (#48118) by @tarekziade in [#48118]

[EsmFold2] Fix stale expected distogram logit values (#48182) by @ydshieh in [#48182]

Fix Apr 05 integration test regressions (cuda sm_86) (#48170) by @ydshieh in [#48170]

Add MLU support to is_flash_linear_attention_available (#46995) by @atri2549 in [#46995]

Fix integration test expected values for cuda sm_86 (Mar 15 regressions) (#48168) by @ydshieh in [#48168]

Fix DynamicCache reconstruction during ExecuTorch export (#47900) by @eladsegal in [#47900]

[LLaVA] Fix pixtral integration tests for cuda sm_86 (#48166) by @ydshieh in [#48166]

Port ESMC and ESMFold2 to Transformers (#46419) by @Rocketknight1 in [#46419]

[Qwen2.5-Omni] Update stale expected values for cuda sm_86 (#48164) by @ydshieh in [#48164]

[Mistral3] Fix batched integration tests: padding_side=left + update expected values (#48161) by @ydshieh in [#48161]

Fix DeepSeek V2 default vocab size (#48159) by @hmellor in [#48159]

[InternVL] Fix stale expected values for Llama integration tests (cuda sm_80) (#48153) by @ydshieh in [#48153]

Retry transient network errors (RemoteDisconnected) in github_utils (#48124) by @ydshieh in [#48124]

fix bugs for clvp model (#47127) by @kaixuanliu in [#47127]

Always tie embeddings for LongT5 and Pop2Piano (#47620) by @jiqing-feng in [#47620]

Use generator with seed for LengthGroupedSampler in Trainer._get_eval_sampler for deterministic eval order with per_device_eval_batch_size > 1 (#48025) by @philipshurpik in [#48025]

Enable mlinter findings artifact for inline PR reviews (#48117) by @ydshieh in [#48117]

Fix CpmAnt loading: size lm_head to vocab_size (#48012) by @jiqing-feng in [#48012]

Delete old mlinter review comments before posting new ones (#48107) by @ydshieh in [#48107]

[Video] Warn and return all frames when num_frames exceeds total_num_frames (#48074) by @carlszk in [#48074]

Accept artifact dir as argument in post_mlinter_review.py (#48106) by @ydshieh in [#48106]

Fix mlinter artifact path (#48088) by @ydshieh in [#48088]

[Video] Fix convert_to_rgb channel slicing and alpha blending for RGBA videos (#48053) by @ in [#48053]

[PE] Skip test_sdpa_can_dispatch_on_flash for TimmWrapper-backed models (#48064) by @ydshieh in [#48064]

fix(pipeline): preserve model.generation_config precedence over pipeline defaults (#47752) (#47953) by @nithin42 in [#47953]

[Gemma3] Update integration test expected values for A10G (#48036) by @ydshieh in [#48036]

Fallback from 'lanczos' to 'bicubic' when on cuda (#48026) by @zucchini-nlp in [#48026]

[serge] Fix 2 integration tests regressed by commit b9090ae (PR #47096) (#48060) by @sergereview[bot] in [#48060]

[Fix] Small FA-related test failures in CB (#47341) by @remi-or in [#47341]

fix: honor empty processor_kwargs={} in multimodal pipelines (#48044) by @ in [#48044]

[CircleCI] Enable CI for private forks, no-op for public repo (#48056) by @ydshieh in [#48056]

Support BatchFeature in length-grouped samplers (#48034) by @qgallouedec in [#48034]

Fix EOS for candidate generators (#47931) by @ in [#47931]

[Gemma3n] Update integration test expected values for A10G + torch 2.13 (#48035) by @ydshieh in [#48035]

fix failed test cases for muse_glimmer (#48011) by @kaixuanliu in [#48011]

Cohere compass tests (#47895) by @zucchini-nlp in [#47895]

docs: use relative paths for README language menus and add fa/ro entries (#47777) by @Priyans-Lathiya in [#47777]

Moving mlinter to 0.1.4 (#47918) by @tarekziade in [#47918]

[MoE] Fix Blackwell GPU crash with torch._grouped_mm on torch <= 2.8 (#48014) by @ in [#48014]

[docs] Muse Glimmer (#47882) by @stevhliu in [#47882]

[Florence2] Fix two integration test failures caused by torch 2.13 and auto-dtype (#48031) by @ydshieh in [#48031]

Proper separation of tests (#47943) by @zucchini-nlp in [#47943]

unpin pytest in the examples_torch deps (#48023) by @tarekziade in [#48023]

[CI] Fix startup failure in pr_build_doc_with_comment workflow by adding missing get-pr-number dependency (#47971) by @ in [#47971]

Let the GPU verify caller turn on the memory probe (#48001) by @tarekziade in [#48001]

Fix MTP config when mlp_layer_types is absent (#48015) by @Cyrilvallez in [#48015]

Remove duplicate block_sparse_moe assignment in GraniteMoeDecoderLayer (#47876) by @Aman2394 in [#47876]

[ModernVBERT] Fix integration test checkpoint (404 since April) (#48009) by @ydshieh in [#48009]

Fix cropping (#48006) by @Cyrilvallez in [#48006]

Fix DFlash candidate token device mismatch with device_map="auto" (#47877) by @sywangyi in [#47877]

🔴 Allow tokenizers 0.23.1 (#46381) by @ArthurZucker in [#46381]

[Whisper] Fix batch decode_with_timestamps in WhisperTokenizer.decode() (#47997) by @ydshieh in [#47997]

[OLMoE] Update expected logits for A10G and add torch.no_grad() (#47989) by @ydshieh in [#47989]

[OLMo] Fix OOM in logits tests by adding torch.no_grad() (#47986) by @ydshieh in [#47986]

[AXK1] Fix expected logits for CUDA A10G (#47980) by @ydshieh in [#47980]

[Gemma] Update expected values for A10G (#47976) by @ydshieh in [#47976]

Fix gemma4 video to device (#47896) by @guarin in [#47896]

Remove stale (None, None) fallback in qwen2_5_vl batch_different_resolutions test (#47972) by @ydshieh in [#47972]

Potential fix for code scanning alert no. 267: Artifact poisoning (#47949) by @tarekziade in [#47949]

Fix GatedDeltaNet A_log dtype to prevent -inf under bfloat16 init (#47944) by @Nkluge-correa in [#47944]

[serge] Fix 2 integration tests for model got_ocr2 failing with other (other (2)) (#47937) by @sergereview[bot] in [#47937]

Fix Jinja block endings in CHAT WITH MODELS' Writing a chat template … (#47960) by @ak1for2business-prog in [#47960]

[tests] Fix expected output for Qwen2.5-VL batch_wo_image on CUDA (#47968) by @ydshieh in [#47968]

[serge] Fix 2 integration tests for model opt failing with other (other (2)) (#47909) by @sergereview[bot] in [#47909]

Scan a diff in trufflehog, not the whole repo history (#47945) by @tarekziade in [#47945]

[serge] Fix 2 integration tests for model vivit failing with output_mismatch (tensor values differ (2)) (#47566) by @sergereview[bot] in [#47566]

Fix Gemma sliding_window being halved on every config save/reload (#47940) by @Bluear7878 in [#47940]

docs: fix incorrect PEFT anchor link in fine-tuning section (#47927) by @dsulot in [#47927]

CI: add vllm-test-init and vllm-test-transformers jobs on dedicated runners (#47934) by @ydshieh in [#47934]

Make muse glimmer exportable (#47871) by @IlyasMoutawwakil in [#47871]

docs: add installation instructions for NVIDIA Spark (ARM64) devices (#47906) by @mfuntowicz in [#47906]

[CohereCompass] Minor docs fixes (#47903) by @calpt in [#47903]

fix: correct checkpoints, config annotations, and create_dummy_models improvements (#47902) by @ydshieh in [#47902]

Transform paths and repeat joining for response parsing (#47648) by @Rocketknight1 in [#47648]

[docs] Update toctree (#47781) by @stevhliu in [#47781]

Add CI_CPU_MEMORY_LIMIT_GB to check_failed_tests workflow (#47884) by @ydshieh in [#47884]

Use tiny Hub checkpoint in Qwen3ASR processor test (#47833) by @ydshieh in [#47833]

Update AutoRound XPU/CPU backend (#47826) by @yiliu30 in [#47826]

docs(tests): fix typos in test comments (#47859) by @zhaoxinyi02 in [#47859]

Update version post release (#47870) by @Cyrilvallez in [#47870]

Significant community contributions

The following contributors have made significant changes to the library over the last release:

@tarekziade

CI: gate the hunyuan-moe slow test (#48330)

CI: fix muse OOMs (#48284)

ignore mlinter ci file (#48267)

Post two CI badges on a PR: CPU PR CI and GPU run-slow (#48190)

Assign a reviewer even when a codeowner has left, and route models by modality (#48085)

Fix gpt_oss runs on GPU (#48118)

Moving mlinter to 0.1.4 (#47918)

unpin pytest in the examples_torch deps (#48023)

Let the GPU verify caller turn on the memory probe (#48001)

Potential fix for code scanning alert no. 267: Artifact poisoning (#47949)

Scan a diff in trufflehog, not the whole repo history (#47945)

@dkrisman

Add an opt-in per-frame pixel cap (cap_pixels_per_frame) to the Qwen3-VL video processor (#48071)

@jiqing-feng

Cpmant fix use cache (#48013)

[xcodec2] Fix flex attention and flash dispatch tests (#48244)

Always tie embeddings for LongT5 and Pop2Piano (#47620)

Fix CpmAnt loading: size lm_head to vocab_size (#48012)

Fix Qwen2.5-Omni / Qwen3-Omni-MoE generation with a compilable cache (#47872)

Bump default flash-attn2 hub kernel version to v3 (#47863)

Declare sdpa support in TimmWrapper (#47939)

@ydshieh

Fix AutoTokenizer returning TokenizersBackend for DeepSeek-R1-Distill-Qwen models (#48211)

[Gemma4] Investigate flaky test_generation_beyond_sliding_window_1_eager (#48236)

[Gemma4] Fix stale expected values in integration tests (#48233)

[TableTransformer, PI0] Fix stale expected values and OOM in integration tests (#48198)

Fix stale expected values in integration tests (cuda sm_86 / Aug04 regressions) (#48171)

[EsmFold2] Fix stale expected distogram logit values (#48182)

Fix Apr 05 integration test regressions (cuda sm_86) (#48170)

Fix integration test expected values for cuda sm_86 (Mar 15 regressions) (#48168)

[LLaVA] Fix pixtral integration tests for cuda sm_86 (#48166)

[Qwen2.5-Omni] Update stale expected values for cuda sm_86 (#48164)

[Mistral3] Fix batched integration tests: padding_side=left + update expected values (#48161)

[InternVL] Fix stale expected values for Llama integration tests (cuda sm_80) (#48153)

Retry transient network errors (RemoteDisconnected) in github_utils (#48124)

[Whisper] Fix speculative decoding: preserve cleared suppress tokens through super().generate() (#48108)

Enable mlinter findings artifact for inline PR reviews (#48117)

Delete old mlinter review comments before posting new ones (#48107)

Accept artifact dir as argument in post_mlinter_review.py (#48106)

Fix mlinter artifact path (#48088)

[Whisper] Fix decoder position IDs for left-padded batches in longform generation (#48028)

[PE] Skip test_sdpa_can_dispatch_on_flash for TimmWrapper-backed models (#48064)

[Gemma3] Update integration test expected values for A10G (#48036)

[CircleCI] Enable CI for private forks, no-op for public repo (#48056)

[Gemma3n] Update integration test expected values for A10G + torch 2.13 (#48035)

[Florence2] Fix two integration test failures caused by torch 2.13 and auto-dtype (#48031)

[ModernVBERT] Fix integration test checkpoint (404 since April) (#48009)

[Whisper] Fix speculative decoding: UnboundLocalError, cache corruption, and speed regression (#48000)

[Whisper] Fix batch decode_with_timestamps in WhisperTokenizer.decode() (#47997)

[Whisper] Fix integration test failures on A10G (dtype, stale values, API changes) (#47995)

[DeepSeekV2] Fix integration tests OOM: use device_map=auto instead of 8-bit quantization (#47991)

[OLMoE] Update expected logits for A10G and add torch.no_grad() (#47989)

[OLMo] Fix OOM in logits tests by adding torch.no_grad() (#47986)

[GPTNeoX] Fix post_processor not overridden when loading from pretrained (OLMo garbage generation) (#47988)

[AXK1] Fix expected logits for CUDA A10G (#47980)

[Gemma] Update expected values for A10G (#47976)

[emu3] 🦮 Black Labrador is back! Fix image generation broken since #37033 (#47948)

Remove stale (None, None) fallback in qwen2_5_vl batch_different_resolutions test (#47972)

[tests] Fix expected output for Qwen2.5-VL batch_wo_image on CUDA (#47968)

CI: add vllm-test-init and vllm-test-transformers jobs on dedicated runners (#47934)

fix: correct checkpoints, config annotations, and create_dummy_models improvements (#47902)

Add CI_CPU_MEMORY_LIMIT_GB to check_failed_tests workflow (#47884)

Use tiny Hub checkpoint in Qwen3ASR processor test (#47833)

@eustlb

gs (#48288)

@eladsegal

Fix DynamicCache reconstruction during ExecuTorch export (#47900)

Support per-layer cache configuration and attention-mask selection (#47901)

@YangKai0616

🚨[wav2vec2] Support attn_implementation=sdpa dispatch (#46196)

@drbh

feat: add nvfp4 quantization (#47883)

@Priyans-Lathiya

docs: use relative paths for README language menus and add fa/ro entries (#47777)

@itazap

[new model] step 3.7 (#46658)

@calpt

[CohereCompass] Minor docs fixes (#47903)

Add CohereCompass modeling (#47878)

Original source

Tip

This release features several new pipelines, including LTX2.5, MiniMax H3, and Wan Animate 2. We're also graduating Modular Diffusers out of the experimental phase and announcing its stable support. Additionally, this release includes minimal support for tensor-parallel. There's a lot more that went down in this release. So, please consult the notes for details.

New Pipelines

MiniMax-H3

MiniMax-H3 generates video and its soundtrack together. A single transformer denoises one packed sequence containing the text conditioning, the conditioning media, and the target video and audio latents — there is no separate vocoder and no post-hoc audio pass. Its conditioner is a Qwen3VLForConditionalGeneration whose unnormalized 50th-decoder-layer hidden state is read instead of the last one.

MiniMax-H3 is integrated as Modular Diffusers blocks only — MiniMaxH3Blocks and their MiniMaxH3ModularPipeline are the whole integration. The conversion ships both checkpoint partitions in one repository and exposes three workflows (t2va, fl2va, ref2va) that can be pruned at from_pretrained time so only that task's components are declared and downloaded.

MiniMax Music 3

MiniMax Music 3 produces complete songs up to five minutes long from lyrics and a music description, with expressive vocals and long-range structure. It is a hybrid of an autoregressive and a diffusion stage: an 8B Qwen3-based global language model predicts one semantic audio token per frame while a small depth decoder fills in seven residual RVQ codebooks, and their fused hidden states condition a 2.4B flow-matching transformer that produces Flow-VAE latents in overlapping chunks. A DAC-style decoder turns the latents into 44.1 kHz stereo audio.

Stable Audio 3

Stable Audio 3 is a text-to-audio model from Stability AI that generates high-quality stereo audio at 44.1 kHz. It uses a rectified-flow DiT conditioned on a frozen T5Gemma text encoder (via cross-attention) and on duration (a float embedded by StableAudio3DurationEmbedder and used for adaptive layer norm), and decodes with the SAME (Semantically-Aligned Music Encoder) autoencoder, AutoencoderSAME.

Three pipelines ship: StableAudio3Pipeline, StableAudio3AudioToAudioPipeline, and StableAudio3InpaintPipeline.

Thanks to @buffett0323 for the contribution (#14119).

LTX-2.5

LTX-2.5 reuses the existing LTX2Pipeline / LTX2VideoTransformer3DModel / AutoencoderKLLTX2Video classes — there is no separate pipeline class. The user-visible difference is the text encoder: LTX-2.5 is paired with a Gemma 4 (gemma4_unified) checkpoint, loaded automatically from a converted LTX-2.5 repo.

Lightricks/LTX-2.5-Diffusers ships both the distilled DiT (transformer/) and the full/SFT DiT (transformer_full/), plus everything two-stage generation needs. Alongside the checkpoint support, this release adds:

• LTX2VideoDiffusionDecoderModel and LTX2VideoDiffusionDecodePipeline — a second video decoder over the same latent space, so latents are interchangeable between decoders.
• A duration_head that predicts shot length from the text-connector output, so num_frames is auto-predicted by default when the loaded pipeline has one.
• Prompt enhancement through a separate off-the-shelf google/gemma-4-E2B-it checkpoint (enable_prompt_enhancement=True).
• LTX25AutoBlocks for Modular Diffusers (#14453).

Wan-Animate-2

Wan-Animate-2 by the Alibaba Wan Team animates a reference character image with the motion of a driving video. The driving video is processed in fixed-length segments: each segment runs a reference-extraction pass that caches the driving segment's K/V in every transformer layer, denoises against that cache, and is decoded inside the loop, because the next segment conditions on the previous segment's decoded tail frames.

Two presets are available — the base checkpoint samples with classifier-free guidance, and the distilled checkpoint samples in few steps without it. Guidance is owned by the pipeline's guider component, so there is no guidance_scale argument.

Thanks to @kelseyee for authoring the integration (#14413).

JoyAI-Image-Edit-Plus

JoyAI-Image-Edit-Plus extends the JoyAI-Image family (an 8B MLLM paired with a 16B MMDiT) to multi-image instruction-guided editing. It accepts 1–5 reference images plus a text instruction and composes elements from the references into a new image.

Thanks to @tangyanf for the contribution (#14032).

Cosmos 3 follow-ups

Cosmos 3 landed in 0.39.0 and gets substantially more coverage in this release:

• A Cosmos 3 Modular pipeline (Cosmos3OmniModularPipeline), with Transfer support for precomputed control videos (edge, blur, depth, segmentation, world-scenario maps) generated autoregressively in chunks and stitched automatically.
• Edge support.
• Distilled (few-step) text-to-image and image-to-video via Cosmos3DistilledModularPipeline.
• Model parallelism.

Thanks to @yzhautouskay and @atharvajoshi10 for the contributions.

Modular Pipeline Support

Modular Diffusers is no longer marked experimental (#14525) — the API warning has been dropped.

Three of this release's new integrations — MiniMax-H3, MiniMax Music 3, and Wan-Animate-2 — ship as Modular blocks only, with no DiffusionPipeline half.

Cosmos 3 (#14110, #14150, #14177) and Krea 2 (#14083, thanks to @Cedric-Perauer) gained modular pipelines, and Anima gained img2img blocks (#13929).

Branch-specific input defaults (#14234): when sibling blocks of a ConditionalPipelineBlocks declare different defaults for the same input, combine_inputs now merges the default to None and records the per-block defaults in a new InputParam.defaults_by_block field. get_block_state falls back to the block's own declared default, so each branch resolves its own default when it actually runs. Docstrings render conflicted defaults as e.g. "defaults to None or 189, depending on the workflow".

Group offloading now works under auto offloading, and a group-offload device mismatch was fixed.

Unused intermediate_inputs were cleaned up, stale auto-docstrings are now detected in CI, and Mellon custom-block required-input handling was fixed.

Important

Please try out Modular Diffusers and let us know about your feedback!

Core Library

Tensor parallelism

Tensor-parallel inference is now supported for model inference on CUDA and AWS Neuron (Trainium/Inferentia), exposed through the same public API already used for context parallelism:

from diffusers import TensorParallelConfig
pipe.transformer.enable_parallelism(config=TensorParallelConfig(mesh=tp_mesh))

Sharding is model-agnostic and driven from a flat _tp_plan, which has been added to the FLUX.1, FLUX.2 and Qwen-Image DiTs. Check out the docs for more details.

New quantization backends

SDNQ — a training-free backend supporting int8 down to 2-bit, FP8 and other low-bit float formats, with optional SVD correction, Hadamard rotation, and quantized INT8/FP8 matmul. Runs on CUDA, ROCm, XPU, MPS, and CPU. SDNQ MiniMax-H3 loading landed too. Thanks to @asomoza.

Nunchaku Lite for loading prequantized checkpoints, with fused-kernel performance documented. Thanks to @rootonchair.

Attention, kernels and offloading

Kernels are downloaded only when users request them

ROCm FlashAttention now uses the kernels-community/aiter-flash-attn-ck Hub kernel, dropping the aiter dependency. Thanks to @Abdennacer-Badaoui.

Sequence-parallel support for the _flash_3_varlen_hub backend and a mask-handling fix. Thanks to @zhtmike.

DiffusionPipeline.device deduction for split-device pipelines

Hook forward signatures are preserved. Thanks to @jloftin-nv.

CLI

The diffusers-cli was reworked for agentic use (#13966) and then cleaned up (#14381): modular_model_index.json is written when saving a custom block so ModularPipeline.from_pretrained can load and run custom blocks as pipelines, auto CPU offload works for Modular pipelines, workflow can be passed to Modular pipelines, and output saving handles multimodal output (e.g. LTX video frames + audio) and batched video.

Skills are now installed through the CLI rather than the Makefile (#14454):

diffusers-cli skills list
diffusers-cli skills add <skill name>

🚨 Breaking changes and deprecations

JAX/Flax support has been removed (#14169) — roughly 12k lines deleted. All Flax* classes and the flax extras are gone.

Mixed-rank LoRAs without alpha keys now load at their intended scale (#14409). Previously get_peft_kwargs took lora_alpha from the first entry of the rank dict and never revisited it, so every module whose rank differed from the first key's rank got an arbitrary, key-order-dependent scale. Ranks are now mirrored into the alphas when a checkpoint brings no alpha information (the diffusers/PEFT convention: alpha == rank, scale 1.0). Adapters with a declared alpha keep it, and uniform-rank adapters are unaffected. Existing mixed-rank, no-alpha LoRAs will now produce different (correct) results.

torch_dtype is deprecated in favour of dtype (#14205, #14313), following transformers. torch_dtype still works but warns, and will be removed in 1.0.0. A torch.dtype alias was added for the docs.

DDUF is deprecated (#14141). dduf_file warns and will be removed in 0.41.0.

Security

Fixed a path-traversal / arbitrary out-of-directory file read via sharded checkpoint index files. Thanks to @buffett0323.

A SECURITY.md was added.

Training

Caption dropout and aspect-ratio buckets for the Krea 2 and Qwen-Image LoRA scripts, plus an aspect-ratio bucketing fix and on-the-fly buckets across the DreamBooth scripts. Thanks to @linoytsaban.

Krea 2 LoRA cards are tagged with the Turbo base model first so the inference widget works.

LoRA

ACE-Step LoRA support. Thanks to @chenyangzhu1.

Fix LoRA hot-swapping recompilation with different_shapes_for_compilation. Thanks to @jiqing-feng.

Only drop an adapter from _merged_adapters when it is unfused from all components.

Fix local LoRA weight auto-discovery in offline mode. Thanks to @fropych.

Notable fixes

Fix Kandinsky 5 I2V conditioning — the image latent is no longer injected into visual_cond channels.

Fix batched DiffusionGemma adaptive stopping

Fix image preprocessing for Cosmos 3 and Cosmos 3 Edge generator K normalization

Fix AuraFlow model-parallelism device mismatch and AuraFlow VAE dtype mismatch on pipeline reuse

Make WanTransformer3DModel and SD3Transformer2DModel hidden states contiguous

Fix the rectified-flow time scheduler

Fix Helios auto-offload decode

Align snapshot_download with the latest huggingface_hub

Reduce FLUX int8 test peak memory with sequential offload

Tests and infrastructure

A large chunk of this release is test modernization: pipeline tests continue migrating to the new mixin structure (Wan, Qwen-Image, FLUX.2, CogVideoX, Stable Diffusion, and the LoRA pipeline tests), tests/others, training tests, and attention-processor tests moved to pytest, model-level and pipeline-level quantization tests were standardized, and an output_shape property was introduced in the pipeline tests. The agent-facing docs and skills under .ai/ were expanded to cover tests, model implementation, and blockset conventions.

All commits

Fix model offloading and training tests + prevent examples timeout by @GiGiKoneti in #14091
fix _flash_3_varlen_hub mask handling by @zhtmike in #14115
Make doc builds faster by @mishig25 in #14131
[tests] fix autoencoderdc ml training fix. by @sayakpaul in #14129
[tests] fix tests for transformers latest stable. by @sayakpaul in #14120
fix rf time scheduler problem by @TheLovesOfLadyPurple in #14011
docs: improve docstring scheduling_lcm.py by @delmalih in #14133
feat: add JoyImage edit plus by @tangyanf in #14032
update licensing for others to 2026. by @sayakpaul in #14128
[fix] missing doc-builder imports by @stevhliu in #14134
add SP support for _flash_3_varlen_hub backend by @zhtmike in #13809
post release v0.39.0 by @sayakpaul in #14116
deprecate dduf. by @sayakpaul (direct commit on v0.40.0-release)
Revert "deprecate dduf." by @sayakpaul (direct commit on v0.40.0-release)
docs: improve docstring scheduling_scm.py by @delmalih in #14136
Cosmos3 ModularPipeline by @yzhautouskay in #14110
docs: improve docstring scheduling_sasolver.py by @delmalih in #14148
Fix Helios auto offload decode by @pzarzycki in #14140
docs: improve docstring scheduling_sde_ve.py by @delmalih in #14172
Add Nunchaku Lite single-file quantization by @rootonchair in #14100
Transfer support for Cosmos3 ModularPipeline by @yzhautouskay in #14150
torch.dtype alias by @stevhliu in #14162
ask to share self-review notes by @yiyixuxu in #14185
Make SD3Transformer2DModel hidden states contiguous by @menglcai in #14186
align snapshot_download to respect hfh latest version by @sayakpaul in #14118
[skills] add notes about model implementation in our skills. by @sayakpaul in #14191
[lora training] tag krea2 LoRA card with Turbo base model first to enable inference widget by @linoytsaban in #14171
docs: improve docstring scheduling_tcd.py by @delmalih in #14192
docs: improve docstring scheduling_unclip.py by @delmalih in #14196
docs: improve docstring scheduling_repaint.py by @delmalih in #14199
[core] deprecate dduf by @sayakpaul in #14141
Fix local LoRA weight auto-discovery in offline mode by @fropych in #14204
Add tests and agent docs for kwargs_type input/output by @yiyixuxu in #14157
[Anima] Add img2img pipeline blocks by @PreethamNoelP in #13929
update agent doc to cover more on tests + include it in review CI scope by @yiyixuxu in #14197
Cosmos3 edge support by @atharvajoshi10 in #14181
Fixing path traversal/arbitrary out-of-directory file read via sharde… by @buffett0323 in #14182
add security.md as per internal notice. by @sayakpaul in #14174
[Quantization] Fix ModelOpt pre-quantized loading by @yzhautouskay in #14188
[tests] Pipeline test refactor by @sayakpaul in #14113
docs: improve docstring scheduling_k_dpm_2_discrete.py by @delmalih in #14212
Cosmos3 Distilled support by @yzhautouskay in #14177
Add not_params to ModularPipelineTesterMixin by @yiyixuxu in #14207
[agents doc] notes on when to create new blocksets for checkpoint variant by @yiyixuxu in #14208
Fix AuraFlow VAE dtype mismatch on pipeline reuse by @IvenHsu01 in #14184
[tests] fix autoencoderkl dtype tests by @sayakpaul in #14190
refactor qwenimage pipeline tests to the new mixin structure by @akshan-main in #14220
Add ace step lora support by @chenyangzhu1 in #14193
[skills] update test skills for standard pipelines. by @sayakpaul in #14223
Make WanTransformer3DModel hidden states contiguous before the block loop by @menglcai in #14236
refactor wan 2.2 image-to-video pipeline tests to the new mixin structure by @akshan-main in #14242
refactor wan vace pipeline tests to the new mixin structure by @akshan-main in #14231
refactor wan 2.2 pipeline tests to the new mixin structure by @akshan-main in #14240
refactor wan image-to-video pipeline tests to the new mixin structure by @akshan-main in #14228
refactor wan pipeline tests to the new mixin structure by @akshan-main in #14224
refactor wan video-to-video pipeline tests to the new mixin structure by @akshan-main in #14235
refactor wan animate pipeline tests to the new mixin structure by @akshan-main in #14239
fix test_pag_sdxl::StableDiffusionXLPAGPipelineIntegrationTests by @sywangyi in #13892
[chore] remove experimental api warning from lora modules. by @sayakpaul in #14248
[tests] fix cosmos3 tests by @sayakpaul in #14229
Fix Cosmos3 Edge generator K normalization by @atharvajoshi10 in #14246
[core] deprecate torch_dtype and prefer dtype following transformers. by @sayakpaul in #14205
Remove JAX/Flax by @DN6 in #14169
[tests] port attention processor tests to use pytest by @sayakpaul in #14161
[Modular Diffusers] Fix group offload device mismatch by @asomoza in #14252
[lora training] fix aspect ratio bucketing in dreambooth scripts (+ caption dropout, on-the-fly buckets) by @linoytsaban in #14158
Reduce FLUX int8 test peak memory with sequential offload by @jiqing-feng in #13776
Update diffusers-cli for agentic use by @DN6 in #13966
[modular] detect stale auto docstrings in CI + regenerate current ones by @yiyixuxu in #14241
Fix Wan and Motif video pipeline fast test failures by @sywangyi in #14269
Krea2 Modular Pipeline Support, Documentation, Test Cases by @Cedric-Perauer in #14083
[ci] ignore the token in vq_diffusion converter. by @sayakpaul in #14275
Cosmos3 Model Parallelism by @atharvajoshi10 in #14054
[tests] component manager tests by @sayakpaul in #13961
[modular] clean up unused intermediate_inputs by @yiyixuxu in #14278
refactor cogvideox video to video pipeline tests to the new mixin structure by @akshan-main in #14289
refactor cogvideox pipeline tests to the new mixin structure by @akshan-main in #14276
refactor cogvideox image to video pipeline tests to the new mixin structure by @akshan-main in #14283
refactor cogvideox fun control pipeline tests to the new mixin structure by @akshan-main in #14284
[tests] migrate training tests to pytest. by @sayakpaul in #14267
[tests] fix some CLI command tests by @sayakpaul in #14302
[lora training] add caption dropout and aspect ratio buckets to krea2 and qwen-image by @linoytsaban in #14089
added the technical blog by @charchit7 in #14306
[core] restrict logging of quant config based on the header size. by @sayakpaul in #14262
docs: improve docstring scheduling_k_dpm_2_ancetstral_discrete.py by @delmalih in #14226
[docs] add a warning note about device_map='cuda' usage. by @sayakpaul in #14261
[docs] minor corrections in the overview of quantization table by @sayakpaul in #14301
[docs] deprecate torch_dtype for dtype by @stevhliu in #14313
refactor qwenimage controlnet pipeline tests to the new mixin structure by @akshan-main in #14309
refactor qwenimage edit plus pipeline tests to the new mixin structure by @akshan-main in #14315
refactor qwenimage edit pipeline tests to the new mixin structure by @akshan-main in #14314
Preserve hook forward signature by @jloftin-nv in #14285
Fix LoRA hot-swapping recompilation with different_shapes_for_compilation by @jiqing-feng in #14297
refactor qwenimage img2img pipeline tests to the new mixin structure by @akshan-main in #14320
[docs] remove image generation benchmarks by @stevhliu in #14310
docs: improve docstring scheduling_ddpm_wuerstchen.py by @delmalih in #14319
[docs] update agentic contribution guidelines by @stevhliu in #14312
[tests] unify keep_in_32 tests in quantization test suites by @sayakpaul in #14257
[quantization] SDNQ core loading by @asomoza in #14277
docs: improve docstring scheduling folder - last batch by @delmalih in #14330
Modular branch specific defaults by @yiyixuxu in #14234
fix Mellon custom block required input handling by @wunianze666-netizen in #13888
gguf: fix dequantization. by @sayakpaul in #14333
[tests] standardize model-level quant tests by @sayakpaul in #14332
Add missing Args: entries to scheduler docstrings by @4ktLuffy in #14354
support group offloading under auto offloading by @yiyixuxu in #14358
refactor flux2 klein inpaint pipeline tests to the new mixin structure by @akshan-main in #14337
[kernels] download kernels when users request for it. by @sayakpaul in #14298
refactor flux2 klein pipeline tests to the new mixin structure by @akshan-main in #14336
refactor flux2 klein kv pipeline tests to the new mixin structure by @akshan-main in #14344
Auto-reply to new model requests with remote code guidance by @yiyixuxu in #14343
Add XPU expected slice for SlowBnb4BitFluxControlWithLoraTests::test_lora_loading by @jiqing-feng in #14202
[tests] Migrate tests/others to pytest. by @sayakpaul in #14299
[Kandinsky 5] Fix I2V conditioning: don't inject the image latent into visual_cond channels by @sashakunitsyn in #14282
use assert_tensors_close in the migrated pipeline tests by @akshan-main in #14369
Add MiniMax-H3 by @apolinario in #14355
[Quantization] SDNQ Minimax H3 loading by @asomoza in #14398
refactor qwenimage inpaint pipeline tests to the new mixin structure by @akshan-main in #14321
refactor flux2 pipeline tests to the new mixin structure by @akshan-main in #14326
[tests] introduce output_shape property in the pipeline tests. by @sayakpaul in #14253
Point the MiniMax-H3 docs install note at main by @apolinario in #14401
[core] support cp in h3. by @sayakpaul in #14407
[tests] tighten tests for unload_components() by @sayakpaul in #14406
[tests] implement todos in the modular tests for H3 followup by @sayakpaul in #14404
[tests] fix torchao tests by @sayakpaul in #14258
fix(lora): only drop adapter from _merged_adapters when unfused from all components by @AloysJehwin in #14385
[tests] guard peft imports in test_lora_loader_utils by @sayakpaul in #14434
[core] DiffusionPipeline.device deduction for split-device pipelines by @JingyaHuang in #14383
🚨 Give mixed-rank LoRAs without alpha keys their intended scale by @apolinario in #14409
Ltx 2.5 by @sayakpaul in #14447
docs: add installation instructions for NVIDIA Spark (ARM64) devices by @mfuntowicz in #14448
Wan-Animate-2 (authored by @kelseyee) by @yiyixuxu in #14413
Move LTX25AutoBlocks to new file with new model name by @dg845 in #14453
[tests] tighten keep_in_32 modules tests by @sayakpaul in #14399
Fix AuraFlow model parallelism device mismatch and update XPU IP-Adap… by @sywangyi in #14273
feat: clean up peft related versioning voodoo. by @sayakpaul in #14463
docs: document fused-kernel performance in Nunchaku Lite guide by @rootonchair in #14458
[Agent docs] some updated based on recent integration by @yiyixuxu in #14452
[CI][tests] Fix LTX-2.5 Diffuson Decoder Tiling CI CPU Timeout by @dg845 in #14462
Add MiniMax Music 3 by @apolinario in #14456
Add LoRA loading support for MiniMax-H3 by @apolinario in #14408
[FA2] Use kernels-community/aiter-flash-attn-ck Hub kernel for ROCm FlashAttention (drop aiter dependency) by @Abdennacer-Badaoui in #14436
split up tests in modular pipelines by @sayakpaul in #14444
svd failures. by @sayakpaul (direct commit on v0.40.0-release)
Revert "svd failures." by @sayakpaul (direct commit on v0.40.0-release)
Bump transformers from 4.47.0 to 5.5.0 in /examples/flux-control by @dependabot[bot] in #14179
FIX LoRA tests warning about unexpected keys by @BenjaminBossan in #14476
[tests] use eval() on t5 for deterministic results. by @sayakpaul in #14472
tests: fix qwen tests from getting oom'd in our CI. by @sayakpaul in #14474
fix SVD tests by @sayakpaul in #14475
tests: fix cuda model tests. by @sayakpaul in #14464
ci: drop stale tokenizers<0.23.0 override by @sayakpaul in #14504
Drop the experimental warning from Modular Diffusers by @yiyixuxu in #14525
[tests] skip parallelism tests for some models. by @sayakpaul in #14505
Fix batched DiffusionGemma adaptive stopping by @kashif in #14386
docs: fix typos in Kandinsky 5.0 video docs by @iridescentWen in #14345
Support DiffSynth-Studio MiniMax-H3 LoRAs by @apolinario in #14484
[tests] refactor stable diffusion pipeline tests by @sayakpaul in #14339
[tests] refactor pipeline-level quantization tests by @sayakpaul in #14435
replace enable_gpa with repeat_interleave for krea2 by @yiyixuxu in #14523
[CLI] Clean up and improvements to diffusers CLI by @DN6 in #14381
[CI] Allow single file tests to run as fast tests by @DN6 in #14324
Feat -- Stable Audio 3 by @buffett0323 in #14119
[core] Support tensor parallelism for model inference (CUDA, Neuron) by @JingyaHuang in #13718
[CI] Fix authentication on forked PRs by @DN6 in #14534
fix: image preprocessing for cosmos3 by @atharvajoshi10 in #14519
[tests] migrate lora tests for pipelines to use new mixins by @sayakpaul in #14268
[distributed] fix corrupted gradient problem under ring CP. by @sayakpaul in #14274
Restructure skill files and reference guides for CLI based installation. by @DN6 in #14454
[CI] Fix build and test step in release workflow by @DN6 in #14543
Release: v0.40.0-release by @sayakpaul (direct commit on v0.40.0-release)

Original source

Patch release v5.15.1

This patch most notably solves a few issues with DFlash and MTP candidate generators, as well as an issue where images could sometimes not be processed on accelerator if using Lanczos filter.

It contains the following commits:

• Fix DFlash candidate token device mismatch with device_map="auto" (#47877) by @sywangyi and @Cyrilvallez
• Align logit distributions for CandidateGenerators using sampling (#48007) by @Cyrilvallez
• Fix MTP config when mlp_layer_types is absent (#48015) by @Cyrilvallez
• Fallback from 'lanczos' to 'bicubic' when on cuda (#48026) by @zucchini-nlp
• Fix gemma4 video to device (#47896) by @guarin

Original source

Release v5.15.0

New Model additions

Meta Muse Glimmer

Muse Glimmer, released today, is Meta’s new multimodal model, especially designed for agentic use cases. Distilled from Muse to 30B parameters, and released under the Apache 2.0 license, it can be deployed to local setups for privacy-aware applications such as coding, document analysis, personal assistants, Claw- or Hermes-like setups.

Muse Glimmer is a dense 30B parameter model consisting of:

• 2B ViT-style encoder for vision (Perception Encoder)
• 28B parameter text decoder

We're covering it in the following blogpost: http://hf.co/blog/muse-glimmer

GraniteMoeSWA & GraniteSWA

Links: Documentation

Add Granite-swa and Granitemoe-swa model support (#47179) by @daviswer in #47179

Links: Documentation

Add Granite-swa and Granitemoe-swa model support (#47179) by @daviswer in #47179

A.X-K1 & A.X-K2

Links: Documentation

Add AXK2 from SKT (#47528) by @vasqu in #47528

Links: Documentation

add_axk1 (#46867) by @kmswin1 in #46867

Cosmos3 Edge

Links: Documentation

Add Cosmos3 Edge model support (#47181) by @atharvajoshi10 in #47181

Breaking changes

Kernels are now opt-in rather than mandatory for linear attention models (Mamba, GDN, Conv-only, etc.), so users who relied on automatic kernel selection must explicitly enable kernels to maintain previous behavior.

🚨 [Kernels] Refactor all linear attn models & native kernels fallback (#47630) by @vasqu

The cache cropping API now only accepts negative values (relative offsets) instead of absolute sizes, so users calling crop methods directly must update their code to pass negative values accordingly.

🚨 [cache] Cropping can only be done with negative values (#47720) by @Cyrilvallez

T5 and its model family (MT5, LongT5, etc.) now support SDPA and other attention backends via ALL_ATTENTION_FUNCTIONS, meaning the default attention implementation may change and users relying on the previous eager-only path should explicitly set attn_implementation="eager" if needed.

🚨 Enable SDPA (and other attention backends) for T5 and propagate to the T5 family (#47014) by @jiqing-feng

Several small private helper functions (e.g., _is_url, _build_image_tokens) have been removed from multimodal processor files, so users or downstream libraries that imported these private functions directly must remove or replace those references.

🚨 Processors update the rest (#46556) by @zucchini-nlp

Attention

This release includes several attention fixes and improvements, including correcting Multi-Head Latent Attention (MLA) cache compression, optimizing Flash Attention max sequence length computation in vision models, and fixing bugs in CTRL flex-attention and SDPA prefill with position bias. Additional changes refactor linear attention models for better maintainability, make Gemma 4's heterogeneous attention config explicit, and improve MPS support via metal-flash-sdpa integration.

[Fix] Fix multi-head latent attention (MLA) (#47761) by @remi-or in [#47761]

Refactor all linear attention models to latest best standards for convolution (#47452) by @Cyrilvallez in [#47452]

Allow metal-flash-sdpa for OpenAIPrivacyFilter on MPS (#46740) by @ArthurZucker in [#46740]

Use new per_layer_config for Gemma 4 so that heterogeneous attention config is explicit (#47384) by @hmellor in [#47384]

add paged attention tests support for XPU (#47163) by @kaixuanliu in [#47163]

Move value padding into the attention interfaces that need it (#47451) by @hmellor in [#47451]

Simplify function dispatch for linear attention (#47450) by @Cyrilvallez in [#47450]

Optimize flash attention max seqlen computation in vision attention (#47170) by @ShareLer in [#47170]

Fix BlockMask crash in CTRL flex-attention generation (#46854) by @jiqing-feng in [#46854]

[CB] Automatically switch attention implementation to flash (#47330) by @remi-or in [#47330]

Fix sdpa prefill with position_bias (#47359) by @Cyrilvallez in [#47359]

Vision

Vision improvements in this release include performance optimizations such as faster image preprocessing for vision-language models (GLM4V, MiniMaxM3-VL, and others) by eliminating redundant tensor copies, and more efficient Flash Attention variable-length paths by precomputing maximum sequence lengths once per forward pass. Several bug fixes were also applied, including correcting dtype alignment in Kosmos2/Kosmos2_5 embedding merges, fixing a position-embedding initialization fallback in Phi4Multimodal, resolving PIL resize parity in Hunyuan-VL, and patching stop-sequence handling in the image-text-to-text pipeline.

Modularize qwen-format vision processors (#47573) by @zucchini-nlp in [#47573]

Update daily CI Docker image to torch 2.13.0 / CUDA 13.0 (#47738) by @ydshieh in [#47738]

Align image feature dtype in kosmos2 and kosmos2_5 embedding merge (#47691) by @ in [#47691]

Speed up image preprocessing for vision-language models (#47453) by @labAxiaoming in [#47453]

Fix vision position-embedding init width fallback in Phi4Multimodal (#47509) by @ in [#47509]

Fix Hunyuan-VL PIL image resize parity with reference preprocessing (#47233) by @IMvision12 in [#47233]

Fix image-text-to-text stop_sequence handling (#47032) by @Sunt-ing in [#47032]

Refactor image loading in tests to use load_test_image helper (#47218) by @LevelVoid in [#47218]

Generation

Several generation improvements and bug fixes were made, including enabling batched audio generation for Qwen2.5/3-Omni, allowing sliding window cache layers to work with speculative decoding, and fixing memory overhead from static cache persistence across generate() calls. Multiple model-specific bugs were also resolved, including crashes in KyutaiSpeechToText, MusicgenForCausalLM, CTRL flex-attention, and assisted decoding for EncoderDecoder cache and OlmoHybrid models.

Align OlmoHybrid to use a native cache in generate (#47604) by @Cyrilvallez in [#47604]

[generate] Stop setting the static cache as an attribute to save memory (#47731) by @Cyrilvallez in [#47731]

Add support for batched Qwen2.5/3-Omni audio generation (#47186) by @IMvision12 in [#47186]

[cache] Allow sliding window layers to be roll-backed for speculative decoding (#47447) by @Cyrilvallez in [#47447]

Fix shape mismatch in KyutaiSpeechToText generate() last window (#46952) by @jiqing-feng in [#46952]

Fix typo in MusicgenForCausalLM.generate() (#46974) by @jiqing-feng in [#46974]

Fix assisted decoding for models with EncoderDecoder cache & OlmoHybrid (#47361) by @Cyrilvallez in [#47361]

Cache

Several cache-related bugs were fixed, including correcting NemotronH's missing "mlp" layer-type mapping, resolving recurrent-layer padding masks being skipped during chunked prefill and cache continuation for hybrid models, and fixing assisted decoding for models with EncoderDecoderCache and OlmoHybrid. Additional improvements include aligning OlmoHybrid to use a native cache, enabling sliding window layers to support speculative decoding rollback, and stopping the static cache from being stored as a model attribute to reduce unexpected memory overhead.

[docs] MPS graph cache (#47304) by @stevhliu in [#47304]

Fix NemotronH: Register "mlp" in the cache layer-type mappings (#47535) by @qgallouedec in [#47535]

Fix recurrent-layer padding mask being skipped on continued forwards (chunked prefill, cache continuation) (#47087) by @abcgco in [#47087]

Kernels

⚠️ The kernels python package will very likely be a required dependency for transformers[torch] in the near future. This will help us deliver maximum performance to all users; kernels will only be downloaded from trusted publishers manually approved by the HF team. Please let us know of any issues you're facing beforehands so that we may solidify our integration.

Improved robustness of the kernels integration by refactoring function handling to use layer repos, fixing CI EROFS fallback patches for kernel downloads via HfApi, resolving a positional argument collision in causal_conv1d_fn, and bumping the FP8 kernels version to prevent NaNs.

[conftest] Fix EROFS fallback for kernel downloads (correct interception point) (#47794) by @ydshieh in [#47794]

[conftest] Fix EROFS fallback for kernel downloads via HfApi (#47791) by @ydshieh in [#47791]

[Kernels] Refactor function handling (#46883) by @vasqu in [#46883]

Kernels and loaders robustification (#47334) by @IlyasMoutawwakil in [#47334]

Fix causal_conv1d_fn positional activation colliding with hub kernel's seq_idx (#47527) by @qgallouedec in [#47527]

[FP8] Bump kernels version (#47344) by @vasqu in [#47344]

[docs] FlashAttention kernel fallback (#47345) by @stevhliu in [#47345]

Quantization

Quantization support was expanded with FP8 kernels for compressed-tensors models, fixes for FP8 module normalization and format-based compression detection, and a multi-device MXFP4 dequantization race condition fix. GPTQ and MXFP4 tests were also extended to cover Intel XPU devices.

extend tests/quantization/gptq/test_gptq.py::GPTQTestCUDA and tests/q… (#47166) by @sywangyi in [#47166]

Compressed tensors fp8 (#47216) by @SunMarc in [#47216]

Fix A.X-K2 fp8 modules_to_not_convert normalization for the gated-norm MLP (#47578) by @kmswin1 in [#47578]

[Quantization]: Refactor is_quantization_compressed for format-based detection (#47152) by @rigen1048 in [#47152]

Fix multi-device mxfp4 dequantization race in _convert_moe_packed_tensors (#47423) by @kaixuanliu in [#47423]

Audio

Batched audio generation is now supported for Qwen2.5/3-Omni, and several bug fixes were applied across audio models, including a dtype mismatch in Gemma4 audio feature merging, a bfloat16 positional embedding error in AudioFlamingo3, and missing backend requirement guards for Voxtral. The VibeVoice ASR processor was also updated to make audio input optional and support multiple audios per prompt.

feat[vLLM x v5]: Make audio optional and support multiple audios in VibeVoice ASR processor (#47483) by @harshaljanjani in [#47483]

Fix Gemma4 audio feature dtype mismatch in masked_scatter (#47482) by @danielhanchen in [#47482]

[fix] fix requirements audio feature and proc (#47113) by @eustlb in [#47113]

[AudioFlamingo3] Fix bfloat16 dtype mismatch in audio encoder positional embedding (#47258) by @snkii in [#47258]

Parallelization

Expanded FSDP support across 94 ForCausalLM model classes with auto-generated FSDP plans, added end-to-end FSDP tests including distributed checkpoint save/load and generation, and introduced a dedicated FSDP CI job. Additionally, fixed a device mismatch bug in create_bidirectional_sliding_window_mask under model parallelism and resolved a tensor parallel inference issue for models with tied embeddings.

skip fsdp tests when backend is mps (#47601) by @3outeille in [#47601]

Fix model parallel device mismatch in create_bidirectional_sliding_window_mask (#47560) by @abcgco in [#47560]

Add FSDP plans to all models (#47165) by @3outeille in [#47165]

Fix TP inference for tied embedding (#47503) by @3outeille in [#47503]

Add FSDP CI and end-to-end FSDP tests + save fsdp (#47357) by @3outeille in [#47357]

Tokenization

This release adds native support for Mistral's "tekken" tokenizer format via AutoTokenizer, fixes a CodeLlama tokenizer bug where leading whitespace was incorrectly dropped during decode, and patches a potential ReDoS vulnerability caused by unescaped tokenizer filenames being used as regex patterns in from_pretrained.

[Mistral] Add native tekken tokenizer support to AutoTokenizer (#47507) by @juliendenize in [#47507]

Fix CodeLlama tokenizer dropping leading whitespace on decode (#47488) by @SuryanshSS1011 in [#47488]

Fix potential ReDoS by escaping tokenizer filename used as regex pattern (#47498) by @hameedibrh in [#47498]

Serve

Improved the serve chat parsing to unify streaming and non-streaming paths under a single response parser that handles tool calls, reasoning, and content, simplifying the addition of new model support. Additionally, hardened daily CI reporting by fixing GitHub API diagnostic output being captured in Slack payloads and adding rate-limit resilience to prevent report failures when paginating large job matrices.

CI: Log GitHub API diagnostics to stderr (#47635) by @tarekziade in [#47635]

Update serve chat parsing (#46267) by @SunMarc in [#46267]

ci: harden daily CI reporting against GitHub API rate limits (#47382) by @tarekziade in [#47382]

Bugfixes and improvements

Fix cached_files silently returning stale file on read-only filesystem (EROFS) (#47852) by @ydshieh in [#47852]

Fix PhimoeIntegrationTest (#46539) by @ydshieh in [#46539]

make examples under doc device agnostic (#47812) by @kaixuanliu in [#47812]

cancel deterministic for XPU in gemma4 tests (#47790) by @kaixuanliu in [#47790]

Serialize post-mlinter-review after post-link to avoid PR description race (#47832) by @ydshieh in [#47832]

Use content hash for mlinter review deduplication (#47830) by @ydshieh in [#47830]

Add new args in auto-docstring (#47737) by @zucchini-nlp in [#47737]

Add check_model_inits.py (#47656) by @guarin in [#47656]

add xpu in installation guide (#47785) by @sywangyi in [#47785]

Fix mlinter review job: checkout before artifact download (#47820) by @ydshieh in [#47820]

Post mlinter findings as inline PR review comments (#47819) by @ydshieh in [#47819]

Fix ci style (#47818) by @vasqu in [#47818]

Hotfix axk2 indexer norm (#47810) by @kmswin1 in [#47810]

open fla support for XPU to benefit from the acceleration (#47799) by @kaixuanliu in [#47799]

[docs] Update BatchEncoding.to() type annotation and docstring (#47789) by @samyuktahegde in [#47789]

Fix linting (#47807) by @Cyrilvallez in [#47807]

Fix patching in some models (#47798) by @zucchini-nlp in [#47798]

Add post-mlinter-review job to post-dashboard-link workflow (#47800) by @ydshieh in [#47800]

Migrate torchao integration off deleted torchao.dtypes (#47797) by @vkuzo in [#47797]

[conftest] Also wrap snapshot_download for EROFS fallback (#47796) by @ydshieh in [#47796]

Fix MI355 CI: bump hf-workflows pin to NUM_SLICES=4 (#47792) by @Abdennacer-Badaoui in [#47792]

[Fix] Wrong type hint in get_number_of_image_patches (#47788) by @remi-or in [#47788]

Fix Dac offload tests (#47775) by @guarin in [#47775]

[Fix] Swapped height and width in KimiK25 (#47786) by @remi-or in [#47786]

Remove dangling files and folders (#47764) by @Cyrilvallez in [#47764]

Fix AI-written conversion mappings (#47755) by @Cyrilvallez in [#47755]

update mistral common version for PR 47507 (#47677) by @itazap in [#47677]

Fix spelling/grammar in model files (batch 3/3) (#47684) by @Rocketknight1 in [#47684]

Fix spelling/grammar in core library, examples, and utils (#47685) by @Rocketknight1 in [#47685]

Fix spelling/grammar in model files (batch 2/3) (#47683) by @Rocketknight1 in [#47683]

update sonicmoe versions (#47769) by @IlyasMoutawwakil in [#47769]

clean up reverse_op fixme in compressed_tensors (#47701) by @DhanushPillay in [#47701]

PR CI with torch 2.13 (#47767) by @ydshieh in [#47767]

Import utils compilation fixes (#47726) by @IlyasMoutawwakil in [#47726]

Fix DBRX MoE hidden size and expert GLU transposes (#47671) by @kaixuanliu in [#47671]

Fix multi token decode merging (#47762) by @IlyasMoutawwakil in [#47762]

Fix: Remove redundant @can_return_tuple conflicting with @capture_out… (#47733) by @guarin in [#47733]

fix processor config nested key fallback (#47628) by @YunzhuLu in [#47628]

[CI - Debug] Skip /transformers-dependent steps for CPU runner (#47759) by @ydshieh in [#47759]

[CI] Add CPU runner support to ssh-runner workflow (#47757) by @ydshieh in [#47757]

Simplify reverse weight conversion (#47725) by @Cyrilvallez in [#47725]

Remove useless linting for inv_freq (#47753) by @Cyrilvallez in [#47753]

Use explicit nn.Buffer everywhere for modular (#47722) by @Cyrilvallez in [#47722]

Remove stale and redundant _no_split_modules entries (#47645) by @guarin in [#47645]

Executorch exporter fixes (#47243) by @IlyasMoutawwakil in [#47243]

Acc fix in xpu (#47500) by @sywangyi in [#47500]

update Dockerfile for xpu torch2.13 (#47502) by @sywangyi in [#47502]

feat[vLLM]: Support text replacement offsets in the remaining old-format processors (#47614) by @harshaljanjani in [#47614]

Fix ImportError in transformers.exporters on torch < 2.8 (#47711) by @Neal006 in [#47711]

Fix failing tests for axk1 and axk2 (#47727) by @kaixuanliu in [#47727]

Fix failing tests for granite_swa and granitemoe_swa (#47723) by @kaixuanliu in [#47723]

skip invalid test cases for inkling tests (#47493) by @kaixuanliu in [#47493]

[docs] Fix BatchEncoding documentation inconsistencies (#47647) by @samyuktahegde in [#47647]

Fix spelling/grammar in model files (batch 1/3) (#47682) by @Rocketknight1 in [#47682]

Fix spelling/grammar in English docs (q → z) (#47681) by @Rocketknight1 in [#47681]

Fix spelling/grammar in English docs (h → p) (#47680) by @Rocketknight1 in [#47680]

Fix spelling/grammar in English docs (a → g) (#47679) by @Rocketknight1 in [#47679]

fix npu check (#47587) by @DhanushPillay in [#47587]

fix: correct text input validation logic in 8 multimodal processors (and → or) (#47663) by @AbdullahRasheed45 in [#47663]

Fix feature dtype mismatch in masked_scatter for seven multimodal models (#47673) by @ in [#47673]

[docs] storing and loading chat templates (#47650) by @stevhliu in [#47650]

adding amd quark config class changes (#47322) by @debasisdwivedy in [#47322]

Fix compressed tensors impl (#47652) by @SunMarc in [#47652]

Hoist special-token lookups in wav2vec2 decode paths and drop a dead filter in wav2vec2_phoneme (#47557) by @ishan-1010 in [#47557]

silencing elastic warning by import distributed lib inside functions (#47665) by @3outeille in [#47665]

Fix missing github_utils.py download in PR CI dashboard workflow (#47668) by @ydshieh in [#47668]

Exportable kimi (#47096) by @IlyasMoutawwakil in [#47096]

better guarding to handle torch compiled with USE_DISTRIBUTED=0 (#47619) by @3outeille in [#47619]

Remove gemma4 warnings (#47664) by @Cyrilvallez in [#47664]

[Chat Parsing] Type inline tool-call arguments from the calling tool's JSON Schema (#47529) by @yonigozlan in [#47529]

Improve Trainer DataLoader Controls for Streaming and Multiprocessing (#47164) by @muyihao in [#47164]

Update maintainer list (#47644) by @Rocketknight1 in [#47644]

Allow position_ids_start=2 on DataCollatorWithFlattening for RoBERTa etc. (#47525) by @tomaarsen in [#47525]

Remove Rotary warning (#47642) by @Cyrilvallez in [#47642]

Drop multimodal inputs natively in prepare_inputs_for_generation if not in prefill (#47622) by @Cyrilvallez in [#47622]

Simplify all Rotary modules (#47598) by @Cyrilvallez in [#47598]

[docs] response_template when serving (#47626) by @stevhliu in [#47626]

[docs] MTP support (#47301) by @stevhliu in [#47301]

Fix GPT-2 c_proj depth scaling initialization (#47459) by @DavidJohnQuinlan in [#47459]

Fix fp8_linear compilability (#47623) by @IlyasMoutawwakil in [#47623]

byebye torch 2.4 (#47609) by @ydshieh in [#47609]

Vectorize NoRepeatNGramLogitsProcessor and remove its host sync (#47571) by @hameedibrh in [#47571]

Fix CUDA Graph breaking host to device copy from scalar tensor allocation (#47547) by @hmellor in [#47547]

Fix some processors (#47608) by @zucchini-nlp in [#47608]

CI: Add serge review relay workflow and review rules (#47610) by @tarekziade in [#47610]

[docs] Exporters (#47374) by @stevhliu in [#47374]

CI: use a single function for GH calls (#47474) by @tarekziade in [#47474]

Remove redundant guarding for distributed (#47570) by @3outeille in [#47570]

Fix modular for mamba packages (#47494) by @Cyrilvallez in [#47494]

Remove deprecated conversion in Kimi (#47581) by @zucchini-nlp in [#47581]

CI: use transformers-ci daily workflow with OTEL (#47360) by @tarekziade in [#47360]

Fix slow tensor path in _check_special_mm_tokens (#47580) by @guan404ming in [#47580]

CI: let's run integration failure cron at 10pm (#47537) by @tarekziade in [#47537]

Better and more extensive tests for RoPE (#46912) by @zucchini-nlp in [#46912]

Fix mamba2 family decode and simplify all reshape ops (#47569) by @Cyrilvallez in [#47569]

[DiffusionGemma] Cast the decoder padding mask to bool (#47295) by @kashif in [#47295]

General maintenance (#47517) by @zucchini-nlp in [#47517]

fixed the benchmark script with DistributedConfig (#47568) by @tarekziade in [#47568]

Fix failing tests for mimo_v2_flash (#47284) by @kaixuanliu in [#47284]

Make tokenization_mistral_common importable without mistral_common installed (#47397) by @juliendenize in [#47397]

Use --flake-runs=1 in check_bad_commit.py for PR comment CI (#47522) by @ydshieh in [#47522]

Deprecate the old response_schema (#47320) by @Rocketknight1 in [#47320]

fix: add pickle support to _LazyConfigMapping for spawn multiprocessing (#46026) by @kfojcik-intel in [#46026]

Delete old deprecations (#47518) by @zucchini-nlp in [#47518]

Run only @slow tests in PR comment CI (#47521) by @ydshieh in [#47521]

Fix incorrect type hint (#47519) by @hmellor in [#47519]

Deprecate CB config in gen configuration (#47291) by @remi-or in [#47291]

[Offloading] [Bugfix] Fix fully offloaded model saving (#47336) by @kylesayrs in [#47336]

CI: fix torchaudio pinning +proper break in rnnt (#47422) by @tarekziade in [#47422]

Fix Qwen2.5-Omni Token2Wav DiT rotary embedding layout (interleaved cos/sin) (#47403) by @HenryVarro666 in [#47403]

CI: add reproduce mode to serge verify caller (#47492) by @tarekziade in [#47492]

[fix][whisper]: fix max_new_tokens handling (#46795) by @eustlb in [#46795]

Isolate MLA KV expansion to make it easier to bypass (#47460) by @hmellor in [#47460]

Fix loss alignment and Trainer token counting for encoder decoder models (#46903) by @OmkumarSolanki in [#46903]

Fix the HunyuanVL's torchvision backend (#47499) by @Mi-Jiazhi in [#47499]

[DiffusionGemma] Support gradient checkpointing (#46572) by @kashif in [#46572]

Fix failing tests for zaya (#47268) by @kaixuanliu in [#47268]

tipsv2_dpt: fix failing tests for XPU (#47292) by @kaixuanliu in [#47292]

Fix some failed test cases related with XPU Expectations (#47173) by @kaixuanliu in [#47173]

fix: guard DTensor import in sharding_utils.py for PyTorch < 2.5 (#47481) by @ in [#47481]

CPU can incur a slow path on non-contiguous magnitudes (#47351) by @vbayanag in [#47351]

Route chat management calls to the service root (#47138) (#47303) by @dhruv7477 in [#47303]

fix: liger unnecessarily materializes logits in VRAM during eval, causing OOM (#45273) by @excepshenal in [#45273]

Fix gradient inflation when combining label smoothing with gradient accumulation (#47261) by @Incheonkirin in [#47261]

Fix MoE expert decompression for non-32-divisible bit widths (#47315) by @KKothuri in [#47315]

[Qwen3ASR] Add hotword parsing, and fix language parsing and training. (#47111) by @ebezzam in [#47111]

Remove deprecated training args and is_fast property (#46917) by @cyyever in [#46917]

Consistent output shape from get_image_features (#46405) by @zucchini-nlp in [#46405]

fix failed test cases for qwen3_omni_moe model (#47449) by @kaixuanliu in [#47449]

Fix double-shifted training loss in GitForCausalLM (#47395) by @ in [#47395]

Fix CohereASR training-loss double-shift (same as Moonshine fix #46784) (#46895) by @sharmax-vikas in [#46895]

Warn when group_by_length is silently ignored for iterable datasets (#47379) by @qgallouedec in [#47379]

Update bug report list (#46607) by @molbap in [#46607]

fix: remove unreachable return in special token builder (#47420) by @hai1222 in [#47420]

Add Harry to slow CI (#47454) by @vasqu in [#47454]

BLT: vectorize patch length processing (#47385) by @sj0618 in [#47385]

Fix TrackioCallback fails to log evaluation metrics after training ends (#46935) by @lewtun in [#46935]

[Kimi] add integration tests (#47383) by @zucchini-nlp in [#47383]

Add distributed runtime utils and DistributedMixin (#47352) by @3outeille in [#47352]

Fix Cosmos 3 Edge Patch packing order (#47399) by @atharvajoshi10 in [#47399]

Fix yarn mscale_all_dim for DeepSeek v2 and Mistral 4 (#47435) by @hmellor in [#47435]

Hoist special-token lookups out of per-token loops in six slow tokenizers (#47425) by @ishan-1010 in [#47425]

fix typos and variable naming in quicktour.md (#47418) by @yashasvi-srivastava21 in [#47418]

Normalize multimodal input keys in AnyToAnyPipeline (#47074) by @Sunt-ing in [#47074]

Fix Aria checkpoint key conversion mapping (#47151) by @sywangyi in [#47151]

extend tests/models/qwen3_next/test_modeling_qwen3_next.py::Qwen3Next… (#47184) by @sywangyi in [#47184]

CI: add serge verify (GPU) caller workflow (#47381) by @tarekziade in [#47381]

Fix orthogonal_ init for low-precision dtypes (bf16/fp16) (#47252) by @janbernloehr in [#47252]

[docs] Fix decode examples and expected output in fast_tokenizers (#47369) by @samyuktahegde in [#47369]

[serge] Fix 20 integration tests for model whisper failing with output_mismatch (list output differs (10), other (6) (#47150) by @sergereview[bot] in [#47150]

[Mistral] Move MistralConverter into integrations/mistral/ package (#46603) by @juliendenize in [#46603]

Fix GLM video frame padding for temporal patches (#47141) by @labAxiaoming in [#47141]

[docs] Inkling (#47350) by @stevhliu in [#47350]

Fix Daily CI reporting issues (#47364) by @tarekziade in [#47364]

[peft] Support key_mapping with PEFT models (#46766) by @tomaarsen in [#46766]

Fix model tests for tipsv2 (#47356) by @kaixuanliu in [#47356]

Fix TimesFM 2.5 window_size AttributeError (#47363) by @kashif in [#47363]

fix: allow num_labels property to return None when id2label is unset (#47069) by @SebTardif in [#47069]

[Tests] Fix slow video tensor creation from list of numpy arrays in SmolVLM (#44731) by @Defalt-Meh in [#44731]

Update dev version on main (#47366) by @vasqu in [#47366]

Fix deepgemm on multiple devices (#47323) by @IlyasMoutawwakil in [#47323]

ci: Handle empty GitHub token in CI run lookup (#47362) by @tarekziade in [#47362]

Fix inkling feature extractor (#47349) by @ArthurZucker in [#47349]

Significant community contributions

The following contributors have made significant changes to the library over the last release:

@ydshieh

Fix cached_files silently returning stale file on read-only filesystem (EROFS) (#47852)

Fix PhimoeIntegrationTest (#46539)

Serialize post-mlinter-review after post-link to avoid PR description race (#47832)

Use content hash for mlinter review deduplication (#47830)

Fix mlinter review job: checkout before artifact download (#47820)

Post mlinter findings as inline PR review comments (#47819)

Add post-mlinter-review job to post-dashboard-link workflow (#47800)

[conftest] Also wrap snapshot_download for EROFS fallback (#47796)

[conftest] Fix EROFS fallback for kernel downloads (correct interception point) (#47794)

[conftest] Fix EROFS fallback for kernel downloads via HfApi (#47791)

PR CI with torch 2.13 (#47767)

[CI - Debug] Skip /transformers-dependent steps for CPU runner (#47759)

[CI] Add CPU runner support to ssh-runner workflow (#47757)

Update daily CI Docker image to torch 2.13.0 / CUDA 13.0 (#47738)

Fix missing github_utils.py download in PR CI dashboard workflow (#47668)

byebye torch 2.4 (#47609)

Use --flake-runs=1 in check_bad_commit.py for PR comment CI (#47522)

Run only @slow tests in PR comment CI (#47521)

@kaixuanliu

make examples under doc device agnostic (#47812)

cancel deterministic for XPU in gemma4 tests (#47790)

open fla support for XPU to benefit from the acceleration (#47799)

Fix DBRX MoE hidden size and expert GLU transposes (#47671)

Fix failing tests for axk1 and axk2 (#47727)

Fix failing tests for granite_swa and granitemoe_swa (#47723)

skip invalid test cases for inkling tests (#47493)

Fix failing tests for mimo_v2_flash (#47284)

Fix failing tests for zaya (#47268)

tipsv2_dpt: fix failing tests for XPU (#47292)

Fix some failed test cases related with XPU Expectations (#47173)

add paged attention tests support for XPU (#47163)

Fix multi-device mxfp4 dequantization race in _convert_moe_packed_tensors (#47423)

fix failed test cases for qwen3_omni_moe model (#47449)

Fix model tests for tipsv2 (#47356)

@vasqu

Fix ci style (#47818)

🚨 [Kernels] Refactor all linear attn models & native kernels fallback (#47630)

[Kernels] Refactor function handling (#46883)

Add AXK2 from SKT (#47528)

Add Harry to slow CI (#47454)

Update dev version on main (#47366)

[FP8] Bump kernels version (#47344)

@kmswin1

Hotfix axk2 indexer norm (#47810)

Fix A.X-K2 fp8 modules_to_not_convert normalization for the gated-norm MLP (#47578)

add_axk1 (#46867)

@remi-or

[Fix] Fix multi-head latent attention (MLA) (#47761)

[Fix] Wrong type hint in get_number_of_image_patches (#47788)

[Fix] Swapped height and width in KimiK25 (#47786)

Deprecate CB config in gen configuration (#47291)

[CB] Automatically switch attention implementation to flash (#47330)

@juliendenize

[Mistral] Add native tekken tokenizer support to AutoTokenizer (#47507)

Make tokenization_mistral_common importable without mistral_common installed (#47397)

[Mistral] Move MistralConverter into integrations/mistral/ package (#46603)

@IMvision12

Add support for batched Qwen2.5/3-Omni audio generation (#47186)

Fix Hunyuan-VL PIL image resize parity with reference preprocessing (#47233)

@jiqing-feng

🚨 Enable SDPA (and other attention backends) for T5 and propagate to the T5 family (#47014)

Fix shape mismatch in KyutaiSpeechToText generate() last window (#46952)

Fix typo in MusicgenForCausalLM.generate() (#46974)

Fix BlockMask crash in CTRL flex-attention generation (#46854)

@tarekziade

CI: Log GitHub API diagnostics to stderr (#47635)

CI: Add serge review relay workflow and review rules (#47610)

CI: use a single function for GH calls (#47474)

CI: use transformers-ci daily workflow with OTEL (#47360)

CI: let's run integration failure cron at 10pm (#47537)

fixed the benchmark script with DistributedConfig (#47568)

CI: fix torchaudio pinning +proper break in rnnt (#47422)

CI: add reproduce mode to serge verify caller (#47492)

CI: add serge verify (GPU) caller workflow (#47381)

ci: harden daily CI reporting against GitHub API rate limits (#47382)

Fix Daily CI reporting issues (#47364)

ci: Handle empty GitHub token in CI run lookup (#47362)

@daviswer

Add Granite-swa and Granitemoe-swa model support (#47179)

@ShareLer

Optimize flash attention max seqlen computation in vision attention (#47170)

@atharvajoshi10

Fix Cosmos 3 Edge Patch packing order (#47399)

Add Cosmos3 Edge model support (#47181)

Original source

The Hugging Face MCP Server has been updated to provide enhanced connectivity to the Hub with fewer tools and tokens.

The main change is the new hf_fs tool which provides a single interface to repositories, storage, documentation, papers and more. It's equipped with search and lets your assistant naturally navigate Hugging Face in just over 1,000 tokens.

Add Sandboxes to give your assistant secure execution environments attached to buckets and repositories. This gives blazing fast code access for dataset analysis, model training, Space creation and much more.

Share your feedback on MCP here and update your settings at https://huggingface.co/settings/mcp

Original source

Patch release v5.14.1

This patch solves a few issues which appeared when integrating Inkling model, most notably an issue affecting models using EncoderDecoderCache during assisted generation. It also fixes an issue that could appear during prefill with StaticCache and sdpa without padding for Inkling which uses a position_bias.

It contains the following commits:

• Fix sdpa prefill with position_bias (#47359) by @Cyrilvallez
• Fix assisted decoding for models with EncoderDecoder cache & OlmoHybrid (#47361) by @Cyrilvallez
• [FP8] Bump kernels version (#47344) by @vasqu
• Fix deepgemm on multiple devices (#47323) by @IlyasMoutawwakil

Original source

Release v5.14.0

New Model additions

Inkling (fresh from Thinking Machines): 975B total, 41B active

Add Inkling model #47347 by @molbap @Cyrilvallez @eustlb and @zucchini-nlp

Inkling is a general-purpose multimodal model that accepts text, image and audio inputs and generates text outputs. It is intended for use in English and other languages, and across multiple coding languages. The model is designed to be used by developers building AI-powered applications, including agentic and tool-use systems, coding assistants, chatbots, and retrieval-augmented generation systems, and is suitable for general-purpose conversational use, instruction-following, and other natural language and multimodal tasks. It is released with open weights to support research, fine-tuning and integration into third-party products by downstream developers.

TIPSv2

Links: Documentation

Add TIPSv2 (#46347) by @Ternura143 in #46347

TIPSv2 DPT

Links: Documentation

Add TIPSv2 (#46347) by @Ternura143 in #46347

🚨 Breaking changes

GPTNeoX now remaps embed_out to lm_head and GPTBigCode has _supports_attention_backend = True enabled for vLLM compatibility; users relying on the previous weight naming or attention backend behavior for these models should update their code accordingly.

🚨 Fix GPTBigCode and GPTNeoX for the Transformers modelling backend for vLLM (#47198) by @hmellor

Kernels

Several kernel-related fixes and improvements were made, including pinning the kernels dependency to a compatible version in the benchmark workflow, removing a deprecated package_name argument from LocalLayerRepository, and making the DeepGEMM Triton fallback more robust when CUDA_HOME is unset or misconfigured. Additionally, SDPA prefill was updated to leverage the FlashAttention kernel with StaticCache, yielding significant performance gains (up to 260% faster for large input sizes).

Pin kernels to compatible version in benchmark workflow (#47339) by @tarekziade in [#47339]

[Fix] Remove deprecated argument from kernels call (#47100) by @remi-or in [#47100]

[Fix] Make DeepGEMM triton fallback more robust (#47126) by @remi-or in [#47126]

[sdpa] Allow prefill to use FA kernel with StaticCache (#47094) by @Cyrilvallez in [#47094]

Generation

Generation improvements include adding Multi-Token Prediction (MTP) decoding support, static ensemble verification for speculative decoding to improve draft token acceptance rates, and a fix for crashes in greedy assisted generation with different tokenizers. A misleading double-negative warning message for synced_gpus in continuous batching mode was also corrected.

[generation] Fix misleading synced_gpus warning in continuous batching (#47158) by @Partha-Shankar in [#47158]

[generate] Add proper MTP support (#46229) by @Cyrilvallez in [#46229]

Fix crash in greedy assisted generation with different tokenizers (#46936) by @Sunt-ing in [#46936]

[Generation] Add static ensemble verification for lossy speculative decoding (#45979) by @kasakh in [#45979]

Performance

Fixed a Flash Attention performance regression affecting models like Qwen3-VL and resolved a MoE decode optimization bug where the grouped-to-batched matrix multiplication switch was not applied to experts residing in submodels (e.g., VLMs with a nested text config).

Fix FA performance regression (#47134) by @andreasgoulas in [#47134]

Fix MoE decode optimization for experts living in a submodel (#47107) by @IlyasMoutawwakil in [#47107]

Make doc builds faster (#47099) by @mishig25 in [#47099]

Cache

Cache dispatch logic was simplified by introducing explicit layer-type mappings for sliding and static layers, reducing complexity in cache routing. Additionally, fixes were made for read-only cache failures in CPU CI environments and for MPS graph cache growth during variable-length batch training on Apple Silicon.

Fix CI read-only cache failures by patching cached_files in conftest (#47043) by @ydshieh in [#47043]

trainer: clear MPS graph cache via torch_empty_cache_steps (#45818) by @anagnorisis2peripeteia in [#45818]

[cache] Simplify cache dispatch based on layer_types (#47118) by @Cyrilvallez in [#47118]

Bugfixes and improvements

ci: cover xet as well (runtime error) (#47338) by @tarekziade in [#47338]

[docs] TokenizersBackend fallback (#47302) by @stevhliu in [#47302]

Resolve continuous batching XPU availability checks at runtime (#47185) by @kaixuanliu in [#47185]

[Nit] Add kernels_fallback_ok kwarg to is_flash_attn_N_available (#47318) by @remi-or in [#47318]

[Nit] Add expectations for gemma4 tests on H100 (#47311) by @remi-or in [#47311]

[docs] DeepGEMM requirements (#47324) by @stevhliu in [#47324]

DeepGEMM shouldn't pad on SM90 (#47313) by @IlyasMoutawwakil in [#47313]

Fix half-precision torch.compile crash in DETR-family sine position embeddings (#47238) by @David-Wu1119 in [#47238]

Fix hardcoded paths in siglip checkpoint/vocab loading (#47178) by @XanxusCrypto in [#47178]

Update AMD CI runner groups to amd-mi300 (#47307) by @Abdennacer-Badaoui in [#47307]

Point to Gemma 4 model in Gemma4ForCausalLM docstring example (#47255) by @lefft in [#47255]

Fix Qwen Omni batched text postprocessing (#47197) by @Sunt-ing in [#47197]

Fix AqlmConfig error messages to say "int" instead of "float" (#47089) by @Sreekant13 in [#47089]

Fix check for interactive stdout in _style function (#47283) by @smart8986 in [#47283]

Fix get_json_schema crash on non-string docstring choices (#47072) by @Sreekant13 in [#47072]

Make MODEL_IDS_TO_TOKENIZERS_BACKEND capture all DeepSeek R1 distills (#47296) by @hmellor in [#47296]

Update doc preprocessing regex to prevent ReDoS (#47187) by @WilliamRoyNelson in [#47187]

Shard on read Dtensor aware (#46717) by @3outeille in [#46717]

Switch AMD daily CI to mi300 runners (#47259) by @Abdennacer-Badaoui in [#47259]

tests: reduce processor test memory usage by using tiny Hub checkpoints (#47213) by @ydshieh in [#47213]

Torch compile backend defaults to "neuron" (#47035) by @michaelbenayoun in [#47035]

Fix flash-attn Docker build broken by setuptools 83 removing pkg_resources (#47251) by @ydshieh in [#47251]

Add heterogeneous config support (per-layer configuration) (#45333) by @eladsegal in [#45333]

[fix] update integration test values (#47146) by @eustlb in [#47146]

Fix DeepSpeed SP loss aggregation and LocalLayerRepository kwargs (#47073) by @sshivampeta in [#47073]

tests only for the top 10 download models (#47244) by @3outeille in [#47244]

Fix InputTokensDetails missing cache_write_tokens for openai>=2.34.0 (#47248) by @ydshieh in [#47248]

Revert "Trigger a scheduled run" (#47249) by @ydshieh in [#47249]

Remove Executorch from CI until latest version is supported and fully tested on CI env (#47242) by @IlyasMoutawwakil in [#47242]

Be more defensive with remap_legacy_layer_types for custom models (#47245) by @hmellor in [#47245]

Fix DistributedConfig docstring for unimplemented sp_plan (#47237) by @3outeille in [#47237]

Switch mlinter to 0.1.2 (#47172) by @tarekziade in [#47172]

Trigger a scheduled run (#47209) by @ydshieh in [#47209]

Make executorch exporter tests always use xnnpack backend (#47201) by @tarekziade in [#47201]

No agent PR descriptions (#45790) by @Rocketknight1 in [#45790]

Clarify that max_steps is required for datasets without len (#47155) by @albertvillanova in [#47155]

Cleanup pipelines, stop materializing generators (#47142) by @Rocketknight1 in [#47142]

Fix device_map computation when the no_split_modules have different sizes (#47203) by @Cyrilvallez in [#47203]

Add native FSDP2 module + migration (#46707) by @3outeille in [#46707]

Fix experts implementation in two spots (#47097) by @remi-or in [#47097]

[Fix] Remove old automatic cross attn pattern from output recorders (#47117) by @remi-or in [#47117]

🌐 [i18n-KO] Translate accelerator_selection.md to Korean (#47157) by @kkwjk2718 in [#47157]

[i18n-KO] Translate optimum.md to Korean and fix Furiosa typo (#47156) by @kkwjk2718 in [#47156]

[docs] fix curly quotes rendering to straight quotes (#47135) by @clijo in [#47135]

Fix custom code which doesn't know about the new linear layer type names (#47174) by @hmellor in [#47174]

Reject path traversal in the transformers_weights config field (#46890) by @LinZiyuu in [#46890]

[docs] Custom code conversion mapping (#47114) by @stevhliu in [#47114]

Add exporters min version requirements and test skip (#47161) by @IlyasMoutawwakil in [#47161]

tests: reduce processor test memory usage and use tiny test assets (#47168) by @ydshieh in [#47168]

Clarify input device placement in the Quicktour inference example (#47136) by @samyuktahegde in [#47136]

Extend continuous batching memory prediction test to XPU (#47159) by @sywangyi in [#47159]

Fix case where _LazyAutoMapping.register is passed a str key (#47148) by @hmellor in [#47148]

[docs] MoE decode switching (#47149) by @stevhliu in [#47149]

add XPU output expectations for minicpm3 tests (#47092) by @kaixuanliu in [#47092]

Diffusion gemma: fix failed test cases (#47025) by @kaixuanliu in [#47025]

add XPU Expectation for cosmos3_omni tests (#46880) by @kaixuanliu in [#46880]

Fix IndexError Bug in XLMRoberta/Camembert ForMultipleChoice by restoring the pooler (#47147) by @pariidanDKE in [#47147]

Skip caching_allocator_warmup on Neuron (no reuse pool to warm; currently OOMs) (#47029) by @dacorvo in [#47029]

[docs] continuous batching (offloading behavior, max batch tokens, block size minimum) (#46925) by @stevhliu in [#46925]

[docs] fix autolinks (#46968) by @stevhliu in [#46968]

revert #47121 (#47144) by @eustlb in [#47144]

Fix output labels for AudioFlamingo3 (and related) models (#47112) by @ebezzam in [#47112]

Fix false len claims in Trainer docstrings (#47131) by @albertvillanova in [#47131]

processor tests: use tiny Hub repos to reduce CI memory (#47115) by @ydshieh in [#47115]

[serge] Fix 12 integration tests for model dac failing with output_mismatch (tensor values differ (6), other (6)) (#47121) by @sergereview[bot] in [#47121]

Fix CLI compatibility with huggingface_hub 1.22 (#47059) (#47064) by @dhruv7477 in [#47064]

we want to run the CI in the release branches (#47125) by @tarekziade in [#47125]

Small improvement (#47128) by @Cyrilvallez in [#47128]

[Model] Support use_cache=False for DeepSeek V4 (#46965) by @kylesayrs in [#46965]

docs-fix: IMDb dataset link in sequence classification guide (#47062) by @abhishekkapoorx in [#47062]

Fix AltCLIP text embedding resize test (#47079) by @IMvision12 in [#47079]

fix mask return-type contract regression and add correctness guard for (#47019) by @kaixuanliu in [#47019]

Fix save_pretrained with offloading and weight conversions (#47018) by @Cyrilvallez in [#47018]

Update dev (#47044) by @vasqu in [#47044]

[Gemma4] Update 1 integration test (#47042) by @vasqu in [#47042]

Significant community contributions

The following contributors have made significant changes to the library over the last release:

@ArthurZucker

v5.14.0

@tarekziade

ci: cover xet as well (runtime error) (#47338)

Pin kernels to compatible version in benchmark workflow (#47339)

Switch mlinter to 0.1.2 (#47172)

Make executorch exporter tests always use xnnpack backend (#47201)

Remove executorch from all-latest-gpu image + add torch smoke test (#47196)

we want to run the CI in the release branches (#47125)

@remi-or

[Nit] Add kernels_fallback_ok kwarg to is_flash_attn_N_available (#47318)

[Nit] Add expectations for gemma4 tests on H100 (#47311)

[Fix] Remove deprecated argument from kernels call (#47100)

[Fix] Make DeepGEMM triton fallback more robust (#47126)

Fix experts implementation in two spots (#47097)

[Fix] Remove old automatic cross attn pattern from output recorders (#47117)

@ydshieh

tests: reduce processor test memory usage by using tiny Hub checkpoints (#47213)

Fix flash-attn Docker build broken by setuptools 83 removing pkg_resources (#47251)

Fix InputTokensDetails missing cache_write_tokens for openai>=2.34.0 (#47248)

Revert "Trigger a scheduled run" (#47249)

Fix CI read-only cache failures by patching cached_files in conftest (#47043)

Trigger a scheduled run (#47209)

tests: reduce processor test memory usage and use tiny test assets (#47168)

processor tests: use tiny Hub repos to reduce CI memory (#47115)

@eladsegal

Add heterogeneous config support (per-layer configuration) (#45333)

@eustlb

[fix] update integration test values (#47146)

revert #47121 (#47144)

@Ternura143

Add TIPSv2 (#46347)

Original source

Fine-Grained Token Presets

You can now create fine-grained access tokens from presets: Read-Only, Inference, Write, CI/CD, or Full Access. Pick a preset to get a curated set of permissions with a clear summary of what the token can do, attach organizations in one click, or switch to Custom to fine-tune individual permissions as before.

Presets are also linkable: URLs like /settings/tokens/new?preset=inference or /settings/tokens/new?preset=read-only&orgs=huggingface take you straight to a pre-configured token, handy for docs and onboarding guides.

Original source