List your company

Image Classification on ImageNet Source to 10 Fine-Grained Targets (test)

96.96Caltech101 AccuracyFGA
92.311293.518194.72595.9319Oct 14, 2024Jan 8, 2025Apr 5, 2025Jun 30, 2025Sep 25, 2025Dec 20, 2025Mar 17, 2026

Evaluation results

MethodLinks
FGABackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0196.9691.2868.9372.1126.9769.2949.7647.5884.9568.1767.6—
LPT2024.109590.8766.573.425.676843.9554.1386.5369.2767.3371.77
EvoPromptTraining Source=ImageNet, Evaluation Mode=Cross-Dataset Transfer2026.0394.990.465.872.524.267.446.9249.487.469.366.8271.63
HeBAYear=Ours, Training Shots=16 shots, Source Training Dataset=ImageNet2026.0394.8192.265.4173.0427.0968.2250.7158.9986.1370.4568.7171.5
LwEIBYear=2025, Training Shots=16 shots, Source Training Dataset=ImageNet2026.0394.5192.566.5873.0327.769.3350.6355.3786.3770.0368.6171.31
ZERO+MaPLeBackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0194.4890.668.5871.6226.2568.245.8642.1786.7769.8766.44—
CoCoOpBackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0194.4390.1465.3271.8822.9467.3645.7339.2383.9768.4464.94—
CoCoOpTraining Source=ImageNet, Evaluation Mode=Cross-Dataset Transfer2026.0394.4390.1465.3271.8822.9467.3645.7345.3786.0668.2165.7471.02
CoCoOpYear=2022, Training Shots=16 shots, Source Training Dataset=ImageNet2026.0394.4390.1465.3271.8822.9467.3645.7345.3786.0668.2165.7471.02
CoCoOp2024.1094.4390.1465.3271.8822.9467.3645.7345.3786.0668.2165.7471.02
TDABackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0194.2488.6367.2871.4223.9167.5447.45886.1470.6667.53—
HPTYear=2024, Training Shots=16 shots, Source Training Dataset=ImageNet2026.0394.292.6366.3374.8425.6868.7550.8747.3686.2170.567.7471.72
TPTBackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0194.1687.7966.8768.9824.7865.547.7542.4484.6768.0465.1—
PromptAlignBackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0194.0190.7668.572.3924.867.5447.2447.8686.6569.4766.92—
TCPTraining Source=ImageNet, Evaluation Mode=Cross-Dataset Transfer2026.0393.9791.2564.6971.2123.4567.1544.3551.4586.6968.7366.2971.4
TCP2024.1093.9791.2564.6971.2123.4567.1544.3551.4586.6968.7366.2971.4
ZERO+CoOpBackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0193.8588.3664.967.2319.1464.7343.6233.5382.6766.6162.46—
MMATraining Source=ImageNet, Evaluation Mode=Cross-Dataset Transfer2026.0393.890.366.1372.0725.3368.1746.5749.2486.1268.3266.6171
MMAYear=2024, Training Shots=16 shots, Source Training Dataset=ImageNet2026.0393.890.366.1372.0725.3368.1746.5749.2486.1268.3266.6171
MMA2024.1093.890.366.1372.0725.3368.1746.5749.2486.1268.3266.6171
TPT+CoOpBackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0193.7588.9367.0668.2525.8966.447.1548.7883.8266.5365.66—
CoOpBackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0193.789.1464.5168.7118.4764.1541.9246.3985.366.5563.88—
CoOpTraining Source=ImageNet, Evaluation Mode=Cross-Dataset Transfer2026.0393.789.1464.5168.7118.4764.1541.9246.3985.366.5563.8871.51
CoOpYear=2022, Training Shots=16 shots, Source Training Dataset=ImageNet2026.0393.789.1464.5168.7118.4764.1541.9246.3985.366.5563.8871.51
CLIP2024.1093.789.165.770.724.962.644.348.385.967.665.2466.7
CoOp2024.1093.789.1464.5168.7118.4764.1541.9246.3985.366.5563.8871.51
ZEROBackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0193.6687.7568.0467.6825.2165.0346.1234.3386.5367.7764.21—
C-TPTBackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0193.688.265.869.82464.84643.283.765.764.48—
PromptSRCTraining Source=ImageNet, Evaluation Mode=Cross-Dataset Transfer2026.0393.690.2565.770.2523.967.146.8745.586.1568.7565.8171.27
P-SRCYear=2023, Training Shots=16 shots, Source Training Dataset=ImageNet2026.0393.690.2565.770.2523.967.146.8745.586.1568.7565.8171.27
PromptSRC2024.1093.690.2565.770.2523.967.146.8745.586.1568.7565.8171.27
TPT+MaPLeBackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0193.5990.7266.572.3724.767.5445.8747.886.6469.1966.5—
MaPLeTraining Source=ImageNet, Evaluation Mode=Cross-Dataset Transfer2026.0393.5390.4965.5772.2324.7467.0146.4948.0686.268.6966.370.72
MaPLeYear=2023, Training Shots=16 shots, Source Training Dataset=ImageNet2026.0393.5390.4965.5772.2324.7467.0146.4948.0686.268.6966.370.72
CLIP-ViT-B/16Backbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0193.3588.2565.4867.4423.6762.5944.2742.0183.6565.1363.58—
CLIPYear=2021, Training Shots=16 shots, Source Training Dataset=ImageNet2026.0392.9889.1365.2971.324.962.5944.5647.8486.1166.8365.1566.72
DiffTPTBackbone=ViT-B/16, Source Dataset=ImageNet, Prompt Tuning Shots=16-shot2025.0192.4988.2267.0170.125.665.744743.1387.2362.6764.92—