List your company

Throughput Measurement on LLaMA-2 13B

19.4Throughput (tokens/s)Token Filtering
2.0326.54111.0515.559Dec 8, 2025

Evaluation results

MethodLinks
Token FilteringPruning Ratio=50%, Batch size=82025.1219.4
FLAPPruning Ratio=50%, Batch size=82025.1219.4
ProbePruning Ratio=50%, Batch size=82025.1219.2
FLAPPruning Ratio=50%, Batch size=162025.1219.1
Token FilteringPruning Ratio=50%, Batch size=162025.1217.8
ProbePruning Ratio=50%, Batch size=162025.1217.2
DensePruning Ratio=0%, Batch size=82025.1216.8
DensePruning Ratio=0%, Batch size=162025.1216.1
FLAPPruning Ratio=50%, Batch size=322025.1215.4
Token FilteringPruning Ratio=50%, Batch size=322025.1214.9
ProbePruning Ratio=50%, Batch size=322025.1214.6
DensePruning Ratio=0%, Batch size=322025.1211.8
FLAPPruning Ratio=50%, Batch size=642025.129.6
Token FilteringPruning Ratio=50%, Batch size=642025.128.8
ProbePruning Ratio=50%, Batch size=642025.127.5
FLAPPruning Ratio=50%, Batch size=1282025.125.8
Token FilteringPruning Ratio=50%, Batch size=1282025.125.2
DensePruning Ratio=0%, Batch size=642025.124.7
ProbePruning Ratio=50%, Batch size=1282025.124.4
DensePruning Ratio=0%, Batch size=1282025.122.7