Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
INT2 Quantization on LLaMA-30B
Loading...
18.59
Memory Footprint (GB)
OPTQ
17.6136
24.2043
30.795
37.3857
Feb 14, 2024
Memory Footprint (GB)
Updated 4d ago
Evaluation Results
Method
Method
Links
Memory Footprint (GB)
OPTQ
Target=layer-wise reco...
2024.02
18.59
Z-FOLD
Target=layer-wise reco...
2024.02
18.59
OPTQ
Target=layer-wise reco...
2024.02
18.59
Z-FOLD
Target=layer-wise reco...
2024.02
18.59
OmniQuant
Target=attention-wise...
2024.02
24.53
OmniQuant
Target=attention-wise...
2024.02
24.53
AffineQuant
Target=attention-wise...
2024.02
38.59
AffineQuant
Target=attention-wise...
2024.02
38.59
aespa
Target=attention-wise...
2024.02
43
aespa
Target=attention-wise...
2024.02
43
Feedback
Search any
task
Search any
task