Update ASR_INTEGRATION_GUIDE.md
Browse files- ASR_INTEGRATION_GUIDE.md +3 -3
ASR_INTEGRATION_GUIDE.md
CHANGED
|
@@ -70,7 +70,7 @@ python examples/asr/asr_cache_aware_streaming/speech_to_text_multitalker_streami
|
|
| 70 |
binary_diar_preds=true \
|
| 71 |
att_context_size='[70,13]' \
|
| 72 |
fifo_len=264 \
|
| 73 |
-
spkcache_update_period=
|
| 74 |
generate_realtime_scripts=false \
|
| 75 |
output_path=./speaker_attributed_output.json
|
| 76 |
```
|
|
@@ -88,7 +88,7 @@ The important settings are:
|
|
| 88 |
| `masked_asr=false` | Uses speaker activity as conditioning for Multitalker Parakeet rather than simply muting the mixed waveform outside one speaker's regions. |
|
| 89 |
| `att_context_size='[70,13]'` | Uses 70 left-context frames and 13 right-context frames. With the model's current geometry, the streaming advance is about 1.12 seconds. |
|
| 90 |
| `cache_gating=true` | Stabilizes speaker-cache updates during streaming. |
|
| 91 |
-
| `spkcache_update_period=
|
| 92 |
|
| 93 |
`max_num_of_spks` is an upper bound, not a promise that every configured speaker
|
| 94 |
will appear. The diarization model discovers active speaker identities from the
|
|
@@ -117,7 +117,7 @@ config = MultitalkerTranscriptionConfig(
|
|
| 117 |
masked_asr=False,
|
| 118 |
cache_gating=True,
|
| 119 |
binary_diar_preds=True,
|
| 120 |
-
spkcache_update_period=
|
| 121 |
fifo_len=264,
|
| 122 |
)
|
| 123 |
|
|
|
|
| 70 |
binary_diar_preds=true \
|
| 71 |
att_context_size='[70,13]' \
|
| 72 |
fifo_len=264 \
|
| 73 |
+
spkcache_update_period=222 \
|
| 74 |
generate_realtime_scripts=false \
|
| 75 |
output_path=./speaker_attributed_output.json
|
| 76 |
```
|
|
|
|
| 88 |
| `masked_asr=false` | Uses speaker activity as conditioning for Multitalker Parakeet rather than simply muting the mixed waveform outside one speaker's regions. |
|
| 89 |
| `att_context_size='[70,13]'` | Uses 70 left-context frames and 13 right-context frames. With the model's current geometry, the streaming advance is about 1.12 seconds. |
|
| 90 |
| `cache_gating=true` | Stabilizes speaker-cache updates during streaming. |
|
| 91 |
+
| `spkcache_update_period=222` | Refreshes the speaker cache frequently enough for the long, eight-speaker sessions used by this demo. Tune this for other workloads. |
|
| 92 |
|
| 93 |
`max_num_of_spks` is an upper bound, not a promise that every configured speaker
|
| 94 |
will appear. The diarization model discovers active speaker identities from the
|
|
|
|
| 117 |
masked_asr=False,
|
| 118 |
cache_gating=True,
|
| 119 |
binary_diar_preds=True,
|
| 120 |
+
spkcache_update_period=222,
|
| 121 |
fifo_len=264,
|
| 122 |
)
|
| 123 |
|