Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Inference Throughput on GPT-J 6B

1,484Inference Throughput (tokens/s)

GPT-J-6B (Proposed Method)

1,212.561,283.031,353.51,423.97May 14, 2026
Updated 19d ago

Evaluation Results

MethodLinks
2026.05
1,484113.8
2026.05
1,430.3108
2026.05
1,426.8106.3
2026.05
1,396.492.7
2026.05
1,38268
2026.05
1,37467.9
2026.05
1,370.2-
2026.05
1,325.764.8
2026.05
1,322.3-
2026.05
1,320.5-
2026.05
1,314.1-
2026.05
1,306.1-
2026.05
1,303.8-
2026.05
1,276.853.8
2026.05
1,260.9-
2026.05
1,223-