List your company

Image Classification on SUN397 (Top-1 Accuracy, Delta)

80.3Top-1 AccuracyWUDI-Merging
61.00866.016571.02576.0335Mar 9, 2026Mar 25, 2026Apr 10, 2026Apr 27, 2026May 13, 2026May 29, 2026Jun 15, 2026

Evaluation results

MethodLinks
WUDI-MergingMerging Method=WUDI-Merging, Quantization Method=None, Bit-width=Full-precision, Model Backbone=CLIP-ViT-L/142026.0580.3—
WUDI-Merging w/ E-PMQMerging Method=WUDI-Merging, Quantization Method=E-PMQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0579.9—
WUDI-Merging w/ GPTQMerging Method=WUDI-Merging, Quantization Method=GPTQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0579.1—
WUDI-Merging w/ RTNMerging Method=WUDI-Merging, Quantization Method=RTN, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0578.9—
SDS-LoRAModel=ViT-Large, Rank=82026.0676.58—
DoRAModel=ViT-Large, Rank=82026.0676.5—
SDS-LoRAModel=ViT-Large, Rank=322026.0676.46—
rsLoRAModel=ViT-Large, Rank=82026.0676.45—
LoRAModel=ViT-Large, Rank=82026.0676.3—
DoRAModel=ViT-Large, Rank=322026.0676.23—
rsLoRAModel=ViT-Large, Rank=322026.0676.22—
Simple Averaging w/ E-PMQMerging Method=Simple Averaging, Quantization Method=E-PMQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0576.2—
Full FT.Model=ViT-Large, Rank=-2026.0676.2—
TIES-Merging w/ E-PMQMerging Method=TIES-Merging, Quantization Method=E-PMQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0576—
Task Arithmetic w/ E-PMQMerging Method=Task Arithmetic, Quantization Method=E-PMQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0575.9—
PiSSAModel=ViT-Large, Rank=82026.0675.41—
LoRA+Model=ViT-Large, Rank=322026.0675.19—
LoRA+Model=ViT-Large, Rank=82026.0675.08—
Fine-tuned (STL)Backbone=CLIP-ViT-B/32, Bit-width=Full-precision, Training=Single-task learning2026.0575—
SDS-LoRARank=32, Backbone=ViT-Base2026.0674.89—
SDS-LoRAModel=ViT-Base, Rank=322026.0674.89—
SDS-LoRARank=8, Backbone=ViT-Base2026.0674.85—
SDS-LoRAModel=ViT-Base, Rank=82026.0674.85—
Full FT.Rank=-, Backbone=ViT-Base2026.0674.76—
Full FT.Model=ViT-Base, Rank=-2026.0674.76—
TIES-MergingMerging Method=TIES-Merging, Quantization Method=None, Bit-width=Full-precision, Model Backbone=CLIP-ViT-L/142026.0574.7—
rsLoRARank=8, Backbone=ViT-Base2026.0674.63—
rsLoRAModel=ViT-Base, Rank=82026.0674.63—
LoRAModel=ViT-Large, Rank=322026.0674.59—
DoRARank=8, Backbone=ViT-Base2026.0674.46—
DoRAModel=ViT-Base, Rank=82026.0674.46—
LoRARank=8, Backbone=ViT-Base2026.0674.35—
LoRAModel=ViT-Base, Rank=82026.0674.35—
BiCLIPShots=16, Base Model=CLIP, Adaptation Method=Bilinear adaptation2026.0374.2710.77
BiSigLIPShots=16, Base Model=SigLIP, Adaptation Method=Bilinear adaptation2026.0374.248.38
WUDI-Merging w/ AWQMerging Method=WUDI-Merging, Quantization Method=AWQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0574.1—
DoRARank=32, Backbone=ViT-Base2026.0674.05—
DoRAModel=ViT-Base, Rank=322026.0674.05—
TIES-Merging w/ GPTQMerging Method=TIES-Merging, Quantization Method=GPTQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0574—
rsLoRARank=32, Backbone=ViT-Base2026.0673.92—
rsLoRAModel=ViT-Base, Rank=322026.0673.92—
PiSSAModel=ViT-Large, Rank=322026.0673.84—
TIES-Merging w/ RTNMerging Method=TIES-Merging, Quantization Method=RTN, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0573.8—
LoRA+Rank=8, Backbone=ViT-Base2026.0672.87—
LoRA+Model=ViT-Base, Rank=82026.0672.87—
PiSSARank=8, Backbone=ViT-Base2026.0672.82—
PiSSAModel=ViT-Base, Rank=82026.0672.82—
Simple AveragingMerging Method=Simple Averaging, Quantization Method=None, Bit-width=Full-precision, Model Backbone=CLIP-ViT-L/142026.0572.5—
Simple Averaging w/ GPTQMerging Method=Simple Averaging, Quantization Method=GPTQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0572—
Task ArithmeticMerging Method=Task Arithmetic, Quantization Method=None, Bit-width=Full-precision, Model Backbone=CLIP-ViT-L/142026.0572—
Simple Averaging w/ RTNMerging Method=Simple Averaging, Quantization Method=RTN, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0571.6—
Task Arithmetic w/ GPTQMerging Method=Task Arithmetic, Quantization Method=GPTQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0571.3—
Task Arithmetic w/ RTNMerging Method=Task Arithmetic, Quantization Method=RTN, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0571—
LoRARank=32, Backbone=ViT-Base2026.0670.82—
LoRAModel=ViT-Base, Rank=322026.0670.82—
TIES-Merging w/ AWQMerging Method=TIES-Merging, Quantization Method=AWQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0570.4—
LR-LoRABackbone=DINOv2 ViT-B/14, Rank=16, Adapter Placement=attention and MLP projections, Evaluation Seeds=32026.0670.2—
PiSSARank=32, Backbone=ViT-Base2026.0669.73—
PiSSAModel=ViT-Base, Rank=322026.0669.73—
RandLoRABackbone=DINOv2 ViT-B/14, Rank=16, Adapter Placement=attention and MLP projections, Evaluation Seeds=32026.0669.3—
LoRABackbone=DINOv2 ViT-B/14, Rank=16, Adapter Placement=attention and MLP projections, Evaluation Seeds=32026.0668.9—
VeRABackbone=DINOv2 ViT-B/14, Rank=16, Adapter Placement=attention and MLP projections, Evaluation Seeds=32026.0668.5—
WUDI-MergingBackbone=CLIP-ViT-B/32, Merging Strategy=WUDI-Merging, Bit-width=Full-precision2026.0568—
WUDI-Merging w/ E-PMQBackbone=CLIP-ViT-B/32, Merging Strategy=WUDI-Merging, Quantization Algorithm=E-PMQ, Bit-width=4-bit2026.0567.9—
Simple Averaging w/ E-PMQBackbone=CLIP-ViT-B/32, Merging Strategy=Simple Averaging, Quantization Algorithm=E-PMQ, Bit-width=4-bit2026.0567.7—
LoRA+Rank=32, Backbone=ViT-Base2026.0667.66—
LoRA+Model=ViT-Base, Rank=322026.0667.66—
TIES-Merging w/ E-PMQBackbone=CLIP-ViT-B/32, Merging Strategy=TIES-Merging, Quantization Algorithm=E-PMQ, Bit-width=4-bit, Lambda=0.32026.0567.6—
WUDI-Merging w/ E-PMQMerging Strategy=WUDI-Merging, Quantization Method=E-PMQ, Model Backbone=CLIP-ViT-B/32, Bit-width=4-bit2026.0567.44—
TIES-MergingBackbone=CLIP-ViT-B/32, Merging Strategy=TIES-Merging, Bit-width=Full-precision, Lambda=0.32026.0567.1—
Task Arithmetic w/ E-PMQBackbone=CLIP-ViT-B/32, Merging Strategy=Task Arithmetic, Quantization Algorithm=E-PMQ, Bit-width=4-bit, Lambda=0.32026.0567—
WUDI-Merging w/ GPTQBackbone=CLIP-ViT-B/32, Merging Strategy=WUDI-Merging, Quantization Algorithm=GPTQ, Bit-width=4-bit2026.0566.6—
Simple Averaging w/ AWQMerging Method=Simple Averaging, Quantization Method=AWQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0566.6—
SigLIPShots=0, Base Model=SigLIP, Adaptation Method=Zero-Shot2026.0365.85—
WUDI-MergingMerging Strategy=WUDI-Merging, Quantization Method=None, Model Backbone=CLIP-ViT-B/32, Bit-width=Full2026.0565.7—
Simple Averaging w/ E-PMQMerging Strategy=Simple Averaging, Quantization Method=E-PMQ, Model Backbone=CLIP-ViT-B/32, Bit-width=4-bit2026.0565.65—
TIES-Merging w/ GPTQBackbone=CLIP-ViT-B/32, Merging Strategy=TIES-Merging, Quantization Algorithm=GPTQ, Bit-width=4-bit, Lambda=0.32026.0565.6—
Simple AveragingBackbone=CLIP-ViT-B/32, Merging Strategy=Simple Averaging, Bit-width=Full-precision2026.0565.4—
TIES-Merging w/ E-PMQMerging Strategy=TIES-Merging, Quantization Method=E-PMQ, Model Backbone=CLIP-ViT-B/32, Bit-width=4-bit2026.0565.28—
Simple AveragingMerging Strategy=Simple Averaging, Quantization Method=None, Model Backbone=CLIP-ViT-B/32, Bit-width=Full2026.0564.8—
Simple Averaging + E-PMQMerging Strategy=Simple Averaging, Quantization Method=E-PMQ, Bit-width=4-bit, Backbone=CLIP-ViT-B/322026.0564.59—
WUDI-Merging + E-PMQMerging Strategy=WUDI-Merging, Quantization Method=E-PMQ, Bit-width=4-bit, Backbone=CLIP-ViT-B/322026.0564.37—
Simple Averaging w/ GPTQBackbone=CLIP-ViT-B/32, Merging Strategy=Simple Averaging, Quantization Algorithm=GPTQ, Bit-width=4-bit2026.0564.3—
Task Arithmetic w/ E-PMQMerging Strategy=Task Arithmetic, Quantization Method=E-PMQ, Model Backbone=CLIP-ViT-B/32, Bit-width=4-bit2026.0564.23—
Simple AveragingMerging Strategy=Simple Averaging, Quantization Method=Full-precision, Bit-width=Full, Backbone=CLIP-ViT-B/322026.0564.2—
TIES-Merging + E-PMQMerging Strategy=TIES-Merging, Quantization Method=E-PMQ, Bit-width=4-bit, Backbone=CLIP-ViT-B/322026.0564.04—
Simple Averaging w/ GPTQMerging Strategy=Simple Averaging, Quantization Method=GPTQ, Model Backbone=CLIP-ViT-B/32, Bit-width=4-bit2026.0563.89—
WUDI-Merging w/ GPTQMerging Strategy=WUDI-Merging, Quantization Method=GPTQ, Model Backbone=CLIP-ViT-B/32, Bit-width=4-bit2026.0563.77—
CLIPShots=0, Base Model=CLIP, Adaptation Method=Zero-Shot2026.0363.5—
Pre-trainedBackbone=CLIP-ViT-B/32, Bit-width=Full-precision2026.0563.2—
TIES-Merging w/ RTNBackbone=CLIP-ViT-B/32, Merging Strategy=TIES-Merging, Quantization Algorithm=RTN, Bit-width=4-bit, Lambda=0.32026.0563.1—
TIES-Merging w/ AWQBackbone=CLIP-ViT-B/32, Merging Strategy=TIES-Merging, Quantization Algorithm=AWQ, Bit-width=4-bit, Lambda=0.32026.0562.9—
WUDI-Merging w/ RTNBackbone=CLIP-ViT-B/32, Merging Strategy=WUDI-Merging, Quantization Algorithm=RTN, Bit-width=4-bit2026.0562.9—
Simple Averaging + GPTQMerging Strategy=Simple Averaging, Quantization Method=GPTQ, Bit-width=4-bit, Backbone=CLIP-ViT-B/322026.0562.79—
Task Arithmetic w/ AWQMerging Method=Task Arithmetic, Quantization Method=AWQ, Bit-width=4-bit, Model Backbone=CLIP-ViT-L/142026.0562.6—
WUDI-Merging w/ AWQBackbone=CLIP-ViT-B/32, Merging Strategy=WUDI-Merging, Quantization Algorithm=AWQ, Bit-width=4-bit2026.0562.5—
TIES-MergingMerging Strategy=TIES-Merging, Quantization Method=None, Model Backbone=CLIP-ViT-B/32, Bit-width=Full2026.0562.2—
Simple Averaging w/ RTNBackbone=CLIP-ViT-B/32, Merging Strategy=Simple Averaging, Quantization Algorithm=RTN, Bit-width=4-bit2026.0562—
Task Arithmetic + E-PMQMerging Strategy=Task Arithmetic, Quantization Method=E-PMQ, Bit-width=4-bit, Backbone=CLIP-ViT-B/322026.0561.83—
WUDI-Merging w/ AWQMerging Strategy=WUDI-Merging, Quantization Method=AWQ, Model Backbone=CLIP-ViT-B/32, Bit-width=4-bit2026.0561.75—