Papers › Riemannian Preconditioned LoRA for Fine-Tuning Foundation Models

Riemannian Preconditioned LoRA for Fine-Tuning Foundation Models

4 Feb 2024arXiv:2402.02347archive 2025-07-28

Fangzhao Zhang, Mert Pilanci

Low-Rank Adaptation (LoRA) emerges as a popular parameter-efficient fine-tuning (PEFT) method, which proposes to freeze pretrained model weights and update an additive low-rank trainable matrix. In this work, we study the enhancement of LoRA training by introducing an r ×r preconditioner in each gradient step where r is the LoRA rank. We theoretically verify that the proposed preconditioner stabilizes feature learning with LoRA under infinite-width NN setting. Empirically, the implementation of this new preconditioner requires a small change to existing optimizer code and creates virtually minuscule storage and runtime overhead. Our experimental results with both large language models and text-to-image diffusion models show that with this new preconditioner, the convergence and reliability of SGD and AdamW can be significantly enhanced. Moreover, the training process becomes much more robust to hyperparameter choices such as learning rate. The new preconditioner can be derived from a novel Riemannian metric in low-rank matrix field. Code can be accessed at https://github.com/pilancilab/Riemannian_Preconditioned_LoRA.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 21 code samples: 18 from 1 repository linked to this paper, 3 identical to code harvested from another paper's repository. Across all 21: 15 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 2 honoured, 1 violated, 6 with no contract checked; 6 where Syntology's instrument failed) · 6 unverified. Where Syntology's instrument failed: 5 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 18 samples from 1 repository, 12 ran (9 with no instrument failure; 3 where Syntology's instrument failed); 3 identical to code first harvested elsewhere. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 5 of the 15 samples that ran have no recorded date.

pilancilab/Riemannian_Preconditioned_LoRA officialmentioned in papermentioned on GitHubpytorchMITSyntology: 18 samples, 12 ran (3 instrument-failed) · 6 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

21 samples harvested: 15 ran (of which 0 constructed an object rather than computing a result; 9 with no instrument failure: 2 honoured, 1 violated, 6 with no contract checked; 6 where Syntology's instrument failed) · 6 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

2ran · honoured contract
1ran · violated contract
5ran · our draft was wrong
1ran; the contract check could not run
6ran
6unverified

Licence: 3 of the 21 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from pilancilab/Riemannian_Preconditioned_LoRA. 3 samples are listed last, in a group without a repository. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

pilancilab/Riemannian_Preconditioned_LoRA official repository 18 samples, 12 ran (3 instrument-failed) · 6 unverified
convert_vae_state_dict Object_Generation/lora_diffusion/to_ckpt_v2.py ran · our draft was wrong MIT (permissive) · 0e6846a624e7d973 · report fixture: signature vae_state_dict:dict/0/? · returned dict · level recorded 2026-09-24
image_grid in 41 papers Object_Generation/lora_diffusion/utils.py ran MIT (permissive) · 2dda05eee1b6e0f7 · report fixture: signature _imgs:list/1/Image, rows:NoneType/0/?, cols:NoneType/0/? · returned Image · level recorded 2026-09-24
inject_trainable_lora in 12 papers Object_Generation/lora_diffusion/lora.py ran · our draft was wrong MIT (permissive) · 4b4a8984a2737ae5 · report fixture: signature model:callable/0/?, target_replace_module:set/0/?, r:int/0/int, loras:NoneType/0/?, verbose:bool/0/bool, dropout_p:float/0/float, scale:float/0/float · returned tuple · level recorded, date not recorded
inject_trainable_lora_extended in 11 papers Object_Generation/lora_diffusion/lora.py ran · honoured contract MIT (permissive) · 3aec51ff694c19af · report fixture: signature model:callable/0/?, target_replace_module:set/0/?, r:int/0/int, loras:NoneType/0/? · returned tuple · level recorded, date not recorded
inpainting_dataloader Object_Generation/lora_diffusion/cli_lora_pti.py ran MIT (permissive) · d793962f13865771 · report fixture: signature train_dataset:InpaintingDataset/0/?, train_batch_size:int/0/int, tokenizer:MinimalTokenizer/0/?, vae:callable/0/?, text_encoder:callable/0/? · returned DataLoader · level recorded 2026-09-22
prepare_clip_model_sets Object_Generation/lora_diffusion/utils.py ran; the contract check could not run (RAISES) MIT (permissive) · 298713e11f263fda · report fixture: signature eval_clip_id:str/0/str · returned tuple · level recorded 2026-09-24
reshape_weight_for_sd Object_Generation/lora_diffusion/to_ckpt_v2.py ran · honoured contract fingerprinted MIT (permissive) · a1d506dd99d38923 · report fixture: signature w:Tensor/2/float · returned Tensor · level recorded 2026-09-24
safe_open Object_Generation/lora_diffusion/safe_open.py ran MIT (permissive) · f999bed53b70c3ab · report fixture: signature filename:str/0/str, framework:str/0/str, device:str/0/str · returned SafetensorsWrapper · level recorded 2026-09-24
swin_ir_sr Object_Generation/lora_diffusion/preprocess_files.py ran MIT (permissive) · ec4e915ba607e9c9 · report fixture: signature images:list/1/Image, model_id:str/0/str, target_size:NoneType/0/?, device:device/0/? · returned list · level recorded 2026-09-24
test_xformers_backwards Object_Generation/lora_diffusion/xformers_utils.py ran · violated contract fingerprinted MIT (permissive) · bdf6d68955356de7 · report fixture: signature size:int/0/int · returned bool · level recorded 2026-09-24
text2img_dataloader Object_Generation/lora_diffusion/cli_lora_pti.py ran MIT (permissive) · aa63f591fb2e9bd7 · report fixture: signature train_dataset:MockDataset/0/?, train_batch_size:int/0/int, tokenizer:MockTokenizer/0/?, vae:callable/0/?, text_encoder:callable/0/?, cached_latents:bool/0/bool · returned DataLoader · level recorded 2026-09-22
text_img_alignment Object_Generation/lora_diffusion/utils.py ran MIT (permissive) · 8dbc9878c5043ebc · report fixture: signature img_embeds:Tensor/2/float, text_embeds:Tensor/2/float, target_img_embeds:Tensor/2/float · returned dict · level recorded 2026-09-24
blip_captioning_dataset Object_Generation/lora_diffusion/preprocess_files.py unverified MIT (permissive) · cbef98786d899b3b · report
clipseg_mask_generator Object_Generation/lora_diffusion/preprocess_files.py unverified MIT (permissive) · 98530605dba303f0 · report
convert_unet_state_dict Object_Generation/lora_diffusion/to_ckpt_v2.py unverified MIT (permissive) · f2b2f19d941175af · report
create_tensor Object_Generation/lora_diffusion/safe_open.py unverified MIT (permissive) · effb6d8bd6b936e2 · report
extract_lora_ups_down in 9 papers Object_Generation/lora_diffusion/lora.py unverified MIT (permissive) · 2f9880b526ff277e · report
lora_join Object_Generation/lora_diffusion/lora_manager.py unverified MIT (permissive) · 47566c65ffdc7fa2 · report
Repository not located 3 samples, 3 ran (3 instrument-failed) · 0 unverified
3 samples are identical to code Syntology harvested from another paper's repository; this paper's own copy was not located, so no repository is named and no licence is recorded.
gelu in 308 papers
census: 43 implementations from 74 papers, 6 distinct outputs on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32
repository not located ran · our draft was wrong fingerprinted licence of this copy not recorded · 8d23fbe2b99b840b · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded, date not recorded
gelu_fast in 39 papers repository not located ran · our draft was wrong fingerprinted licence of this copy not recorded · a4475703ff58ecf9 · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded, date not recorded
gelu_new in 55 papers
census: 3 implementations from 7 papers, 1 distinct output on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32
repository not located ran · our draft was wrong fingerprinted licence of this copy not recorded · dc9ffc0f29e7fa3d · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded, date not recorded

Tasks

parameter-efficient fine-tuning

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

AdamWDiffusionSGD

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections