PEFT

AI

Parameter-efficient fine-tuning — LoRA and friends — for large pretrained models.

Latest v0.20.0 · by Hugging FaceWebsitehuggingface/peft

Release activity

Release activity — 6 releases across 6 days in the last year. Each cell is one day; darker means more releases that day. Older weeks are hidden at this screen width.
MayJunJulAug
SundayNo releases on Apr 26, 2026No releases on May 3, 2026No releases on May 10, 2026No releases on May 17, 2026No releases on May 24, 2026No releases on May 31, 2026No releases on Jun 7, 2026No releases on Jun 14, 2026No releases on Jun 21, 2026No releases on Jun 28, 2026No releases on Jul 5, 2026No releases on Jul 12, 2026No releases on Jul 19, 2026No releases on Jul 26, 2026No releases on Aug 2, 2026No releases on Aug 9, 2026
MondayNo releases on Apr 27, 2026No releases on May 4, 2026No releases on May 11, 2026No releases on May 18, 2026No releases on May 25, 2026No releases on Jun 1, 2026No releases on Jun 8, 2026No releases on Jun 15, 2026No releases on Jun 22, 2026No releases on Jun 29, 2026No releases on Jul 6, 2026No releases on Jul 13, 2026No releases on Jul 20, 2026No releases on Jul 27, 2026No releases on Aug 3, 2026No releases on Aug 10, 2026
TuesdayNo releases on Apr 28, 2026No releases on May 5, 2026No releases on May 12, 2026No releases on May 19, 2026No releases on May 26, 2026No releases on Jun 2, 2026No releases on Jun 9, 2026No releases on Jun 16, 2026No releases on Jun 23, 2026No releases on Jun 30, 2026No releases on Jul 7, 2026No releases on Jul 14, 2026No releases on Jul 21, 20261 release on Jul 28, 2026No releases on Aug 4, 2026No releases on Aug 11, 2026
WednesdayNo releases on Apr 29, 2026No releases on May 6, 2026No releases on May 13, 2026No releases on May 20, 2026No releases on May 27, 2026No releases on Jun 3, 2026No releases on Jun 10, 2026No releases on Jun 17, 2026No releases on Jun 24, 2026No releases on Jul 1, 2026No releases on Jul 8, 2026No releases on Jul 15, 2026No releases on Jul 22, 2026No releases on Jul 29, 2026No releases on Aug 5, 2026
ThursdayNo releases on Apr 30, 2026No releases on May 7, 2026No releases on May 14, 2026No releases on May 21, 2026No releases on May 28, 2026No releases on Jun 4, 2026No releases on Jun 11, 2026No releases on Jun 18, 2026No releases on Jun 25, 2026No releases on Jul 2, 2026No releases on Jul 9, 2026No releases on Jul 16, 2026No releases on Jul 23, 2026No releases on Jul 30, 2026No releases on Aug 6, 2026
FridayNo releases on May 1, 2026No releases on May 8, 2026No releases on May 15, 2026No releases on May 22, 2026No releases on May 29, 2026No releases on Jun 5, 2026No releases on Jun 12, 2026No releases on Jun 19, 2026No releases on Jun 26, 2026No releases on Jul 3, 2026No releases on Jul 10, 2026No releases on Jul 17, 2026No releases on Jul 24, 2026No releases on Jul 31, 2026No releases on Aug 7, 2026
SaturdayNo releases on May 2, 2026No releases on May 9, 2026No releases on May 16, 2026No releases on May 23, 2026No releases on May 30, 2026No releases on Jun 6, 2026No releases on Jun 13, 2026No releases on Jun 20, 2026No releases on Jun 27, 2026No releases on Jul 4, 2026No releases on Jul 11, 2026No releases on Jul 18, 2026No releases on Jul 25, 2026No releases on Aug 1, 2026No releases on Aug 8, 2026

6 releases in the last year

Changelog

v0.20.0

Added 10
  • Add HiRA (Hadamard High-Rank Adaptation) method which multiplies low-rank product with frozen base weight using Hadamard product instead of additive combination
  • Add GLoRA (Generalized LoRA) method extending LoRA with configurable weight, activation, and bias adaptation
  • Add BEFT (Bias-Efficient Fine-Tuning) method that trains only bias terms, targeting the value projection by default for roughly 0.01% trainable parameters
  • Add MonteCLoRA method treating low-rank parameters as distributions and using Monte Carlo estimation for more robust training
  • Add VeLoRA method compressing intermediate activations with rank-1 sub-token projections to reduce activation memory
  • Add Uni-LoRA method using single global projection across the model for cross-layer parameter sharing

Highlights

This release adds no less than nine new PEFT methods and puts a lot of work into the surrounding infrastructure, for example adding a new image generation benchmark for the method comparison suite and greatly improving the documentation structure.

New Methods
HiRA

@hqsiswiliam added "HiRA: Parameter-Efficient Hadamard High-Rank Adaptation for Large Language Models" to PEFT (#2668). Instead of adding the low-rank product BA to the base weight, HiRA multiplies it elementwise (Hadamard product) with the frozen base weight. Because the base weight itself is full rank, the resulting update is no longer constrained to be low rank, while the trainable parameter count stays the same as LoRA's.

GLoRA

@not-lain contributed GLoRA: "One-for-All: Generalized LoRA for Parameter-Efficient Fine-Tuning" in #3098. It is a flexible PEFT method that extends LoRA with configurable weight, activation, and bias adaptation, delivering richer fine-tuning with no extra inference cost. Use it when you need per-layer flexibility or stronger adaptation than vanilla LoRA. Skip it for non-Linear layers (e.g. Conv/Embedding) or when standard LoRA is already sufficient and simplicity matters.

BEFT

@whubaichuan added "BEFT: Bias-Efficient Fine-Tuning of Language Models" in #3195. BEFT builds on the observation that fine-tuning bias terms alone can be competitive in low-data regimes, but goes further: rather than training all biases, it targets the value projection by default, as the authors found this to be most efficient. This brings the trainable parameter count down to roughly 0.01% of the total parameters.

MonteCLoRA

@victor7246 integrated MonteCLoRA, "Robust and Efficient Fine-tuning of LLMs with Bayesian Reparameterization of Low-Rank Adaptation" in #2943. LoRA is known to be sensitive to hyperparameters like learning rate and batch size. This new LoRA variant addresses this by treating the low-rank parameters as a distribution rather than a point estimate, using Monte Carlo estimation to obtain a low-variance posterior estimate. This should make training noticeably more robust, at the cost of only O(rank) additional parameters.

VeLoRA

@roymiles added VeLoRA: "Memory Efficient Training using Rank-1 Sub-Token Projections" in #3159. Unlike most PEFT methods, this LoRA variant targets activation memory rather than parameter count: intermediate activations are split into sub-tokens and compressed with a fixed rank-1 projection before being cached for the backward pass, then reconstructed during backpropagation. Conceptually, it is similar to gradient checkpointing, with a lower memory saving but running faster.

Uni-LoRA

@KaiyangLi1992 contributed Uni-LoRA: "One Vector is All You Need" in #3257. The paper shows that parameter-efficient LoRA variants like VeRA and VB-LoRA can all be described as projecting the full LoRA parameter space down to a much smaller subspace, differing only in the choice of projection. Uni-LoRA uses a single global projection across the whole model instead of layer-wise ones, which allows cross-layer parameter sharing and thus very low parameter counts.

FRoD

@Bane-Elvin added FRoD, "Full-Rank Efficient Fine-Tuning with Rotational Degrees for Fast Convergence", in #3270. Instead of adding low-rank deltas like LoRA, it reconstructs selected weights with shared rotational subspaces and sparse trainable coefficients. It is especially useful when fast convergence and a higher full-rank capacity ceiling are important, and its large sparse rotational subspace may also be promising for model merging. The main tradeoffs are the costly joint-decomposition initialization and slightly slower forward/backward passes than LoRA, so it may be less attractive for a one-off single-task fine-tune.

MiCA

@sr-networks contributed MiCA, "MiCA Learns More Knowledge Than LoRA and Full Fine-Tuning", in #3260. MiCA is a LoRA variant that initializes B from the SVD of the base weight, taking the left singular vectors belonging to the smallest singular values, and then trains only A. The idea is that adapting these underused directions integrates new knowledge more effectively than adapting the dominant subspace, with less interference with existing capabilities.

DEFT

@MAXNORM8650 added DEFT, "Decompositional Efficient Fine-Tuning for Text-to-Image Models", in #3342. This PEFT method splits a weight update into two learned low-rank parts: a projection that removes a sub-space of the frozen weight, and a low-rank update that injects new content in its place. Use DEFT for personalizing a text-to-image model from a few images while retaining the base model's instruction-following/editability with minimal forgetting. It is less of a fit if you don't need to preserve the base model's other capabilities (a plain additive adapter is simpler) or for layers beyond Linear/Conv1D. The PaRa method is also supported as a special case of DEFT.

Enhancements
Automatic LoRA target selection

@oswaldoludwig added KappaTuneSelector in #3106 based on "The Condition Number as a Scale-Invariant Proxy for Information Encoding in Neural Units". The PR adds a new function, find_kappa_target_modules, which automatically identifies which modules best to target based on the condition number of the base weight matrices. This removes some of the guesswork from choosing target_modules when configuring LoRA or other PEFT methods.

Broader quantization support

#3117 adds generic quantization support to BOFT, MiSS, and VeRA, and #3321 does the same for SHiRA. Previously, quantization support had to be implemented per method and per backend; more methods now work with the common quantization backends out of the box. Note that as a consequence, VeRA now uses the generic quantization backend instead of its previous bitsandbytes quantization backend. If you encounter issues because of that, please let us know.

Multiple adapters with target_parameters

target_parameters, which allows targeting nn.Parameters directly (useful for MoE models that pack experts into a single parameter), now supports adding and loading multiple adapters (#3350).

AdaLoRA for Conv2d

@Anai-Guo added SVDConv2d in #3196, so AdaLoRA can now target torch.nn.Conv2d layers.

Documentation restructure

The PEFT documentation was reorganized in #3300 and #3325, which should make it considerably easier to navigate now that PEFT supports so many methods. Each PEFT method also provides an overview of how well it fares in the PEFT benchmarks. There is also a long tail of documentation improvements from many contributors this release, thanks to everyone who worked on those.

Changes
AutoGPTQ deprecation

Following the switch to GPT-QModel as the backend in the previous release, AutoGPTQ is now formally deprecated (#3190) thanks to @Qubitium.

Benchmarks
Image generation benchmark

The method comparison suite now has an image generation benchmark (#3082). This complements the existing MetaMathQA benchmark and allows to compare PEFT methods on a completely different task. This gives us a more complete picture of the capabilities of different PEFT methods. To check the results, visit the Gradio Space and on the top left, select the task "image-gen".

A look at the results from the image generation benchmark

MetaMathQA

Since the last release, we also made several improvements to the MetaMathQA benchmark. You should check the updated results in the Gradio Space. With the help of the community, we now cover a much broader range of experiments. Moreover, we discovered a bug in the forgetting metric -- the new numbers now accurately reflect how much fine-tuning retains of the original knowledge of the base model.

PEFT Shop

#3317 adds a new Gradio Space with the goal of making it very easy to compare different PEFT methods not only with regard to performance but also capabilities (e.g. if quantization is supported). Give this Space a visit to find out which PEFT method is the best fit for your use case.

Browse the PEFT shop for the best PEFT method for your use case

All Changes
New Contributors

Full Changelog: https://github.com/huggingface/peft/compare/v0.19.0...v0.20.0

View originalPermalink
How v0.20.0 went

v0.19.0

Added 10
  • GraLoRA method that subdivides base weights into smaller blocks and applies LoRA to them for increased expressiveness and improved performance at higher ranks
  • BD-LoRA method that implements LoRA weights in block-diagonal form to reduce communication overhead when using tensor parallelism
  • Cartridges method to train a prefix that compresses long context to short size and save on tokens
  • PVeRA method as an extension of VeRA that adds probabilistic element by sampling from shared parameters
  • PSOFT method that improves efficiency of orthogonal fine-tuning by constraining adaptation to low-rank principal subspace
  • Lily method with sophisticated parameter sharing scheme where A parameters are shared blockwise and B parameters are chosen in data-dependent way through a router

Highlights

This PEFT release contains no less than nine new PEFT methods, described below. It also contains numerous enhancements that should make PEFT more useful to many users.

New Methods
GraLoRA

@yeonjoon-jung01 added "GraLoRA: Granular Low-Rank Adaptation for Parameter-Efficient Fine-Tuning" to PEFT (#2851). This method subdivides the base weight into smaller blocks and applies LoRA to those. This more granular adaptation promises to increase expressiveness and improve performance, especially at higher ranks (64+), closing the gap to full fine-tuning.

BD-LoRA

@Conzel contributed BD-LoRA: "Block-Diagonal LoRA for Eliminating Communication Overhead in Tensor Parallel LoRA Serving" (#2895). With BD-LoRA, the LoRA weights are implemented in a block-diagonal way. This allows to reduce communication overhead when using tensor parallelism (TP) and thus faster serving.

There is an experiment branch for BD-LoRA support in vLLM: vllm-project/vllm#28136.

Cartridges

Thanks to @kashif, PEFT now also supports Cartridges (#2953). The main purpose of this method is to train a prefix to compress a long context to a short size and thus save on tokens. On a low level, this is similar to prefix tuning. The PR also added an example recipe to quickly get started.

PVeRA

"PVeRA: Probabilistic Vector-Based Random Matrix Adaptation" was added to PEFT by @leofillioux in #2952. It is an extension of VeRA, a PEFT method that uses weight sharing between layers to be especially parameter efficient. PVeRA builds on top of that by adding a probabilistic element, sampling from the shared parameters and promising better performance overall.

PSOFT

@fei407 added PSOFT, "Efficient Orthogonal Fine-Tuning with Principal Subspace Adaptation", to PEFT in #3037. Orthogonal fine-tuning techniques like OFT and BOFT are good at preserving the structure and thus capabilities of the underlying base model. PSOFT improves efficiency of this technique by constraining the adaptation to low-rank principal subspace.

Lily

@yibozhong added Lily: "Low-Rank Interconnected Adaptation across Layers" to PEFT in #2563. Lily is on the surface similar to LoRA but has a sophisticated parameter sharing scheme. The A parameters are shared blockwise (e.g. 4 consecutive q_proj layers share the same A). There is a pool of B parameters that is shared globally, the actual B's are chosen in a data-dependent way through a router. This allows Lily to use higher ranks than LoRA while maintaining a low trainable parameter count.

PEANuT

In #3084, "PEANuT: Parameter-Efficient Adaptation with Weight-aware Neural Tweakers" was added to PEFT, again by @yibozhong. PEANuT adds a small, neural net (so called weight-aware neural tweakers) to the base model. Compared to LoRA, this increases expressivity for the same trainable parameter count or allows to greatly lower the parameter count without sacrificing expressivity. This comes at the expensive of a higher memory requirement for the same parameter count and decreased speed.

TinyLoRA

We have another serial contributor in @kashif, who also contributed TinyLoRA: "Learning to Reason in 13 Parameters" in #3024. This is a PEFT method that allows to train an extremely small number of parameters, much lower than what could be achieved even with LoRA rank 1. The paper shows that in particular with reinforcement learning, it can often be enough to train just a few parameters to achieve good results.

AdaMSS

@LonglongaaaGo added "AdaMSS: Adaptive Multi-Subspace Approach for Parameter-Efficient Fine-Tuning" to PEFT. This method segments the base weights of the model into smaller subspaces that are targeted for fine-tuning. Moreover, it's possible to dynamically assign a lower parameter budget to less important subspaces during training, similar to what AdaLoRA does. This promises to provide higher expressiveness and better generalization than similar PEFT methods.

Enhancements
Convert non-LoRA adapters to LoRA

In #2939, we added functions to PEFT to allow converting checkpoints of many non-LoRA methods into LoRA checkpoints. This can be useful because many other packages support only LoRA but not other PEFT methods, e.g. Diffusers and vLLM. With the new conversions tools, more PEFT methods than just LoRA can thus be used with those packages. Conversion is lossy but empirical testing showed that with a sufficiently high LoRA rank, the error can be quite low.

LoRA-GA

@sambhavnoobcoder added a new way to initialize LoRA weights with "LoRA-GA: Low-Rank Adaptation with Gradient Approximation" (#2926). This allows you to initialize the LoRA weights in a way that aligns the gradients with full fine-tuning and should lead to faster training convergence.

Reducing intruder dimensions

In "LoRA vs Full Fine-tuning: An Illusion of Equivalence", the authors showed that LoRA fine-tuning can introduce so-called "intruder dimensions" which contribute to forgetting. We now have a utility function to remove intruder dimension in PEFT, reduce_intruder_dimension. When calling this on a fine-tuned LoRA model, forgetting should be reduced while the fine-tuned task performance should remain almost the same.

Transformer Engine

In #3048, @balvisio added support for Transformer Engine, a quantization method by NVIDIA, to PEFT.

Tensor Parallel Support

In a series of PRs (#3079, #3091, #3096), @michaelbenayoun added support for Tensor Parallelism to LoRA.

Weight tying improvements

In many LLMs, the embedding and the LM head have tied weights to save on parameter count. This can, however, lead to tricky situations when trying to fine-tune those layers. Through a series of PRs (#2803, #2922, #2870, #2879, #3126), we improved the user experience when doing so. Most notably, users can now pass ensure_weight_tying=True to their PEFT config to force weight tying to be upheld. Please check the PEFT weight tying docs for how weight tying is now being handled. Thanks to @romitjain, @sambhavnoobcoder, and @Cursx for their contributions.

Low precsion floating type support

#3055 makes LoRA work with base models that use very low precision floats like torch.float8_e4m3fn. An example of that would be MiniMax-M2.5.

Zero init for PrefixTuning

#3128 introduces zero init to Prefix Tuning which, according to our benchmarks, reduced the result variance significantly and yielded good task accuracy without the need for prompt engineering.

LoftQ + int8 quantization

With #3088 the LoftQ implementation now supports correcting errors for int8 quantization without utilizing activation thresholding alongside the already existing nf4 quantization.

Changes
Removal of Bone

The Bone PEFT method was removed in #3115. Users are directed to use MiSS instead, which is the improved replacement for Bone. Use this Bone-to-MiSS conversion script if you want to port old Bone checkpoints.

AutoGPTQ and AutoAWQ

These two quantization methods now use GPTQModel as their backend (#2932) thanks to @ZX-ModelCloud.

Handling of requires_grad in modules_to_save

Previously, PEFT would enable requires_grad on the original module if the corresponding modules_to_save was disabled. This is almost never desirable and was thus fixed. Although this change is technically backwards-incompatible, it's an extreme niche case, so we don't expect any user to be negatively affected by it.

All Changes

New Contributors

Full Changelog: https://github.com/huggingface/peft/compare/v0.18.1...v0.19.0

View originalPermalink
How v0.19.0 went
v0.18.1

0.18.1

Fixed 3
  • Small fixes required for some special cases to work with the upcoming transformers v5 release
  • Enable PEFT to run with AMD ROCm
  • Fix a regression that inadvertently required transformers >= 4.52

Small patch release containing the following changes:

  • #2934: Small fixes required for some special cases to work with the upcoming transformers v5 release
  • #2963: Fix to enable PEFT to run with AMD ROCm thanks to @vladmandic
  • #2976: Fix a regression that inadvertently required transformers >= 4.52
View originalPermalink
How v0.18.1 went
v0.18.0

0.18.0: RoAd, ALoRA, Arrow, WaveFT, DeLoRA, OSF, and more

Added 10
  • Added RoAd (2D Rotary Adaptation) method that learns 2D rotation matrices applied via element-wise multiplication and supports mixed adapter batches
  • Added Activated LoRA (ALoRA) technique for causal language models allowing selective enabling of LoRA adapters based on token invocation sequences
  • Added Arrow dynamic routing algorithm for switching between multiple loaded LoRAs
  • Added GenKnowSub technique built upon Arrow that modifies the library of LoRAs by subtracting general knowledge adapters
  • Added Wavelet Fine-Tuning (WaveFT) method that trains sparse updates in the wavelet domain of residual matrices
  • Added DeLoRA (Decoupled Low-rank Adaptation) method that decouples angle and magnitude of learned adapter weights with norm constraints via delora_lambda parameter

Highlights

FIXME update list of all changes, so some more commits were added

New Methods
RoAd

@ppetrushkov added RoAd: 2D Rotary Adaptation to PEFT in #2678. RoAd learns 2D rotation matrices that are applied using only element-wise multiplication, thus promising very fast inference with adapters in unmerged state.

Remarkably, besides LoRA, RoAd is the only PEFT method that supports mixed adapter batches. This means that when you have loaded a model with multiple RoAd adapters, you can use all of them for different samples in the same batch, which is much more efficient than switching adapters between batches:

model = PeftModel.from_pretrained(base_model, <path-to-road-adapter-A>, adapter_name="adapter-A")
model.add_adapter("adapter-B", <path-to-road-adapter-B>)

inputs = ...  # input with 3 samples
# apply adapter A to sample 0, adapter B to sample 1, and use the base model for sample 2:
adapter_names = ["adapter-A", "adapter-B", "__base__"]
output_mixed = model(**inputs, adapter_names=adapter_names)
gen_mixed = model.generate(**inputs, adapter_names=adapter_names)
ALoRA

Activated LoRA is a technique added by @kgreenewald in #2609 for causal language models, allowing to selectively enable LoRA adapters depending on a specific token invocation sequence in the input. This has the major benefit of being able to re-use most of the KV cache during inference when the adapter is only used to generate part of the response, after which the base model takes over again.

Arrow & GenKnowSub

@TheTahaaa contributed not only support for Arrow, a dynamic routing algorithm between multiple loaded LoRAs in #2644, but also GenKnowSub, a technique built upon Arrow where the 'library' of LoRAs available to Arrow is first modified by subtracting general knowledge adapters (e.g., trained on subsets of Wikipedia) to enhance task-specific performance.

WaveFT

Thanks to @Bilican, Wavelet Fine-Tuning (WaveFT) was added to PEFT in #2560. This method trains sparse updates in the wavelet domain of residual matrices, which is especially parameter efficient. It is very interesting for image generation, as it promises to generate diverse outputs while preserving subject fidelity.

DeLoRA

Decoupled Low-rank Adaptation (DeLoRA) was added by @mwbini in #2780. This new PEFT method is similar to DoRA in so far as it decouples the angle and magnitude of the learned adapter weights. However, DeLoRA implements this in a way that promises to better prevent divergence. Moreover, it constrains the deviation of the learned weight by imposing an upper limit of the norm, which can be adjusted via the delora_lambda parameter.

OSF

Orthogonal Fine-Tuning (OSF) was added by @NikhilNayak-debug in #2685. By freezing the high-rank subspace of the targeted weight matrices and projecting gradient updates to a low-rank subspace, OSF achieves good performance on continual learning tasks. While it is a bit memory intensive for standard fine-tuning processes, it is definitely worth checking out on tasks where performance degradation of previously learned tasks is a concern.

Enhancements
Text generation benchmark

In #2525, @ved1beta added the text generation benchmark to PEFT. This is a framework to determine and compare metrics with regard to text generation of different PEFT methods, e.g. runtime and memory usage. Right now, this benchmark is still lacking experimental settings and a visualization, analogous to what we have in the MetaMathQA benchmark. If this is something that interests you, we encourage you to let us know or, even better, contribute to this benchmark.

Reliable interface for integrations

PEFT has integrations with other libraries like Transformers and Diffusers. To facilitate this integration, PEFT now provides a stable interface of functions that should be used if applicable. For example, the set_adapter function can be used to switch between PEFT adapters on the model, even if the model is not a PeftModel instance. We commit to keeping these functions backwards compatible, so it's safe for other libraries to build on top of those.

Handling of weight tying

Some Transformers models can have tied weights. This is especially prevalent when it comes to the embedding and the LM head. Currently, the way that this is handled in PEFT is not obvious. We thus drafted an issue to illustrate the intended behavior in #2864. This shows what our goal is, although not everything is implemented yet.

In #2803, @romitjain added the ensure_weight_tying argument to LoraConfig. This argument, if set to True, enforces weight tying of the modules targeted with modules_to_save. Thus, if embedding and LM head are tied, they will share weights, which is important to allow, for instance, weight merging. Therefore, for most users, we recommend to enable this setting if they want to fully fine-tune the embedding and LM head. For backwards compatability, the setting is off by default though.

Note that in accordance with #2864, the functionality of ensure_weight_tying=True will be expanded to also include trainable tokens (#2870) and LoRA (tbd.) in the future.

Support Conv1d and 1x1 Conv2 layers in LoHa and LoKr

@grewalsk extended LoHa and LoKr to support nn.Conv1d layers, as well as nn.Conv2d with 1x1 kernels, in #2515.

New prompt tuning initialization

Thanks to @macmacmacmac, we now have a new initialization option for prompt tuning, random discrete initialization (#2815). This option should generally work better than random initialization, as corroborated on our PEFT method comparison suite. Give it a try if you use prompt tuning.

Combining LoRA adapters with negative weights

If you use multiple LoRA adapters, you can merge them into a single adapter using model.add_weighted_adapter. However, so far, this only worked with positive weights per adapter. Thanks to @sambhavnoobcoder and @valteu, it is now possible to pass negative weights too.

Changes
Transformers compatibility

At the time of writing, the Transformers v5 release is imminent. This Transformers version will be incomptabile with PEFT < 0.18.0. If you plan to use Transformers v5 with PEFT, please upgrade PEFT to 0.18.0+.

Python version

This PEFT version no longer supports Python 3.9, which has reached its end of life. Please use Python 3.10+.

Updates to OFT

The OFT method has been updated to make it slightly faster and to stabilize the numerics in #2805. This means, however, that existing checkpoints may give slightly different results after upgrading to PEFT 0.18.0. Therefore, if you use OFT, we recommend to retrain the adapter.

All Changes

New Contributors

Full Changelog: https://github.com/huggingface/peft/compare/v0.17.1...v0.18.0

View originalPermalink
How v0.18.0 went
v0.17.1

0.17.1

Changed 1
  • PEFT now raises an error when attempting to add multiple adapters to prevent incorrect behavior
Fixed 1
  • PEFT no longer removes existing parametrizations from the parameter when using the target_parameters feature

This patch release contains a few fixes (via #2710) for the newly introduced target_parameters feature, which allows LoRA to target nn.Parameters directly (useful for mixture of expert layers). Most notably:

  • PEFT no longer removes possibly existing parametrizations from the parameter.
  • Adding multiple adapters (via model.add_adapter or model.load_adapter) did not work correctly. Since a solution is not trivial, PEFT now raises an error to prevent this situation.
View originalPermalink
How v0.17.1 went
v0.17.0

0.17.0: SHiRA, MiSS, LoRA for MoE, and more

Added 4
  • Add Sparse High Rank Adapters (SHiRA) method for improved performance over LoRAs, especially when using multiple adapters
  • Add MiSS (Matrix Shard Sharing) method as an evolution of Bone with improved performance and memory efficiency
  • Enable LoRA to target nn.Parameter directly using target_parameters config attribute, useful for models with Mixture of Expert layers
  • Support injecting adapters based on a state_dict without needing to specify target_modules in the config
Fixed 5
  • Fix bug in prompt learning methods where modules_to_save was ignored, preventing classification layers from being trained or stored
  • Fix create mask function signature change in transformers 4.53.1
  • Fix faulty OFT parameter device test
  • Allow peft_type to be a string
  • Fix prefix tuning after transformers PR 38635
Deprecated 1
  • Deprecate Bone method in favor of MiSS, to be removed in PEFT v0.19.0

Highlights

New Methods
SHiRA

@kkb-code contributed Sparse High Rank Adapters (SHiRA, paper) which promise to offer a potential gain in performance over LoRAs - especially the concept loss when using multiple adapters is improved. Since the adapters only train on 1-2% of the weights and are inherently sparse, switching between adapters may be cheaper than with LoRAs. (#2584)

MiSS

@JL-er added a new PEFT method, MiSS (Matrix Shard Sharing) in #2604. This method is an evolution of Bone, which, according to our PEFT method comparison benchmark, gives excellent results when it comes to performance and memory efficiency. If you haven't tried it, you should do so now.

At the same time, Bone will be deprecated in favor of MiSS and will be removed in PEFT v0.19.0. If you already have a Bone checkpoint, you can use scripts/convert-bone-to-miss.py to convert it into a MiSS checkpoint and proceed with training using MiSS.

Enhancements
LoRA for nn.Parameter

LoRA is now able to target nn.Parameter directly (#2638, #2665)! Ever had this complicated nn.Module with promising parameters inside but it was too custom to be supported by your favorite fine-tuning library? No worries, now you can target nn.Parameters directly using the target_parameters config attribute which works similarly to target_modules.

This option can be especially useful for models with Mixture of Expert (MoE) layers, as those often use nn.Parameters directly and cannot be targeted with target_modules. For example, for the Llama4 family of models, use the following config to target the MoE weights:

config = LoraConfig(
    ...,
    target_modules=[],  # <= prevent targeting any modules
    target_parameters=["feed_forward.experts.down_proj", "feed_forward.experts.gate_up_proj"],
)

Note that this feature is still experimental as it comes with a few caveats and therefore might change in the future. Also, MoE weights with many experts can be quite huge, so expect a higher memory usage than compared to targeting normal nn.Linear layers.

Injecting adapters based on a state_dict

Sometimes, it is possible that there is a PEFT adapter checkpoint but the corresponding PEFT config is not known for whatever reason. To inject the PEFT layers for this checkpoint, you would usually have to reverse-engineer the corresponding PEFT config, most notably the target_modules argument, based on the state_dict from the checkpoint. This can be cumbersome and error prone. To avoid this, it is also possible to call inject_adapter_in_model and pass the loaded state_dict as an argument:

from safetensors.torch import load_file
from peft import LoraConfig, inject_adapter_in_model

model = ...
state_dict = load_file(<path-to-safetensors-file>)
lora_config = LoraConfig()  # <= no need to specify further
model = inject_adapter_in_model(lora_config, model, state_dict=state_dict)

Find more on state_dict based injection in the docs.

Changes
Compatibility

A bug in prompt learning methods caused modules_to_save to be ignored. Especially classification tasks are affected since they usually add the classification/score layer to modules_to_save. In consequence, these layers were neither trained nor stored after training. This has been corrected now. (#2646)

All Changes
New Contributors

Full Changelog: https://github.com/huggingface/peft/compare/v0.16.0...v0.17.0

View originalPermalink
How v0.17.0 went
v0.16.0

0.16.0: LoRA-FA, RandLoRA, C³A, and much more

Added 8
  • Add LoRA-FA optimizer based on AdamW to increase memory efficiency of LoRA training
  • Add RandLoRA method that uses non-learnable random low rank matrices combined through learnable matrices, with support for quantized models with bitsandbytes
  • Add Circular Convolution Adaptation (C3A) as a new PEFT method to overcome low rank adaptation limits
  • Add support for LoRA with Conv2d layers where groups != 1 (requires rank r divisible by groups)
  • Add Intel Neural Compressor (INC) quantization support to LoRA
  • Add DoRA support for Conv1d layers
  • Add orthogonal weight initialization for LoRA via init_lora_weights='orthogonal'
  • Add Quantization-Aware LoRA training with GPTQ support
Changed 4
  • Refactor Orthogonal Finetuning (OFT) for improved speed and memory efficiency, but this is incompatible with old OFT checkpoints
  • Set required_grads_ of modules_to_save to True when used directly with inject_adapter for PEFT integrations
  • Update compatibility with Transformers vision language model refactor; apply prompt learning methods to vlm directly instead of vlm.language_model
  • Update prompt learning methods for Transformers attention mask refactoring, which may result in numerical differences for prefix tuning and affect models using 4d attention masks like Gemma
Fixed 1
  • Fix LoRA hotswapping with compiled models to no longer cause CUDA graph re-records

Highlights

peft-v0 16 0

New Methods
LoRA-FA

In #2468, @AaronZLT added the LoRA-FA optimizer to PEFT. This optimizer is based on AdamW and it increases memory efficiency of LoRA training. This means that you can train LoRA with less memory, or, with the same memory budget, use higher LoRA ranks, potentially getting better results.

RandLoRA

Thanks to @PaulAlbert31, a new PEFT method called RandLoRA was added to PEFT (#2464). Similarly to VeRA, it uses non-learnable random low rank matrices that are combined through learnable matrices. This way, RandLoRA can approximate full rank updates of the weights. Training models quantized with bitsandbytes is supported.

C³A

@Phoveran added Circular Convolution Adaptation, C3A, in #2577. This new PEFT method can overcome the limit of low rank adaptations as seen e.g. in LoRA while still promising to be fast and memory efficient.

Enhancements

Thanks to @gslama12 and @SP1029, LoRA now supports Conv2d layers with groups != 1. This requires the rank r being divisible by groups. See #2403 and #2567 for context.

@dsocek added support for Intel Neural Compressor (INC) quantization to LoRA in #2499.

DoRA now supports Conv1d layers thanks to @EskildAndersen (#2531).

Passing init_lora_weights="orthogonal" now enables orthogonal weight initialization for LoRA (#2498).

@gapsong brought us Quantization-Aware LoRA training in #2571. This can make QLoRA training more efficient, please check the included example. Right now, only GPTQ is supported.

There has been a big refactor of Orthogonal Finetuning, OFT, thanks to @zqiu24 (#2575). This makes the PEFT method run more quickly and require less memory. It is, however, incompatible with old OFT checkpoints. If you have old OFT checkpoints, either pin the PEFT version to <0.16.0 or retrain it with the new PEFT version.

Thanks to @keepdying, LoRA hotswapping with compiled models no longer leads to CUDA graph re-records (#2611).

Changes

Compatibility
  • #2481: The value of required_grads_ of modules_to_save is now set to True when used directly with inject_adapter. This is relevant for PEFT integrations, e.g. Transformers or Diffusers.
  • Due to a big refactor of vision language models (VLMs) in Transformers, the model architecture has been slightly adjusted. One consequence of this is that if you use a PEFT prompt learning method that is applied to vlm.language_model, it will no longer work, please apply it to vlm directly (see #2554 for context). Morever, the refactor results in different checkpoints. We managed to ensure backwards compatability in PEFT, i.e. old checkpoints can be loaded successfully. There is, however, no forward compatibility, i.e. loading checkpoints trained after the refactor is not possible with package versions from before the refactor. In this case, you need to upgrade PEFT and transformers. More context in #2574.
  • #2579: There have been bigger refactors in Transformers concerning attention masks. This required some changes on the PEFT side which can affect prompt learning methods. For prefix tuning specifically, this can result in numerical differences but overall performance should be the same. For other prompt learning methods, numerical values should be the same, except if the base model uses 4d attention masks, like Gemma. If you load old prompt learning checkpoints, please double-check that they still perform as expected, especially if they're trained on Gemma or similar models. If not, please re-train them or pin PEFT and transformers to previous versions (<0.16.0 and <4.52.0, respectively).
All Changes
New Contributors

Full Changelog: https://github.com/huggingface/peft/compare/v0.15.2...v0.16.0

View originalPermalink
How v0.16.0 went

v0.15.2

Fixed 1
  • fix a bug that prevented prompt learning methods like P-tuning from working

This patch fixes a bug that resulted in prompt learning methods like P-tuning not to work (#2477).

View originalPermalink
How v0.15.2 went

v0.15.1

Fixed 1
  • Fix handling of modules_to_save when used in conjunction with DeepSpeed ZeRO stage 3 to prevent those modules from being placeholder values in saved checkpoints

This patch includes a fix for #2450. In this bug modules_to_save was not handled correctly when used in conjunction with DeepSpeed ZeRO stage 3 which resulted in those modules being placeholder values in the saved checkpoints.

Full Changelog: https://github.com/huggingface/peft/compare/v0.15.0...v0.15.1

View originalPermalink
How v0.15.1 went

More in AI

View all
codex

codex release notes.

codex
Ollama

Get up and running with large language models locally.

Ollama
openclaw

Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

openclaw
superpowers

An agentic skills framework & software development methodology that works.

obra
ECC

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

affaan-m
hermes-agent

The agent that grows with you

NousResearch

Discussion