vh
a6c1d3c454
A/B of the live STT seat (fv-ml1 GPU 0, sherpa-onnx int8 v3) against nvidia/parakeet-unified-en-0.6b, measured on GPU 3 with the seat's own image, k2-fsa's published unified int8 export, fp32/fp16 exports made with k2-fsa's recipe, v2 int8, and NeMo 3.0.0 (fp32, bf16 autocast, bf16 weights). - Seat int8 graph runs on one CPU thread (cpu/wall 1.00, GPU 2-9%). - unified-en under NeMo: -121/-234/-530 ms vs the seat at 1-3/3-8/8-20 s (paired, n=120/bin; floor <=6 ms; +50 ms positive control reads +52-54). - unified-en WER lower in every runtime: -0.7 pp clean, -1.5 pp other, -3.2 to -4.4 pp AMI (paired CIs exclude 0). - Seat defects found: hard 400 s input ceiling (HTTP 500), truncation after a quiet 1.5 s pause, and severe long-window dropouts (int8 v3 only). - B-bf16w needs +0.8 to +1.5 GB over the seat's 1,690 MiB on GPU 0. Raw requests, hypotheses, manifests and the full harness under services/parakeet-ab-2026-09-30/. No deploy; live seat untouched apart from 240 light test requests.
15 lines
593 B
Bash
Executable File
15 lines
593 B
Bash
Executable File
#!/usr/bin/env bash
|
|
# Long-window run with a GPU 3 headroom guard (Scriberr is an on-demand tenant of GPU 3).
|
|
cd /tank/spikes/parakeet-ab
|
|
AB_WINS=${AB_WINS:-long} python3 code/longwin.py run "$@" > out/longwin-run.log 2>&1 &
|
|
run=$!
|
|
while kill -0 $run 2>/dev/null; do
|
|
f=$(nvidia-smi -i 3 --query-gpu=memory.free --format=csv,noheader,nounits)
|
|
if [ "$f" -lt 8192 ]; then
|
|
echo "GUARD fired: GPU3 free $f MiB"; for c in $(docker ps --filter label=ab=parakeet-2026-09-30 --format "{{.Names}}"); do docker rm -f $c >/dev/null; done; break
|
|
fi
|
|
sleep 0.5
|
|
done
|
|
wait $run
|
|
echo "LONGWIN DONE"
|