Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

VLM prefill

Benchmarks

Task NameDataset NameSOTA ResultTrend
Inference LatencyVLM prefill 1024 tokens
Latency (ms)92.7
2
Inference LatencyVLM prefill 512 tokens
Prefill Latency (ms)59.4
2
Showing 2 of 2 rows