QDrop
QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training QuantizationLLM quantization · first seen Mar 11, 2022
superseded — cited as a baseline and beaten by newer methods
0 papers critique it · 4 beat it on benchmarks
Beaten on benchmarks
Head-to-head results where a newer method reports beating QDrop. Values are copied from the source paper's tables — verify against the cited paper.
- FIMA-Q: Post-Training Quantization for Vision Transformers by Fisher Information Matrix Approximation
FIMA-Q (Ours) beats QDrop · top-1 accuracy [W/A 3/3]
64.09 vs 41.05
- FIMA-Q: Post-Training Quantization for Vision Transformers by Fisher Information Matrix Approximation
FIMA-Q (Ours) beats QDrop · top-1 accuracy [W/A 4/4]
76.68 vs 71.84
- FIMA-Q: Post-Training Quantization for Vision Transformers by Fisher Information Matrix Approximation
FIMA-Q (Ours) beats QDrop · AP^b (Mask R-CNN Swin-T) [W/A 4/4]
38.7 vs 36.2
- FIMA-Q: Post-Training Quantization for Vision Transformers by Fisher Information Matrix Approximation
FIMA-Q (Ours) beats QDrop · AP^m (Mask R-CNN Swin-T) [W/A 4/4]
37.8 vs 35.4
- MGRQ: Post-Training Quantization For Vision Transformer With Mixed Granularity Reconstruction
MGRQ beats QDrop · Top-1 accuracy [W4/A4]
70.02 vs 21.24
- MGRQ: Post-Training Quantization For Vision Transformer With Mixed Granularity Reconstruction
MGRQ beats QDrop · Top-1 accuracy [W6/A6]
80.39 vs 70.25
- ADFQ-ViT: Activation-Distribution-Friendly Post-Training Quantization for Vision Transformers
ADFQ-ViT (Ours) beats QDrop · Top-1 accuracy on ImageNet [4-bit (W-bit=4, A-bit=4)]
72.14 vs 21.24
Newer alternatives
Recent methods in the same sub-problem, not yet superseded in the knowledge base.
- COD-TDQWhen W4A4 Breaks Camouflaged Object Detection: Token-Group Dual-Constraint Activation QuantizationApr 18, 2026
- Joint Post-Training Quantization of Vision Transformers with Learned Prompt-Guided Data GenerationJoint Post-Training Quantization of Vision Transformers with Learned Prompt-Guided Data GenerationFeb 21, 2026