---
name: timm (pytorch-image-models)
slug: timm-pytorch-image-models
type: github
source_url: https://github.com/huggingface/pytorch-image-models
organization: Hugging Face
organization_slug: hugging-face
total_releases: 81
latest_version: v1.0.31
latest_date: 2026-11-03
last_updated: 2026-11-15
tracking_since: 2019-05-31
canonical: https://releases.sh/hugging-face/timm-pytorch-image-models
organization_url: https://releases.sh/hugging-face
---

<Summary type="rolling" window-days="90" release-count="2">
timm shifted focus toward production robustness and distributed training compatibility. Security hardening of checkpoint loading now defaults to `weights_only=True`, while attention handling across ViT and EVA models received refinement for masked and causal tasks. The library also tuned optimizers (Muon, AdamP, SGDP) for distributed scenarios like FSDP2 and DTensor, and removed torch.jit usage ahead of its official deprecation. A breaking change in `ParallelScalingBlock` QKV bias handling landed in v1.0.25, though it doesn't affect released model weights.
</Summary>

<Summary type="monthly" period="March 2026" release-count="1">
Focused on robustness and efficiency across core vision transformers. Tightened pickle loading security with `weights_only=True` by default and overhauled attention mask handling for ViT/EVA models to properly resolve boolean masks and propagate causal flags for self-supervised tasks. Added Patch Representation Refinement as a pooling option and optimized Hiera's scaled dot-product attention to unlock Flash Attention kernels.
</Summary>

<Release version="v1.0.31" date="November 3, 2026" published="2026-11-03T19:22:10.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.31">
## Release v1.0.31

## November 3, 2026

* `reset_classifier()` and checkpoint reconstruction: fixes across ViT, EVA, BEiT, NaFlexViT, RepViT and iFormer
  preserve normalization placement and distillation settings when exporting / reloading models after changing the
  classifier or pooling. Also fix unpooled PiT export and EfficientFormerV2 with distillation disabled.
* NaFlex: fix ViT / EVA norm placement during conversion, support SwiGLU alignment for DINOv3 / Lingbot checkpoints,
  and preserve grayscale / non-RGB input channels in data loaders.
* Fix layer-wise LR decay grouping for RepViT, HGNet, MViTv2 and pre-norm ViTs, and InceptionNeXt feature sizes with
  `output_stride=8/16`.
* Release 1.0.31.

## October 24, 2026

* `forward_intermediates()` and `prune_intermediate_layers()` are now supported by every model. Newly added for CSPNet, NFNet, DLA, DenseNet (#2870, thanks [ghaith-mq](https://github.com/ghaith-mq)), CoaT, ConViT, ConvMixer, CrossViT, HRNet, Inception-v3, Inception-ResNet-v2, NASNet, PNASNet, SelecSLS, SENet, Sequencer2d, VGG, Visformer, Xception and Aligned Xception, with `prune_intermediate_layers()` for NaFlexViT and MobileNetV5 encoders. CoaT, ConViT, ConvMixer, CrossViT and Visformer gain `features_only` support.
* `prune_intermediate_layers()` fixes
  * `prune_head=True` removes the whole head (head norm, pre-logits layers, attention pooling) so the pruned model leaves no unused params (DDP w/o `find_unused_parameters`), except CoaT, CrossViT and HRNet when only some of their parallel scales / branches are taken. Pass `prune_head=False` to keep it.
  * Pruning a hierarchical model to an earlier stage removes the final norm and head sized for the last stage, `forward_features()` ends at the last kept feature and `num_features` / `head_hidden_size` match it. Pruning to the stem feature removes the stem pooling after it.
  * Fixes for grad checkpointing, `stop_early`, class token / post network removal, and TorchScript of pruned / `feature_cls='getter'` models across many families.
* `head_hidden_size` is the actual pre-logits (pooled) width in all models, including concat pooling (`catavgmax`) and after `reset_classifier()`. Pool types a model can't apply now raise at creation / reset instead of being silently ignored or failing in forward.
* Other fixes: VOLO `num_classes=0` output is the class token embedding (matching `forward_head(pre_logits=True)`), `vit_pe_core` `reset_classifier()` w/ a non `map` pool removes the attention pool, pool type export via `save_for_hf` reads the head's current pooling, and headless (`num_classes=0`) Swin, SwinV2, DaViT and NextViT checkpoints reload correctly.

## October 13, 2026

* Hyperparameter sweeps via a new `sweep.py`: grid / random search over any `train.py` argument, one effective `--config` per job, run on a pool of local GPUs and resumable. WSD sweeps train one stable trunk per hyperparameter combination and branch independent cooldowns from its milestone checkpoints, comparing annealing at several budgets for a fraction of the cost of full runs. See the [training docs](https://huggingface.co/docs/timm/training_script#hyperparameter-sweeps).
* `train.py`: `--resume auto` resumes in place from the experiment's own `last.pth.tar` (checkpoint history, `model_best` and `summary.csv` stay consistent), `--stop-epoch N` stops without changing the schedule, `--save-epochs N [N ...]` keeps `milestone-N.pth.tar` snapshots, and YAML `--config` values are validated against the script's arguments.
* Deprecate `PlateauLRScheduler` (`--sched plateau`). `timm` schedulers are stateless functions of the epoch or update count, the metric-driven plateau schedule never fit that and its state is not checkpointed, so resume was never exact. Use `cosine` or `wsd`.
* Docs: new [Model API](https://huggingface.co/docs/timm/model_api) page, generated model family pages for every model module replacing the old per-model pages, and expanded data / loss / models reference pages.
* Model API consistency fixes: `reset_classifier()` keeps the current pooling in all models, unknown constructor args raise `TypeError` everywhere, Swin family `dynamic_img_size`, LeViT layer-wise lr decay, `no_weight_decay` names that matched no params (NesT, SwinV2, XCiT, FocalNet), resize state recorded for all `set_input_size` families.
* Fixes: GCViT global query mixed images within a batch, gradient checkpointing for RDNet and MobileNetV5, `save_for_hf` stale test input size, EfficientViT (MSRA) rectangular resize, ImageDataset image conversion errors (#2846). `reparameterize_model` moved to `timm.models`.

## October 9, 2026

* Add [CSPNeXt](https://arxiv.org/abs/2212.07784) (RTMDet backbone) `cspnext_{tiny,s,m,l,x}` with ImageNet-1k classification weights converted from OpenMMLab. Thanks [Munehiro Kobayashi](https://github.com/munehiro-k).
* Optimizer improvements
  * Foreach (multi-tensor) impls for RMSpropTF, LaProp, LAMB, LARS, AdaBelief. Foreach steps are chunked by device L2 cache size (`TIMM_FOREACH_CHUNK_SIZE` to override).
  * RMS update clipping (`clipping_threshold`) for AdamWLegacy / NAdamW, new decoupled / corrected weight decay NovoGrad variants (`novogradw`, `novogradc`).
  * Correctness fixes across many optimizers, including corrected weight decay w/ in-place updated tensor lr, FP16 update stalls at small lr, low precision state handling (Kron, AdaBelief, NovoGrad), and no more in-place `p.grad`
modification.
  * `torch.compile` / CUDA graph fixes for Adopt, SGDW, AdafactorBV.
  * New optimizer docs covering usage, naming conventions, and families. 
* Add WSD (warmup-stable-decay) LR scheduler (`--sched wsd`, `--lr-decay-fn`, `--lr-stable-timescale`).
* Model EMA fixes for bfloat16 / float16 models. ModelEmaV3 bfloat16 EMAs use stochastic rounding by default, float16 models keep a float32 EMA.
* Model resizing via `set_input_size()`
  * Expanded support for windowed models (MaxViT, CoAtNet, BEiT, EfficientViT-MSRA, relpos ViT, Swin family).
  * Swin / SwinV2 / SwinV2-CR keep each stage's window to grid ratio on resize, a call at the current size is now a no-op. Per-stage `window_size` / `window_ratio` supported. Fixes Swin S3 and SwinV2 `window16` / `window12` models
changing outputs on a same-size call.
  * Fix SwinV2 NaN w/ a window size of 1, odd stage feature sizes, and hybrid ViT resize on CUDA / low precision.
* Fix grad accumulation remainder handling (#2821), distributed sampling / repeated augmentation on small datasets, pretrained source overrides and distillation teacher setup, `train.py` w/ `--sched none`, and eval center crop
interpolation for non-square inputs.
* ONNX export compatibility fixes across PyTorch versions, dropout consistency fixes, and more meta-device init fixes.
  
## September 25, 2026

* Multi-label classification support
  * Add `--task multilabel` to `train.py` / `validate.py`, trained with BCE (`MultiLabelClassificationTask`). See the
    [training docs](https://huggingface.co/docs/timm/training_script) for details.
  * Targets can be lists of class indices, dense multi-hot vectors (`--target-format multihot`), or one binary field
    per class (`--target-key a,b,c`), with `/` paths into nested fields. Supported by the `hfds`, `hfids`, `tfds`
    (multi-field keys), and `wds` readers.
  * Mixup / CutMix and NaFlex loaders support multi-label targets. Label smoothing moves each target towards 0.5
    and defaults to 0 for multi-label.
  * Validation reports mAP (the default `--eval-metric`) plus micro, macro, and sample F1, with an F1 threshold set by
    `--multilabel-threshold`.
  * New dataset-level evaluators compute metrics over the full validation set across distributed ranks, excluding
    distributed sampler padding.
  * See the [timm multi-label image datasets](https://huggingface.co/collections/timm/timm-multi-label-image-datasets)
    collection on the Hub for example / benchmark datasets that work out of the box.
* New loss options via `--loss`
  * Poly-1, Two-way, ZLPR, and Distribution-Balanced losses.
  * Class-frequency weighting for BCE / ASL, computed from label counts gathered by `class_weights.py`.
  * ASL is now fp16-safe and defaults to `batchmean` reduction
* RoPE Refactor
  * Extend training-only RoPE coordinate augmentation (shift, jitter, rescale) to axial, mixed, and MRoPE embeddings,
    sharing helpers with DINOv3 while preserving its existing augmentation behavior.
  * Add `grid_type` and `normalize_coords` options to the Fourier / rotary builders (`build_fourier_pos_embed`,
    `build_rotary_pos_embed`) and axial RoPE modules, and configurable MRoPE sections in EVA.
  * Support `rotate_half` in the tuple-output `RotaryEmbedding` and the `base` factory variant.
  * Correct rotation-layout forwarding in EVA, NaFlexViT, and MRoPE attention pooling.
  * Keep non-learned RoPE frequency and coordinate buffers at least float32 in half/bfloat16 models, and compute mixed
    RoPE phases in float32 under autocast. These precision fixes can change low-precision model outputs.
  * Fix batched RoPE embeddings for pixel grids and reference-shape rescaling. DINOv3 `grid_indexing='xy'` now swaps
    coordinate channels as requested; outputs for these previously incorrect configurations change.
  * Make `apply_rot_embed`, `apply_rot_embed_cat`, and `apply_rot_embed_list` return each input's dtype, casting after
    the rotation without first downcasting sin/cos.
  * Consolidate the two axial RoPE modules behind a private base. Existing calls remain compatible, but the `bands`
    buffer is now retained in both cached and dynamic modes (`rope.bands` is no longer `None` in cached mode).
* Train/eval mode switch and dtype change fixes for LeViT and related models (EfficentFormer*, EfficientViT-MSRA, TinyViT) w/ cached attention biases.


## What's Changed
* Fix convit GPSA rel_indices staying on stale device/dtype after .to() by @YHC66 in https://github.com/huggingface/pytorch-image-models/pull/2798
* multi-label classification by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2793
* Fix metric-less CheckpointSaver history tracking by @aswanth-07 in https://github.com/huggingface/pytorch-image-models/pull/2761
* fix(distributed): bind the device for `npu:<index>` in init_distributed_device_so() by @li-lizhe in https://github.com/huggingface/pytorch-image-models/pull/2801
* Fix attention bias caches after dtype conversion by @aswanth-07 in https://github.com/huggingface/pytorch-image-models/pull/2802
* Refactor RoPE helpers, improve precision, and extend coordinate augme… by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2797
* Add loss type selection, new losses, and class-frequency weighting by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2803
* Fix sampler small dataset by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2807
* Pretrained source custom load by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2808
* Fix forward_intermediates grad checkpointing and head pruning in EfficientViT (MIT) by @yupengtang in https://github.com/huggingface/pytorch-image-models/pull/2810
* Harden _registry functions. Fix #2816 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2818
* Address onnx export API compat issues across torch versions.  by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2819
* Use the requested interpolation for non-square eval center crop by @Arthur031221 in https://github.com/huggingface/pytorch-image-models/pull/2824
* Fix/swin window resize by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2820
* Fix/dropout cleanup by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2826
* Fix #2821 grad accum remainder bug without args. use by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2827
* Pass --filename-col to inference results writers by @Arthur031221 in https://github.com/huggingface/pytorch-image-models/pull/2825
* Handle sliced Hugging Face dataset splits by @Arthur031221 in https://github.com/huggingface/pytorch-image-models/pull/2828
* Optim fixes / correctness by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2832
* Bump doc-builder workflow pin by @paulinebm in https://github.com/huggingface/pytorch-image-models/pull/2834
* A few more optimizers followup fixes by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2836
* Fix MKL thread chunks breakign exact float tests on GitHub CI runners by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2837
* Read class map .txt and write hub README as UTF-8 by @Aditya17-bot in https://github.com/huggingface/pytorch-image-models/pull/2839
* Fix model EMA for bfloat16 / float16 models by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2841
* Fix train.py crash with --sched none, raise on unknown scheduler names by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2843
* Improve model resize support via set_input_size() by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2842
* CSPNeXT  by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2844
* Add CSPNeXt (RTMDet backbone) with ImageNet-1k classification weights by @munehiro-k in https://github.com/huggingface/pytorch-image-models/pull/2831
* Bump trufflesecurity/trufflehog from 3.97.4 to 3.97.9 in the actions group across 1 directory by @dependabot[bot] in https://github.com/huggingface/pytorch-image-models/pull/2835
* Clean up DPN model without changing the state dict by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2845
* Fix/model input config by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2785
* Fix GCViT global query mixing images within a batch by @tritsystem in https://github.com/huggingface/pytorch-image-models/pull/2848
* Sweep by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2853
* Fix/community prs 2849 2852 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2854
* Fix features_only backward with grad checkpointing for in-place stem activations by @tritsystem in https://github.com/huggingface/pytorch-image-models/pull/2849
* Apply MobileNetV5 fusion to final intermediate features by @aswanth-07 in https://github.com/huggingface/pytorch-image-models/pull/2850
* Freeze root parameters in whole-model freeze and unfreeze by @aswanth-07 in https://github.com/huggingface/pytorch-image-models/pull/2851
* More sweep improvements, fix some failure cases, shared filesystem use by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2855
* Fix/community prs 2856 2858 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2859
* Bunch of little fixes by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2862
* Only lowercase the dataset/reader prefix, keep HF / TFDS names case-sensitive by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2867
* Fix attention pooling for rectangular feature grids by @mottopanikeiku in https://github.com/huggingface/pytorch-image-models/pull/2866
* RADIO (C-RADIOv3/v4) support for NaFlexVit and VisionTransformer by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2864
* Fix InceptionNeXt MlpClassifierHead with num_classes=0 by @ghaith-mq in https://github.com/huggingface/pytorch-image-models/pull/2868
* Clean state dict prefixes by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2871
* A few hparam / train cmd doc additions by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2872
* Fix group_matcher layer groups for RepViT, HGNet, MViTv2 and pre-norm ViTs by @mottopanikeiku in https://github.com/huggingface/pytorch-image-models/pull/2876
* Fix InceptionNeXt feature map size with output_stride 8 / 16 by @mottopanikeiku in https://github.com/huggingface/pytorch-image-models/pull/2875
* Forward intermediates all by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2879
* Add forward_intermediates to CSPNet, NFNet, DLA and DenseNet by @ghaith-mq in https://github.com/huggingface/pytorch-image-models/pull/2870
* Fix reset classifier export by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2881
* Fix reloading saved models after reset_classifier in ViT, EVA, BEiT, NaFlexViT and RepViT by @patrasq in https://github.com/huggingface/pytorch-image-models/pull/2880
* Bump doc-builder to get the docs machine-translated by @mishig25 in https://github.com/huggingface/pytorch-image-models/pull/2884
* Preserve input channels in NaFlex data loaders by @aswanth-07 in https://github.com/huggingface/pytorch-image-models/pull/2886
* Bump the actions group across 1 directory with 5 updates by @dependabot[bot] in https://github.com/huggingface/pytorch-image-models/pull/2885
* Prep for release 1.0.31 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2888

## New Contributors
* @li-lizhe made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2801
* @yupengtang made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2810
* @Arthur031221 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2824
* @Aditya17-bot made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2839
* @munehiro-k made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2831
* @tritsystem made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2848
* @mottopanikeiku made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2866
* @ghaith-mq made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2868
* @patrasq made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2880

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.30...v1.0.31
</Release>

<Release version="v1.0.30" date="September 22, 2026" published="2026-09-22T20:45:03.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.30">
## Release v1.0.30

## September 22, 2026

* Add Qwen-Drive-1.0-4B vision tower weights.
* Fix prefetch loader stream reuse/synchronization and AugMix for non-square and single-band images.
* Fix AdaMuon conv LR scaling, AdafactorBigVision clipping, MARS gradient history, and Kron checkpoint loading.
* Fix stale eval attention bias caches, pruned model feature metadata, dilated max-pool padding, and asymmetric loss gradient mode.
* Release 1.0.30

## September 10-11, 2026

* Add Qwen3-VL / Qwen3.5 / Qwen3.8 ViT classifier and encoder variants, including classifiers w/ and w/o the native spatial merger.
* Add Sapiens2 ViTs with EVA and NaFlexViT support. Both model additions include native timm Hub weights. Thanks [Yonghye Kwon](https://github.com/developer0hye).
* Add iFormer with native timm Hub weights. Thanks [Ryan Hou](https://github.com/brianhou0208).
* Add EfficientViM with native timm Hub weights. Thanks [smellslikeml](https://github.com/smellslikeml).
* Add DeepSeek-V4 / V4.1 ViT classifiers and encoders with native timm Hub weights. Thanks [Yonghye Kwon](https://github.com/developer0hye).
* Switch default NaFlex SigLIP position interpolation, `inference.py` input-size selection, and improve `--model-kwargs` parsing.
* Fix meta-device construction and device/dtype factory support for a few models; move config tests to meta for faster CI, meta context coverage.


## What's Changed
* Fix ParseKwargs crash on a --model-kwargs value containing '=' or a non-literal by @winklemad in https://github.com/huggingface/pytorch-image-models/pull/2758
* Sapiens2 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2768
* Add Sapiens2 (Meta human-centric ViT) to Eva: GQA, attn-only LayerScale, 5 model defs by @developer0hye in https://github.com/huggingface/pytorch-image-models/pull/2763
* Fix inference test-time preprocessing defaults by @Pdevadiga45 in https://github.com/huggingface/pytorch-image-models/pull/2769
* Qwen3 vit by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2771
* Add Qwen3.5 / Qwen3.8 (Qwen3-VL) vision encoder as qwen3_vit by @developer0hye in https://github.com/huggingface/pytorch-image-models/pull/2765
* Integrate iformer & efficientvim by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2776
* Add iFormer by @brianhou0208 in https://github.com/huggingface/pytorch-image-models/pull/2757
* Add EfficientViM by @smellslikeml in https://github.com/huggingface/pytorch-image-models/pull/2767
* Integrate deepseek vit by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2777
* Add DeepSeek-V4 / V4.1 vision encoder as deepseek_vit by @developer0hye in https://github.com/huggingface/pytorch-image-models/pull/2775
* Fix models that cannot be created with 'meta' device context. by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2778
* Fix #2781, CUDA stream allocation by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2784
* fix augmix mixing buffer shape for non-square and single-band images by @caiotheodoro in https://github.com/huggingface/pytorch-image-models/pull/2783
* chore: enable Dependabot weekly GitHub Actions bumps by @hf-dependantbot-rollout[bot] in https://github.com/huggingface/pytorch-image-models/pull/2703
* Bump the actions group with 7 updates by @dependabot[bot] in https://github.com/huggingface/pytorch-image-models/pull/2787
* Preserve caller gradient mode in AsymmetricLossMultiLabel by @TomasGuija in https://github.com/huggingface/pytorch-image-models/pull/2760
* Add Qwen-Drive-1.0-4B vision tower as a qwen3_vit pretrained tag by @developer0hye in https://github.com/huggingface/pytorch-image-models/pull/2774
* Copy the previous gradient in Mars instead of keeping a reference to p.grad by @Nicholas022400701 in https://github.com/huggingface/pytorch-image-models/pull/2786
* Fix SAME padding for dilated max pooling by @trueoneplusone in https://github.com/huggingface/pytorch-image-models/pull/2779
* Fix feature_info channels for pruned models (#2370) by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2789
* Fix RMS clipping direction in AdafactorBigVision by @Nicholas022400701 in https://github.com/huggingface/pytorch-image-models/pull/2790
* Fix AdaMuon LR scale for conv weights by @Nicholas022400701 in https://github.com/huggingface/pytorch-image-models/pull/2791
* Fix shadowed Kron __setstate__ so loaded checkpoints get corrected_weight_decay by @Nicholas022400701 in https://github.com/huggingface/pytorch-image-models/pull/2792
* Clear the eval-mode attention bias cache on every train()/eval() call by @YHC66 in https://github.com/huggingface/pytorch-image-models/pull/2795

## New Contributors
* @Pdevadiga45 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2769
* @smellslikeml made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2767
* @caiotheodoro made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2783
* @hf-dependantbot-rollout[bot] made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2703
* @TomasGuija made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2760
* @Nicholas022400701 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2786
* @trueoneplusone made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2779
* @YHC66 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2795

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.29...v1.0.30
</Release>

<Release version="v1.0.29" date="August 28, 2026" published="2026-08-28T15:05:27.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.29">
## Release v1.0.29

## August 27, 2026
* Add [LowFormer](https://github.com/altair199797/lowformer) model and weights
* Fix or merge a host of real but small or not commonly encountered bugs surfaced by agents
* Release 1.0.29

## August 11, 2026
* Add model defs and pretrained weights for
  * [CPUBone](https://github.com/altair199797/CPUBone), an efficient vision backbone for devices with low parallelization capabilities.
  * [PP-LCNetV2](https://github.com/PaddlePaddle/PaddleClas/blob/release/2.6/docs/en/models/PP-LCNetV2_en.md).
  * [LingBot-Vision](https://arxiv.org/abs/2607.05247).
* Add per-batch image and batch size scheduling for non-NaFlex training, including progressive small-to-large resolution schedules.
* Improve NaFlexViT with key-only attention masks, compile-friendly and memory-efficient position sampling, and cached variable-patch interpolation.
* Improve model factory path and checkpoint handling; fix equal-token positional embedding resampling.

## What's Changed
* Add optional key-only masks for NaFlexVit self-attention by @chenghuichen in https://github.com/huggingface/pytorch-image-models/pull/2719
* Make NaFlex position embedding sampling compile-friendly and memory-efficient by @chenghuichen in https://github.com/huggingface/pytorch-image-models/pull/2720
* Variable train  image sizes by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2722
* Improved model factory path handling. Fix #2725 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2727
* Naflex interp cache by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2728
* Cpubone by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2733
* Fix positional embedding resampling for equal-token grid changes by @aswanth-07 in https://github.com/huggingface/pytorch-image-models/pull/2737
* Add LingBot-Vision models by @belfner in https://github.com/huggingface/pytorch-image-models/pull/2730
* Add PP-LCNetV2 (lcnetv2_small / lcnetv2_base / lcnetv2_large) by @developer0hye in https://github.com/huggingface/pytorch-image-models/pull/2736
* Fix CutMix minmax border sampling by @tandede in https://github.com/huggingface/pytorch-image-models/pull/2739
* A few more mixup / cutmix / bbox range fixups by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2740
* Return the fused RMSNorm2d result directly by @Robertboy18 in https://github.com/huggingface/pytorch-image-models/pull/2741
* Fix custom input channels for pruned EfficientNet models by @aswanth-07 in https://github.com/huggingface/pytorch-image-models/pull/2742
* Preserve classifier head device and dtype on reset by @aswanth-07 in https://github.com/huggingface/pytorch-image-models/pull/2743
* Fix load_pretrained helper in NaFlexVit, make purpose clear by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2746
* Fix AttributeError in auto-augment factories when hparams is None by @winklemad in https://github.com/huggingface/pytorch-image-models/pull/2747
* Fix AdafactorBigVision dropping the factored row normalization for tall matrices by @winklemad in https://github.com/huggingface/pytorch-image-models/pull/2749
* Fix auto-augment 'inc' and 'b' config flags ignoring the documented 0 value by @winklemad in https://github.com/huggingface/pytorch-image-models/pull/2748
* Fix SGDW foreach path applying weight decay once per (device, dtype) group by @winklemad in https://github.com/huggingface/pytorch-image-models/pull/2750
* Fix OptimizerRegistry.create_optimizer crash on layer_decay without an explicit min scale by @winklemad in https://github.com/huggingface/pytorch-image-models/pull/2754
* Fix device / dtype change for all classifier resets by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2752
* Add LowFormer by @brianhou0208 in https://github.com/huggingface/pytorch-image-models/pull/2755

## New Contributors
* @chenghuichen made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2719
* @aswanth-07 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2737
* @tandede made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2739
* @Robertboy18 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2741
* @winklemad made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2747

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.28...v1.0.29
</Release>

<Release version="v1.0.28" date="July 11, 2026" published="2026-07-11T17:25:29.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.28">
## Release v1.0.28

## July 10, 2026
* Improve optimizer `torch.compile` and tensor learning-rate support.
* Extend NaFlexViT patch-layout (for NaFlex-CLAP), and `forward_intermediates` (NaFlex dict input) support.
* Harden pickle loading and improve custom-label inference.
* Release 1.0.28

## May 27, 2026
* Add model defs and pretrained weights for EUPE ViT (DINOv3-style) and ConvNeXt models. See the [Efficient Universal Perception Encoder paper](https://arxiv.org/abs/2603.22387).
* Add TIPSv2 model defs and pretrained weights for (DINOv2-style) ViTs. See the [TIPSv2 paper](https://arxiv.org/abs/2604.12012).


## What's Changed
* validate.py: add --seed for reproducible validation by @janbernloehr in https://github.com/huggingface/pytorch-image-models/pull/2702
* Add tipsv2 vision encoders by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2705
* Initial EUPE model defs, slight tweaks for weight conversion. by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2706
* Support better torch.compile use by using torch.tensor for step count by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2704
* Preliminary naflexvit support for genlip image encoder (gated attn + mrope)) by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2707
* Address a few pickle security concerns. by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2708
* Fix ZeroDivisionError in AverageMeter.update when count is 0 by @CharlesCNorton in https://github.com/huggingface/pytorch-image-models/pull/2709
* Adding patchify_channels_last flag to unify support across data pipeline & naflex cfg/model by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2710
* Fix reg_token not being added to no-weight decay and stem/embed lr group by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2713
* Make doc builds faster by @mishig25 in https://github.com/huggingface/pytorch-image-models/pull/2714
* Fix inference.py labels for models fine-tuned on non-ImageNet datasets by @javierdejesusda in https://github.com/huggingface/pytorch-image-models/pull/2715
* Support NaFlex dict inputs for forward_intermediates by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2716
* Fixes for 1.0.28 release by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2717

## New Contributors
* @janbernloehr made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2702
* @CharlesCNorton made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2709
* @javierdejesusda made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2715

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.27...v1.0.28
</Release>

<Release version="v1.0.27" date="May 8, 2026" published="2026-05-08T20:06:58.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.27">
## Release v1.0.27

## April 23, 2026
* Add Gemma4 ViT encoders w/ NaFlex pipeline support (variable aspect/size per image). Thanks [Yonghye Kwon](https://github.com/developer0hye)
* Support DINOv3 weights in NaFlexVit. Thanks [Yonghye Kwon](https://github.com/developer0hye)
* Some improvements to Muon fallback (AdamW/NadamW) lr behavior

## What's Changed
* 🔒 Pin GitHub Actions to commit SHAs by @paulinebm in https://github.com/huggingface/pytorch-image-models/pull/2689
* Improve fallback (adamw/nadamw) LR handling for Muon optimizer by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2688
* Fix NaFlexVit DINOv3 support: propagate rope_type and rotate_half by @developer0hye in https://github.com/huggingface/pytorch-image-models/pull/2692
* chore: bump doc-builder SHA for PR upload workflow by @rtrompier in https://github.com/huggingface/pytorch-image-models/pull/2694
* Gemma4 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2697
* Add encoder_pool option to gemma4 classification model to toggle soft… by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2698
* Fix some performance regressions with torch.compile + Tasks. Fix #2693 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2699

## New Contributors
* @paulinebm made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2689

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.26...v1.0.27
</Release>

<Release version="v1.0.26" date="March 23, 2026" published="2026-03-23T18:13:40.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.26">
## Release v1.0.26

## March 23, 2026
* Improve pickle checkpoint handling security. Default all loading to `weights_only=True`, add safe_global for ArgParse.
* Improve attention mask handling for core ViT/EVA models & layers. Resolve bool masks, pass `is_causal` through for SSL tasks.
* Fix class & register token uses with ViT and no pos embed enabled.
* Add Patch Representation Refinement (PRR) as a pooling option in ViT. Thanks Sina (https://github.com/sinahmr).
* Improve consistency of output projection / MLP dimensions for attention pooling layers.
* Hiera model F.SDPA optimization to allow Flash Attention kernel use.
* Caution added to SGDP optimizer.
* Release 1.0.26. First maintenance release since my departure from Hugging Face.

## What's Changed
* fix: replace 5 bare except clauses with except Exception by @haosenwang1018 in https://github.com/huggingface/pytorch-image-models/pull/2672
* Add timmx model export tool to README by @Boulaouaney in https://github.com/huggingface/pytorch-image-models/pull/2673
* Enhance SGDP optimizer with caution parameter by @Yuan-Jinghui in https://github.com/huggingface/pytorch-image-models/pull/2675
* Fix CLS and Reg tokens usage when pos_embed is disabled by @sinahmr in https://github.com/huggingface/pytorch-image-models/pull/2676
* default weights_only=True for load fns by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2679
* Fix Hiera global attention to use 4D tensors for efficient SDPA dispatch by @Raiden129 in https://github.com/huggingface/pytorch-image-models/pull/2680
* Improve 2d and latent attention pool dimension handling. Fix #2682 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2684
* Improve attention mask handling for vision_transformer and eva and related blocks by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2686
* Implement PRR as a pooling module. Alternative to #2678 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2685

## New Contributors
* @haosenwang1018 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2672
* @Raiden129 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2680

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.25...v1.0.26
</Release>

<Release version="v1.0.25" date="February 23, 2026" published="2026-02-23T17:22:16.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.25">
## Release v1.0.25

## Feb 23, 2026
* Add token distillation training support to distillation task wrappers
* Remove some torch.jit usage in prep for official deprecation
* Caution added to AdamP optimizer
* Call reset_parameters() even if meta-device init so that buffers get init w/ hacks like init_empty_weights
* Tweak Muon optimizer to work with DTensor/FSDP2 (clamp_ instead of clamp_min_, alternate NS branch for DTensor)
* Release 1.0.25

## Jan 21, 2026
* **Compat Break**: Fix oversight w/ QKV vs MLP bias in `ParallelScalingBlock` (& `DiffParallelScalingBlock`)
  * Does not impact any trained `timm` models but could impact downstream use.

## What's Changed
* Token distill task & distill task refactoring by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2647
* Fix distilled head dropout using wrong token in PiT forward_head by @hassonofer in https://github.com/huggingface/pytorch-image-models/pull/2649
* Fix #2653, no models with weights impacted so just a clean fix by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2654
* Add the cautious optimizer to AdamP. by @Yuan-Jinghui in https://github.com/huggingface/pytorch-image-models/pull/2657
* Enhance the numerical stability of the Cautious Optimizer by @Yuan-Jinghui in https://github.com/huggingface/pytorch-image-models/pull/2658
* Some misc fixes for torch.jit deprecation and meta device init by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2664
* fix(optim): replace bare except with Exception in Lion optimizer by @llukito in https://github.com/huggingface/pytorch-image-models/pull/2666
* Change clamp_min_ to clamp_(min=) as former doesn't work with DTensor / FSDP2 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2668
* Add DTensor compatible NS impl for Muon by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2669

## New Contributors
* @Yuan-Jinghui made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2657
* @llukito made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2666

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.24...v1.0.25
</Release>

<Release version="v1.0.24" date="January 7, 2026" published="2026-01-07T00:28:47.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.24">
## Release v1.0.24

## Jan 5 & 6, 2025
* Patch Release 1.0.24 (fix for 1.0.23)
* Add new benchmark result csv files for inference timing on all models w/ RTX Pro 6000, 5090, and 4090 cards w/ PyTorch 2.9.1
* Fix moved module error in deprecated timm.models.layers import path that impacts legacy imports
* Release 1.0.23

## Dec 30, 2025
* Add better NAdaMuon trained `dpwee`, `dwee`, `dlittle` (differential) ViTs with a small boost over previous runs
  * https://huggingface.co/timm/vit_dlittle_patch16_reg1_gap_256.sbb_nadamuon_in1k (83.24% top-1)
  * https://huggingface.co/timm/vit_dwee_patch16_reg1_gap_256.sbb_nadamuon_in1k  (81.80% top-1)
  * https://huggingface.co/timm/vit_dpwee_patch16_reg1_gap_256.sbb_nadamuon_in1k (81.67% top-1)
* Add a ~21M param `timm` variant of the CSATv2 model at 512x512 & 640x640
  * https://huggingface.co/timm/csatv2_21m.sw_r640_in1k (83.13% top-1)
  * https://huggingface.co/timm/csatv2_21m.sw_r512_in1k (82.58% top-1)
* Factor non-persistent param init out of `__init__` into a common method that can be externally called via `init_non_persistent_buffers()` after meta-device init. 
  
## Dec 12, 2025
* Add CSATV2 model (thanks https://github.com/gusdlf93) -- a lightweight but high res model with DCT stem & spatial attention. https://huggingface.co/Hyunil/CSATv2
* Add AdaMuon and NAdaMuon optimizer support to existing `timm` Muon impl. Appears more competitive vs AdamW with familiar hparams for image tasks.
* End of year PR cleanup, merge aspects of several long open PR
  * Merge differential attention (`DiffAttention`), add corresponding `DiffParallelScalingBlock` (for ViT), train some wee vits
    * https://huggingface.co/timm/vit_dwee_patch16_reg1_gap_256.sbb_in1k
    * https://huggingface.co/timm/vit_dpwee_patch16_reg1_gap_256.sbb_in1k
  * Add a few pooling modules, `LsePlus` and `SimPool`
  * Cleanup, optimize `DropBlock2d` (also add support to ByobNet based models)
* Bump unit tests to PyTorch 2.9.1 + Python 3.13 on upper end, lower still PyTorch 1.13 + Python 3.10
  
## Dec 1, 2025
* Add lightweight task abstraction, add logits and feature distillation support to train script via new tasks.
* Remove old APEX AMP support

## What's Changed
* Add val-interval argument by @t0278611 in https://github.com/huggingface/pytorch-image-models/pull/2606
* Add coord attn and some variants that I had lying around by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2617
* Distill fixups by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2598
* A simplification and some fixes for DropBlock2d.  by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2620
* Other pooling...  by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2621
* Experimenting with differential attention by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2314
* Differential + parallel attn by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2625
* AdaMuon impl w/ a few other ideas based on recent reading by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2626
* Csatv2 contribution by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2627
* Add HParams sections to hfdocs by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2630
* Upgrade GitHub Actions for Node 24 compatibility by @salmanmkc in https://github.com/huggingface/pytorch-image-models/pull/2633
* [BUG] Modify autocasting in fast normalization functions to handle optional weight params safely by @tesfaldet in https://github.com/huggingface/pytorch-image-models/pull/2631
* 'init_non_persistent_buffers' scheme by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2632
* Add docstrings to layer helper functions and modules by @raimbekovm in https://github.com/huggingface/pytorch-image-models/pull/2634
* refactor(scheduler): add type hints to CosineLRScheduler by @haru-256 in https://github.com/huggingface/pytorch-image-models/pull/2640
* A few misc weights to close out 2025 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2639
* Update typing in other scheduler classes. Add unit tests. by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2641

## New Contributors
* @t0278611 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2606
* @salmanmkc made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2633
* @tesfaldet made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2631
* @raimbekovm made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2634
* @haru-256 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2640

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.22...v1.0.24
</Release>

<Release version="v1.0.23" date="January 5, 2026" published="2026-01-05T21:42:22.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.23">
## Release v1.0.23

## Dec 30, 2025
* Add better NAdaMuon trained `dpwee`, `dwee`, `dlittle` (differential) ViTs with a small boost over previous runs
  * https://huggingface.co/timm/vit_dlittle_patch16_reg1_gap_256.sbb_nadamuon_in1k (83.24% top-1)
  * https://huggingface.co/timm/vit_dwee_patch16_reg1_gap_256.sbb_nadamuon_in1k  (81.80% top-1)
  * https://huggingface.co/timm/vit_dpwee_patch16_reg1_gap_256.sbb_nadamuon_in1k (81.67% top-1)
* Add a ~21M param `timm` variant of the CSATv2 model at 512x512 & 640x640
  * https://huggingface.co/timm/csatv2_21m.sw_r640_in1k (83.13% top-1)
  * https://huggingface.co/timm/csatv2_21m.sw_r512_in1k (82.58% top-1)
* Factor non-persistent param init out of `__init__` into a common method that can be externally called via `init_non_persistent_buffers()` after meta-device init. 
  
## Dec 12, 2025
* Add CSATV2 model (thanks https://github.com/gusdlf93) -- a lightweight but high res model with DCT stem & spatial attention. https://huggingface.co/Hyunil/CSATv2
* Add AdaMuon and NAdaMuon optimizer support to existing `timm` Muon impl. Appears more competitive vs AdamW with familiar hparams for image tasks.
* End of year PR cleanup, merge aspects of several long open PR
  * Merge differential attention (`DiffAttention`), add corresponding `DiffParallelScalingBlock` (for ViT), train some wee vits
    * https://huggingface.co/timm/vit_dwee_patch16_reg1_gap_256.sbb_in1k
    * https://huggingface.co/timm/vit_dpwee_patch16_reg1_gap_256.sbb_in1k
  * Add a few pooling modules, `LsePlus` and `SimPool`
  * Cleanup, optimize `DropBlock2d` (also add support to ByobNet based models)
* Bump unit tests to PyTorch 2.9.1 + Python 3.13 on upper end, lower still PyTorch 1.13 + Python 3.10
  
## Dec 1, 2025
* Add lightweight task abstraction, add logits and feature distillation support to train script via new tasks.
* Remove old APEX AMP support

## What's Changed
* Add val-interval argument by @t0278611 in https://github.com/huggingface/pytorch-image-models/pull/2606
* Add coord attn and some variants that I had lying around by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2617
* Distill fixups by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2598
* A simplification and some fixes for DropBlock2d.  by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2620
* Other pooling...  by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2621
* Experimenting with differential attention by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2314
* Differential + parallel attn by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2625
* AdaMuon impl w/ a few other ideas based on recent reading by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2626
* Csatv2 contribution by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2627
* Add HParams sections to hfdocs by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2630
* Upgrade GitHub Actions for Node 24 compatibility by @salmanmkc in https://github.com/huggingface/pytorch-image-models/pull/2633
* [BUG] Modify autocasting in fast normalization functions to handle optional weight params safely by @tesfaldet in https://github.com/huggingface/pytorch-image-models/pull/2631
* 'init_non_persistent_buffers' scheme by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2632
* Add docstrings to layer helper functions and modules by @raimbekovm in https://github.com/huggingface/pytorch-image-models/pull/2634
* refactor(scheduler): add type hints to CosineLRScheduler by @haru-256 in https://github.com/huggingface/pytorch-image-models/pull/2640
* A few misc weights to close out 2025 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2639
* Update typing in other scheduler classes. Add unit tests. by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2641

## New Contributors
* @t0278611 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2606
* @salmanmkc made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2633
* @tesfaldet made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2631
* @raimbekovm made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2634
* @haru-256 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2640

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.22...v1.0.23
</Release>

<Release version="v1.0.22" date="November 5, 2025" published="2025-11-05T04:08:03.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.22">
## Release v1.0.22

Patch release for priority LayerScale initialization regression in 1.0.21

## What's Changed
* Add some weights for efficientnet_x / efficientnet_h models by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2602
* Update result csvs by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2603
* Fix LayerScale ignoring init_values by @Ilya-Fradlin in https://github.com/huggingface/pytorch-image-models/pull/2605

## New Contributors
* @Ilya-Fradlin made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2605

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.21...v1.0.22
</Release>

<Release version="v1.0.21" date="October 24, 2025" published="2025-10-24T22:39:31.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.21">
## Release v1.0.21

## Oct 16-20, 2025
* Add an impl of the Muon optimizer (based on https://github.com/KellerJordan/Muon) with customizations
  * extra flexibility and improved handling for conv weights and fallbacks for weight shapes not suited for orthogonalization
  * small speedup for NS iterations by reducing allocs and using fused (b)add(b)mm ops
  * by default uses AdamW (or NAdamW if `nesterov=True`) updates if muon not suitable for parameter shape (or excluded via param group flag)
  * like torch impl, select from several LR scale adjustment fns via `adjust_lr_fn`
  * select from several NS coefficient presets or specify your own via `ns_coefficients`
* First 2 steps of 'meta' device model initialization supported
  * Fix several ops that were breaking creation under 'meta' device context
  * Add device & dtype factory kwarg support to all models and modules (anything inherting from nn.Module) in `timm`
* License fields added to pretrained cfgs in code
* Release 1.0.21

## What's Changed
* Add calculate_drop_path_rates helper by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2589
* Review `huggingface_hub` integration by @Wauplin in https://github.com/huggingface/pytorch-image-models/pull/2592
* Adding device/dtype factory_kwargs to modules and models by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2591
* Consistent license handling throughout timm by @alexanderdann in https://github.com/huggingface/pytorch-image-models/pull/2585
* Add impl of Muon optimizer. Fix #2580 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2596
* Rename 'simple' flag for Muon to 'fallback' by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2599

## New Contributors
* @alexanderdann made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2585

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.20...v1.0.21
</Release>

<Release version="v1.0.20" date="September 21, 2025" published="2025-09-21T17:28:50.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.20">
## Release v1.0.20

## Sept 21, 2025
* Remap DINOv3 ViT weight tags from `lvd_1689m` -> `lvd1689m` to match (same for `sat_493m` -> `sat493m`)
* Release 1.0.20

## Sept 17, 2025
* DINOv3 (https://arxiv.org/abs/2508.10104) ConvNeXt and ViT models added. ConvNeXt models were mapped to existing `timm` model. ViT support done via the EVA base model w/ a new `RotaryEmbeddingDinoV3` to match the DINOv3 specific RoPE impl
  * HuggingFace Hub: https://huggingface.co/collections/timm/timm-dinov3-68cb08bb0bee365973d52a4d
* MobileCLIP-2 (https://arxiv.org/abs/2508.20691) vision encoders. New MCI3/MCI4 FastViT variants added and weights mapped to existing FastViT and B, L/14 ViTs.
* MetaCLIP-2 Worldwide (https://arxiv.org/abs/2507.22062) ViT encoder weights added.
* SigLIP-2 (https://arxiv.org/abs/2502.14786) NaFlex ViT encoder weights added via timm NaFlexViT model.
* Misc fixes and contributions

## What's Changed
* Pass init_values at hieradet_sam2 by @hassonofer in https://github.com/huggingface/pytorch-image-models/pull/2559
* Add mobileclip2 encoder weights by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2560
* Add support for Gemma 3n MobileNetV5 encoder weight loading by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2561
* Fix #2562, add siglip2 naflex vit encoder weights by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2564
* fix: create results_dir if missing before saving results by @zhima771 in https://github.com/huggingface/pytorch-image-models/pull/2576
* feat(validate): add precision, recall, and F1 metrics by @ha405 in https://github.com/huggingface/pytorch-image-models/pull/2568
* Allow user to ask for features other than image and label in ImageDataset by @grodino in https://github.com/huggingface/pytorch-image-models/pull/2571
* Add MobileCLIP2 image encoders by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2578
* Add DINOv3 support by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2579

## New Contributors
* @hassonofer made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2559
* @zhima771 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2576
* @ha405 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2568

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.19...v1.0.20
</Release>

<Release version="v1.0.19" date="July 24, 2025" published="2025-07-24T03:06:41.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.19">
## Release v1.0.19

Patch release for Python 3.9 compat break in 1.0.18

## July 23, 2025
* Add `set_input_size()` method to EVA models, used by OpenCLIP 3.0.0 to allow resizing for timm based encoder models.
* Release 1.0.18, needed for PE-Core S & T models in OpenCLIP 3.0.0
* Fix small typing issue that broke Python 3.9 compat. 1.0.19 patch release.

## July 21, 2025
* ROPE support added to NaFlexViT. All models covered by the EVA base (`eva.py`) including EVA, EVA02, Meta PE ViT, `timm` SBB ViT w/ ROPE, and Naver ROPE-ViT can be now loaded in NaFlexViT when `use_naflex=True` passed at model creation time
* More Meta PE ViT encoders added, including small/tiny variants, lang variants w/ tiling, and more spatial variants.
* PatchDropout fixed with NaFlexViT and also w/ EVA models (regression after adding Naver ROPE-ViT)
* Fix XY order with grid_indexing='xy', impacted non-square image use in 'xy' mode (only ROPE-ViT and PE impacted).

## What's Changed
* Add ROPE support to NaFlexVit (axial and mixed), and support most (all?) EVA based vit models & weights by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2552
* Support set_input_size() in EVA models by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2554


**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.17...v1.0.18
</Release>

<Release version="v1.0.18" date="July 23, 2025" published="2025-07-23T20:03:48.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.18">
## Release v1.0.18

## July 23, 2025
* Add `set_input_size()` method to EVA models, used by OpenCLIP 3.0.0 to allow resizing for timm based encoder models.
* Release 1.0.18, needed for PE-Core S & T models in OpenCLIP 3.0.0

## July 21, 2025
* ROPE support added to NaFlexViT. All models covered by the EVA base (`eva.py`) including EVA, EVA02, Meta PE ViT, `timm` SBB ViT w/ ROPE, and Naver ROPE-ViT can be now loaded in NaFlexViT when `use_naflex=True` passed at model creation time
* More Meta PE ViT encoders added, including small/tiny variants, lang variants w/ tiling, and more spatial variants.
* PatchDropout fixed with NaFlexViT and also w/ EVA models (regression after adding Naver ROPE-ViT)
* Fix XY order with grid_indexing='xy', impacted non-square image use in 'xy' mode (only ROPE-ViT and PE impacted).

## What's Changed
* Add ROPE support to NaFlexVit (axial and mixed), and support most (all?) EVA based vit models & weights by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2552
* Support set_input_size() in EVA models by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2554


**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.17...v1.0.18
</Release>

<Release version="v1.0.17" date="July 10, 2025" published="2025-07-10T16:04:42.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.17">
## Release v1.0.17

## July 7, 2025
* MobileNet-v5 backbone tweaks for improved Google Gemma 3n behaviour (to pair with updated official weights)
  * Add stem bias (zero'd in updated weights, compat break with old weights)
  * GELU -> GELU (tanh approx). A minor change to be closer to JAX
* Add two arguments to layer-decay support, a min scale clamp and 'no optimization' scale threshold
* Add 'Fp32' LayerNorm, RMSNorm, SimpleNorm variants that can be enabled to force computation of norm in float32
* Some typing, argument cleanup for norm, norm+act layers done with above
* Support Naver ROPE-ViT (https://github.com/naver-ai/rope-vit) in `eva.py`, add RotaryEmbeddingMixed module for mixed mode, weights on HuggingFace Hub

|model                                             |img_size|top1  |top5  |param_count|
|--------------------------------------------------|--------|------|------|-----------|
|vit_large_patch16_rope_mixed_ape_224.naver_in1k  |224     |84.84 |97.122|304.4      |
|vit_large_patch16_rope_mixed_224.naver_in1k      |224     |84.828|97.116|304.2      |
|vit_large_patch16_rope_ape_224.naver_in1k        |224     |84.65 |97.154|304.37     |
|vit_large_patch16_rope_224.naver_in1k            |224     |84.648|97.122|304.17     |
|vit_base_patch16_rope_mixed_ape_224.naver_in1k   |224     |83.894|96.754|86.59      |
|vit_base_patch16_rope_mixed_224.naver_in1k       |224     |83.804|96.712|86.44      |
|vit_base_patch16_rope_ape_224.naver_in1k         |224     |83.782|96.61 |86.59      |
|vit_base_patch16_rope_224.naver_in1k             |224     |83.718|96.672|86.43      |
|vit_small_patch16_rope_224.naver_in1k            |224     |81.23 |95.022|21.98      |
|vit_small_patch16_rope_mixed_224.naver_in1k      |224     |81.216|95.022|21.99      |
|vit_small_patch16_rope_ape_224.naver_in1k        |224     |81.004|95.016|22.06      |
|vit_small_patch16_rope_mixed_ape_224.naver_in1k  |224     |80.986|94.976|22.06      |
* Some cleanup of ROPE modules, helpers, and FX tracing leaf registration
* Preparing version 1.0.17 release

## What's Changed
* Adding Naver rope-vit compatibility to EVA ViT by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2529
* Update no_grad usage to inference_mode if possible by @GuillaumeErhard in https://github.com/huggingface/pytorch-image-models/pull/2534
* Add a min layer-decay scale clamp, and no optimization threshold to exclude groups from optimization by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2537
* Add stem_bias option to MNV5. Resolve the norm layer so can pass string. by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2538
* Add flag to enable float32 computation for normalization (norm + affine) by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2536
* fix: mnv5 conv_stem bias and GELU with approximate=tanh by @RyanMullins in https://github.com/huggingface/pytorch-image-models/pull/2533
* Fixup casting issues for weights/bias in fp32 norm layers by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2539
* Fix H, W ordering for xy indexing in ROPE by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2541
* Fix 3 typos in README.md by @robin-ede in https://github.com/huggingface/pytorch-image-models/pull/2544

## New Contributors
* @GuillaumeErhard made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2534
* @RyanMullins made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2533
* @robin-ede made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2544

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.16...v1.0.17
</Release>

<Release version="v1.0.16" date="June 26, 2025" published="2025-06-26T18:44:53.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.16">
## Release v1.0.16

## June 26, 2025
* MobileNetV5 backbone (w/ encoder only variant) for [Gemma 3n](https://ai.google.dev/gemma/docs/gemma-3n#parameters) image encoder
* Version 1.0.16 released

## June 23, 2025
* Add F.grid_sample based 2D and factorized pos embed resize to NaFlexViT. Faster when lots of different sizes (based on example by https://github.com/stas-sl).
* Further speed up patch embed resample by replacing vmap with matmul (based on snippet by https://github.com/stas-sl).
* Add 3 initial native aspect NaFlexViT checkpoints created while testing, ImageNet-1k and 3 different pos embed configs w/ same hparams.

 | Model | Top-1 Acc | Top-5 Acc | Params (M) | Eval Seq Len |
 |:---|:---:|:---:|:---:|:---:|
 | [naflexvit_base_patch16_par_gap.e300_s576_in1k](https://hf.co/timm/naflexvit_base_patch16_par_gap.e300_s576_in1k) | 83.67 | 96.45 | 86.63 | 576 |
 | [naflexvit_base_patch16_parfac_gap.e300_s576_in1k](https://hf.co/timm/naflexvit_base_patch16_parfac_gap.e300_s576_in1k) | 83.63 | 96.41 | 86.46 | 576 |
 | [naflexvit_base_patch16_gap.e300_s576_in1k](https://hf.co/timm/naflexvit_base_patch16_gap.e300_s576_in1k) | 83.50 | 96.46 | 86.63 | 576 |
* Support gradient checkpointing for `forward_intermediates` and fix some checkpointing bugs. Thanks https://github.com/brianhou0208
* Add 'corrected weight decay' (https://arxiv.org/abs/2506.02285) as option to AdamW (legacy), Adopt, Kron, Adafactor (BV), Lamb, LaProp, Lion, NadamW, RmsPropTF, SGDW optimizers
* Switch PE (perception encoder) ViT models to use native timm weights instead of remapping on the fly
* Fix cuda stream bug in prefetch loader
  
## June 5, 2025
* Initial NaFlexVit model code. NaFlexVit is a Vision Transformer with:
  1. Encapsulated embedding and position encoding in a single module
  2. Support for nn.Linear patch embedding on pre-patchified (dictionary) inputs
  3. Support for NaFlex variable aspect, variable resolution (SigLip-2: https://arxiv.org/abs/2502.14786)
  4. Support for FlexiViT variable patch size (https://arxiv.org/abs/2212.08013)
  5. Support for NaViT fractional/factorized position embedding (https://arxiv.org/abs/2307.06304)
* Existing vit models in `vision_transformer.py` can be loaded into the NaFlexVit model by adding the `use_naflex=True` flag to `create_model`
  * Some native weights coming soon
* A full NaFlex data pipeline is available that allows training / fine-tuning / evaluating with variable aspect / size images
  * To enable in `train.py` and `validate.py` add the `--naflex-loader` arg, must be used with a NaFlexVit
* To evaluate an existing (classic) ViT loaded in NaFlexVit model w/ NaFlex data pipe:
  * `python validate.py /imagenet --amp -j 8 --model vit_base_patch16_224 --model-kwargs use_naflex=True --naflex-loader --naflex-max-seq-len 256` 
* The training has some extra args features worth noting
  * The `--naflex-train-seq-lens'` argument specifies which sequence lengths to randomly pick from per batch during training
  * The `--naflex-max-seq-len` argument sets the target sequence length for validation
  * Adding `--model-kwargs enable_patch_interpolator=True --naflex-patch-sizes 12 16 24` will enable random patch size selection per-batch w/ interpolation
  * The `--naflex-loss-scale` arg changes loss scaling mode per batch relative to the batch size, `timm` NaFlex loading changes the batch size for each seq len

## May 28, 2025
* Add a number of small/fast models thanks to https://github.com/brianhou0208
  * SwiftFormer - [(ICCV2023) SwiftFormer: Efficient Additive Attention for Transformer-based Real-time Mobile Vision Applications](https://github.com/Amshaker/SwiftFormer) 
  * FasterNet - [(CVPR2023) Run, Don’t Walk: Chasing Higher FLOPS for Faster Neural Networks](https://github.com/JierunChen/FasterNet)
  * SHViT - [(CVPR2024) SHViT: Single-Head Vision Transformer with Memory Efficient](https://github.com/ysj9909/SHViT)
  * StarNet - [(CVPR2024) Rewrite the Stars](https://github.com/ma-xu/Rewrite-the-Stars)
  * GhostNet-V3 [GhostNetV3: Exploring the Training Strategies for Compact Models](https://github.com/huawei-noah/Efficient-AI-Backbones/tree/master/ghostnetv3_pytorch)
* Update EVA ViT (closest match) to support Perception Encoder models (https://arxiv.org/abs/2504.13181) from Meta, loading Hub weights but I still need to push dedicated `timm` weights
  * Add some flexibility to ROPE impl
* Big increase in number of models supporting `forward_intermediates()` and some additional fixes thanks to https://github.com/brianhou0208
  * DaViT, EdgeNeXt, EfficientFormerV2, EfficientViT(MIT), EfficientViT(MSRA), FocalNet, GCViT, HGNet /V2, InceptionNeXt, Inception-V4, MambaOut, MetaFormer, NesT, Next-ViT, PiT, PVT V2, RepGhostNet, RepViT, ResNetV2, ReXNet, TinyViT, TResNet, VoV
* TNT model updated w/ new weights `forward_intermediates()` thanks to https://github.com/brianhou0208
* Add `local-dir:` pretrained schema, can use `local-dir:/path/to/model/folder` for model name to source model / pretrained cfg & weights Hugging Face Hub models (config.json + weights file) from a local folder.
* Fixes, improvements for onnx export

## What's Changed
* Fix arg merging of sknet, old seresnet. Fix #2470 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2471
* Fix onnx export by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2475
* Add local-dir: schema support for model loading (config + weights) from folder by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2476
* Fix: Allow img_size to be int or tuple in PatchEmbed by @sddongxh in https://github.com/huggingface/pytorch-image-models/pull/2477
* Add LightlyTrain Integration for Pretraining Support by @yutong-xiang-97 in https://github.com/huggingface/pytorch-image-models/pull/2474
* Check forward_intermediates features against forward_features output by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2483
* More models support forward_intermediates by @brianhou0208 in https://github.com/huggingface/pytorch-image-models/pull/2482
* Update README.md by @atharva-pathak in https://github.com/huggingface/pytorch-image-models/pull/2484
* remove `download` argument from torch_kwargs for torchvision `ImageNet` class by @ryan-caesar-ramos in https://github.com/huggingface/pytorch-image-models/pull/2486
* Update TNT-(S/B) model weights and add feature extraction support by @brianhou0208 in https://github.com/huggingface/pytorch-image-models/pull/2480
* Add EVA ViT based PE (Perceptual Encoder) impl by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2487
* Add SwiftFormer, SHViT, StarNet, FasterNet and GhostNetV3 by @brianhou0208 in https://github.com/huggingface/pytorch-image-models/pull/2499
* A cleaned up beit3 remap onto vision_transformer.py vit by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2503
* Initial NaFlex ViT model and training support by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2466
* Forgot to compact attention pool branches after verifying by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2507
* Throw exception on non-directory path for pretrained weights by @emmanuel-ferdman in https://github.com/huggingface/pytorch-image-models/pull/2510
* Add corrected_weight decay to several optimizers by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2511
* Doing some Claude enabled docstring, type annotation and other cleanup by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2504
* Fix #2513, be explicit about stream devices by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2515
* Update legacy AdamW impl so it has a multi-tensor impl like NAdamW (n… by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2517
* Fix `head_dim` reference in `AttentionRope` class of `attention.py` by @amorehead in https://github.com/huggingface/pytorch-image-models/pull/2519
* Refactor patch and pos embed resampling based on feedback from https://github.com/stas-sl by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2518
* Add initial weights for my first 3 naflexvit_base models by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2523
* Support gradient checkpointing in `forward_intermediates()` by @brianhou0208 in https://github.com/huggingface/pytorch-image-models/pull/2501
* Update README: add references for additional supported models by @brianhou0208 in https://github.com/huggingface/pytorch-image-models/pull/2526
* MobileNetV5 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2527

## New Contributors
* @sddongxh made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2477
* @yutong-xiang-97 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2474
* @atharva-pathak made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2484
* @ryan-caesar-ramos made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2486
* @emmanuel-ferdman made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2510
* @amorehead made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2519

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.15...v1.0.16
</Release>

<Release version="v1.0.15" date="February 23, 2025" published="2025-02-23T05:07:06.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.15">
## Release v1.0.15

## Feb 21, 2025
* SigLIP 2 ViT image encoders added (https://huggingface.co/collections/timm/siglip-2-67b8e72ba08b09dd97aecaf9)
  * Variable resolution / aspect NaFlex versions are a WIP
* Add 'SO150M2' ViT weights trained with SBB recipes, great results, better for ImageNet than previous attempt w/ less training.
  * `vit_so150m2_patch16_reg1_gap_448.sbb_e200_in12k_ft_in1k` - 88.1% top-1
  * `vit_so150m2_patch16_reg1_gap_384.sbb_e200_in12k_ft_in1k` - 87.9% top-1
  * `vit_so150m2_patch16_reg1_gap_256.sbb_e200_in12k_ft_in1k` - 87.3% top-1
  * `vit_so150m2_patch16_reg4_gap_256.sbb_e200_in12k`
* Updated InternViT-300M '2.5' weights
* Release 1.0.15

## Feb 1, 2025
* FYI PyTorch 2.6 & Python 3.13 are tested and working w/ current main and released version of `timm`

## Jan 27, 2025
* Add Kron Optimizer (PSGD w/ Kronecker-factored preconditioner) 
  * Code from https://github.com/evanatyourservice/kron_torch
  * See also https://sites.google.com/site/lixilinx/home/psgd

## What's Changed
* Fix metavar for `--input-size` by @JosuaRieder in https://github.com/huggingface/pytorch-image-models/pull/2417
* Add arguments to the respective argument groups by @JosuaRieder in https://github.com/huggingface/pytorch-image-models/pull/2416
* Add missing training flag to convert_sync_batchnorm by @collinmccarthy in https://github.com/huggingface/pytorch-image-models/pull/2423
* Fix num_classes update in reset_classifier and RDNet forward head call by @brianhou0208 in https://github.com/huggingface/pytorch-image-models/pull/2421
* timm: add __all__ to __init__ by @adamjstewart in https://github.com/huggingface/pytorch-image-models/pull/2399
* Fiddling with Kron (PSGD) optimizer by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2427
* Try to force numpy<2.0 for torch 1.13 tests, update newest tested torch to 2.5.1 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2429
* Kron flatten improvements + stochastic weight decay by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2431
* PSGD: unify RNG by @ClashLuke in https://github.com/huggingface/pytorch-image-models/pull/2433
* Add vit so150m2 weights by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2439
* adapt_input_conv: add type hints by @adamjstewart in https://github.com/huggingface/pytorch-image-models/pull/2441
* SigLIP 2 by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2440
* timm.models: explicitly export attributes by @adamjstewart in https://github.com/huggingface/pytorch-image-models/pull/2442

## New Contributors
* @collinmccarthy made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2423
* @ClashLuke made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2433

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.14...v1.0.15
</Release>

<Release version="v1.0.14" date="January 19, 2025" published="2025-01-19T23:05:30.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.14">
## Release v1.0.14

## Jan 19, 2025
* Fix loading of LeViT safetensor weights, remove conversion code which should have been deactivated
* Add 'SO150M' ViT weights trained with SBB recipes, decent results, but not optimal shape for ImageNet-12k/1k pretrain/ft
  * `vit_so150m_patch16_reg4_gap_256.sbb_e250_in12k_ft_in1k` - 86.7% top-1
  * `vit_so150m_patch16_reg4_gap_384.sbb_e250_in12k_ft_in1k` - 87.4% top-1
  * `vit_so150m_patch16_reg4_gap_256.sbb_e250_in12k`
* Misc typing, typo, etc. cleanup
* 1.0.14 release to get above LeViT fix out

## What's Changed
* Fix nn.Module type hints by @adamjstewart in https://github.com/huggingface/pytorch-image-models/pull/2400
* Add missing paper title by @JosuaRieder in https://github.com/huggingface/pytorch-image-models/pull/2405
* fix 'timm recipe scripts' link by @JosuaRieder in https://github.com/huggingface/pytorch-image-models/pull/2404
* fix typo in EfficientNet docs by @JosuaRieder in https://github.com/huggingface/pytorch-image-models/pull/2403
* disable abbreviating csv inference output with ellipses by @JosuaRieder in https://github.com/huggingface/pytorch-image-models/pull/2402
* fix incorrect LaTeX formulas by @JosuaRieder in https://github.com/huggingface/pytorch-image-models/pull/2406
* VGG ConvMlp: fix layer defaults/types by @adamjstewart in https://github.com/huggingface/pytorch-image-models/pull/2409
* Implement --no-console-results in inference.py by @JosuaRieder in https://github.com/huggingface/pytorch-image-models/pull/2408
* LeViT safetensors load is broken by conversion code that wasn't deactivated by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2412
* A few more weights by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2413
* Fix typos by @JosuaRieder in https://github.com/huggingface/pytorch-image-models/pull/2415

## New Contributors
* @adamjstewart made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2400

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.13...v1.0.14
</Release>

<Release version="v1.0.13" date="January 9, 2025" published="2025-01-09T18:49:44.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.13">
## Release v1.0.13

## Jan 9, 2025
* Add support to train and validate in pure `bfloat16` or `float16`
* `wandb` project name arg added by https://github.com/caojiaolong, use arg.experiment for name
* Fix old issue w/ checkpoint saving not working on filesystem w/o hard-link support (e.g. FUSE fs mounts)
* 1.0.13 release

## Jan 6, 2025
* Add `torch.utils.checkpoint.checkpoint()` wrapper in `timm.models` that defaults `use_reentrant=False`, unless `TIMM_REENTRANT_CKPT=1` is set in env.

## Dec 31, 2024
* `convnext_nano` 384x384 ImageNet-12k pretrain & fine-tune. https://huggingface.co/models?search=convnext_nano%20r384
* Add AIM-v2 encoders from https://github.com/apple/ml-aim, see on Hub: https://huggingface.co/models?search=timm%20aimv2
* Add PaliGemma2 encoders from https://github.com/google-research/big_vision to existing PaliGemma, see on Hub: https://huggingface.co/models?search=timm%20pali2
* Add missing L/14 DFN2B 39B CLIP ViT, `vit_large_patch14_clip_224.dfn2b_s39b`
* Fix existing `RmsNorm` layer & fn to match standard formulation, use PT 2.5 impl when possible. Move old impl to `SimpleNorm` layer, it's LN w/o centering or bias. There were only two `timm` models using it, and they have been updated.
* Allow override of `cache_dir` arg for model creation
* Pass through `trust_remote_code` for HF datasets wrapper
* `inception_next_atto` model added by creator
* Adan optimizer caution, and Lamb decoupled weighgt decay options
* Some feature_info metadata fixed by https://github.com/brianhou0208
* All OpenCLIP and JAX (CLIP, SigLIP, Pali, etc) model weights that used load time remapping were given their own HF Hub instances so that they work with `hf-hub:` based loading, and thus will work with new Transformers `TimmWrapperModel`

## What's Changed
* Punch cache_dir through model factory / builder / pretrain helpers by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2356
* Yuweihao inception next atto merge by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2360
* Dataset trust remote tweaks by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2361
* Add --dataset-trust-remote-code to the train.py and validate.py scripts by @grodino in https://github.com/huggingface/pytorch-image-models/pull/2328
* Fix feature_info.reduction by @brianhou0208 in https://github.com/huggingface/pytorch-image-models/pull/2369
* Add caution to Adan. Add decouple decay option to LAMB. by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2357
* Switching to timm specific weight instances for open_clip image encoders by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2376
* Fix broken image link in `Quickstart` doc by @ariG23498 in https://github.com/huggingface/pytorch-image-models/pull/2381
* Supporting aimv2 encoders by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2379
* fix: minor typos in markdowns by @ruidazeng in https://github.com/huggingface/pytorch-image-models/pull/2382
* Add 384x384 in12k pretrain and finetune for convnext_nano by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2384
* Fixed unfused attn2d scale by @laclouis5 in https://github.com/huggingface/pytorch-image-models/pull/2387
* Fix MQA V2 by @laclouis5 in https://github.com/huggingface/pytorch-image-models/pull/2388
* Wrap torch checkpoint() fn to default use_reentrant flag to False and allow env var override by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2394
* Add half-precision (bfloat16, float16) support to train & validate scripts by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2397
* Merging wandb project name chages w/ addition by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2398

## New Contributors
* @brianhou0208 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2369
* @ariG23498 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2381
* @ruidazeng made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2382
* @laclouis5 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2387

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.12...v1.0.13
</Release>

<Release version="v1.0.12" date="December 3, 2024" published="2024-12-03T19:05:39.000Z" url="https://github.com/huggingface/pytorch-image-models/releases/tag/v1.0.12">
## Release v1.0.12

## Nov 28, 2024
* More optimizers
  * Add MARS optimizer (https://arxiv.org/abs/2411.10438, https://github.com/AGI-Arena/MARS)
  * Add LaProp optimizer (https://arxiv.org/abs/2002.04839, https://github.com/Z-T-WANG/LaProp-Optimizer)
  * Add masking from 'Cautious Optimizers' (https://arxiv.org/abs/2411.16085, https://github.com/kyleliang919/C-Optim) to Adafactor, Adafactor Big Vision, AdamW (legacy), Adopt, Lamb, LaProp, Lion, NadamW, RMSPropTF, SGDW
  * Cleanup some docstrings and type annotations re optimizers and factory
* Add MobileNet-V4 Conv Medium models pretrained on in12k and fine-tuned in1k @ 384x384
  * https://huggingface.co/timm/mobilenetv4_conv_medium.e250_r384_in12k_ft_in1k
  * https://huggingface.co/timm/mobilenetv4_conv_medium.e250_r384_in12k
  * https://huggingface.co/timm/mobilenetv4_conv_medium.e180_ad_r384_in12k
  * https://huggingface.co/timm/mobilenetv4_conv_medium.e180_r384_in12k
* Add small cs3darknet, quite good for the speed
  * https://huggingface.co/timm/cs3darknet_focus_s.ra4_e3600_r256_in1k

## Nov 12, 2024
* Optimizer factory refactor
  * New factory works by registering optimizers using an OptimInfo dataclass w/ some key traits
  * Add `list_optimizers`, `get_optimizer_class`, `get_optimizer_info` to reworked `create_optimizer_v2` fn to explore optimizers, get info or class
  * deprecate `optim.optim_factory`, move fns to `optim/_optim_factory.py` and `optim/_param_groups.py` and encourage import via `timm.optim`
* Add Adopt (https://github.com/iShohei220/adopt) optimizer
* Add 'Big Vision' variant of Adafactor (https://github.com/google-research/big_vision/blob/main/big_vision/optax.py) optimizer
* Fix original Adafactor to pick better factorization dims for convolutions
* Tweak LAMB optimizer with some improvements in torch.where functionality since original, refactor clipping a bit
* dynamic img size support in vit, deit, eva improved to support resize from non-square patch grids, thanks https://github.com/wojtke

## Oct 31, 2024
Add a set of new very well trained ResNet & ResNet-V2 18/34 (basic block) weights. See https://huggingface.co/blog/rwightman/resnet-trick-or-treat

## Oct 19, 2024
* Cleanup torch amp usage to avoid cuda specific calls, merge support for Ascend (NPU) devices from [MengqingCao](https://github.com/MengqingCao) that should work now in PyTorch 2.5 w/ new device extension autoloading feature. Tested Intel Arc (XPU) in Pytorch 2.5 too and it (mostly) worked.

## What's Changed
* mambaout.py: fixed bug by @NightMachinery in https://github.com/huggingface/pytorch-image-models/pull/2305
* Cleanup some amp related behaviour to better support different (non-cuda) devices by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2308
* Add NPU backend support for val and inference by @MengqingCao in https://github.com/huggingface/pytorch-image-models/pull/2109
* Update some clip pretrained weights to point to new hub locations by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2311
* ResNet vs MNV4 v1/v2 18 & 34 weights by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2316
* Replace deprecated positional argument with --data-dir by @JosuaRieder in https://github.com/huggingface/pytorch-image-models/pull/2322
* Fix typo in train.py: bathes > batches by @JosuaRieder in https://github.com/huggingface/pytorch-image-models/pull/2321
* Fix positional embedding resampling for non-square inputs in ViT by @wojtke in https://github.com/huggingface/pytorch-image-models/pull/2317
* Add trust_remote_code argument to ReaderHfds by @grodino in https://github.com/huggingface/pytorch-image-models/pull/2326
* Extend train epoch schedule by warmup_epochs if warmup_prefix enabled by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2325
* Extend existing unit tests using Cover-Agent by @mrT23 in https://github.com/huggingface/pytorch-image-models/pull/2331
* An impl of adafactor as per big vision (scaling vit) changes by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2320
* Add py.typed file as recommended by PEP 561 by @antoinebrl in https://github.com/huggingface/pytorch-image-models/pull/2252
* Add CODE_OF_CONDUCT.md and CITATION.cff files by @AlinaImtiaz018 in https://github.com/huggingface/pytorch-image-models/pull/2333
* Add some 384x384 small model weights by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2334
* In dist training, update loss running avg every step, sync on log by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2340
* Improve WandB logging by @sinahmr in https://github.com/huggingface/pytorch-image-models/pull/2341
* A few weights to merge Friday by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2343
* Update timm torchvision resnet weight urls to the updated urls in torchvision by @JohannesTheo in https://github.com/huggingface/pytorch-image-models/pull/2346
* More optimizer updates, add MARS, LaProp, add Adopt fix and more by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2347
* Cautious optimizer impl plus some typing cleanup. by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2349
* Add cautious mars, improve test reliability by skipping grad diff for… by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2351
* See if we can avoid some model / layer pickle issues with the aa attr in ConvNormAct by @rwightman in https://github.com/huggingface/pytorch-image-models/pull/2353

## New Contributors
* @MengqingCao made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2109
* @JosuaRieder made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2322
* @wojtke made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2317
* @grodino made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2326
* @AlinaImtiaz018 made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2333
* @sinahmr made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2341
* @JohannesTheo made their first contribution in https://github.com/huggingface/pytorch-image-models/pull/2346

**Full Changelog**: https://github.com/huggingface/pytorch-image-models/compare/v1.0.11...v1.0.12
</Release>

<Pagination cursor="2024-12-03T19:05:39.000Z|2026-04-07T17:29:48.759Z|rel_LARxbI_ltAMDaLu5CJQDB" next="https://releases.sh/hugging-face/timm-pytorch-image-models.md?cursor=2024-12-03T19%3A05%3A39.000Z%7C2026-04-07T17%3A29%3A48.759Z%7Crel_LARxbI_ltAMDaLu5CJQDB&limit=20" />
