Open fotstrt opened 1 year ago
When running the efficientnet_v2_m model, Orion adds significant overhead (Inference, BS 4: ~21 ms running alone, ~26 ms with Orion). Investigate
When running the efficientnet_v2_m model, Orion adds significant overhead (Inference, BS 4: ~21 ms running alone, ~26 ms with Orion). Investigate