Audio encoder conv positional layer stack
Complete fused audio preprocessing pipeline that performs two strided 1D convolutions with GELU activations on mel spectrogram input, adds learned positional embeddings, and processes through the first encoder layer's self-attention and FFN blocks. This represents the critical path from raw audio features to contextualized audio embeddings.
Reported evidence · last observed 2026-08-05. Reported by source; not independently reproduced. No public source. License unknown.
Current records
Workloadsuite of 15 cases · mean latencyd_model = 5120 · head_dim = 256 · num_heads = 20 · num_mel_bins = 80 · input_seq_len = 3000 · output_seq_len = 1500 · encoder_ffn_dim = 2048015 cases
Not measured on H100 for this workload. Challenges →
112.3ms87% SOL1.00×Reported · license unknown · no source2026-08-05
Measured exactly what you asked. Reported by source; not independently reproduced.
212.6ms85% SOL1.02×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
313.2ms79% SOL1.08×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
413.4ms78% SOL1.09×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
514.9ms66% SOL1.21×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
615.0ms66% SOL1.22×Reported · license unknown · no source2026-08-03
Measured exactly what you asked. Reported by source; not independently reproduced.
715.0ms66% SOL1.22×Reported · license unknown · no source2026-08-15
Measured exactly what you asked. Reported by source; not independently reproduced.
815.0ms65% SOL1.22×Reported · license unknown · no source2026-08-15
Measured exactly what you asked. Reported by source; not independently reproduced.
915.5ms63% SOL1.26×Reported · license unknown · no source2026-07-18
Measured exactly what you asked. Reported by source; not independently reproduced.
1015.6ms62% SOL1.27×Reported · license unknown · no source2026-08-15
Measured exactly what you asked. Reported by source; not independently reproduced.
1115.8ms61% SOL1.29×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
1215.9ms61% SOL1.29×Reported · license unknown · no source2026-08-23
Measured exactly what you asked. Reported by source; not independently reproduced.
1316.0ms60% SOL1.30×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
1416.1ms60% SOL1.31×Reported · license unknown · no source2026-08-17
Measured exactly what you asked. Reported by source; not independently reproduced.
1516.1ms60% SOL1.31×Reported · license unknown · no source2026-07-17
Measured exactly what you asked. Reported by source; not independently reproduced.
1616.1ms59% SOL1.31×Reported · license unknown · no source2026-08-16
Measured exactly what you asked. Reported by source; not independently reproduced.
1716.2ms59% SOL1.32×Reported · license unknown · no source2026-08-22
Measured exactly what you asked. Reported by source; not independently reproduced.
1816.2ms59% SOL1.32×Reported · license unknown · no source2026-08-12
Measured exactly what you asked. Reported by source; not independently reproduced.
1916.2ms59% SOL1.32×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
2016.2ms59% SOL1.32×Reported · license unknown · no source2026-08-19
Measured exactly what you asked. Reported by source; not independently reproduced.
2116.4ms58% SOL1.33×Reported · license unknown · no source2026-08-20
Measured exactly what you asked. Reported by source; not independently reproduced.
2216.5ms58% SOL1.34×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
2316.5ms58% SOL1.34×Reported · license unknown · no source2026-08-23
Measured exactly what you asked. Reported by source; not independently reproduced.
2416.5ms58% SOL1.34×Reported · license unknown · no source2026-08-19
Measured exactly what you asked. Reported by source; not independently reproduced.
2517.4ms54% SOL1.41×Reported · license unknown · no source2026-08-06
Measured exactly what you asked. Reported by source; not independently reproduced.
2617.4ms54% SOL1.42×Reported · license unknown · no source2026-07-12
Measured exactly what you asked. Reported by source; not independently reproduced.
2717.9ms52% SOL1.45×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
2819.2ms48% SOL1.56×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
2925.0ms34% SOL2.03×Reported · license unknown · no source2026-07-25
Measured exactly what you asked. Reported by source; not independently reproduced.