File size: 36,210 Bytes
7a12a78
 
 
 
 
 
 
 
 
 
 
 
 
 
 
87842ef
 
7a12a78
 
 
 
 
 
 
 
 
 
 
 
7965430
af9c3d5
 
7a12a78
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
"""WebSocket ASR server for Nemotron-Speech with true incremental streaming and timestamps."""

import asyncio
import argparse
import hashlib
import json
import os
from dataclasses import dataclass, field
from typing import Any, Optional, Tuple

import numpy as np
import torch
from aiohttp import web, WSMsgType
from loguru import logger

from nemo.collections.asr.parts.utils.transcribe_utils import normalize_timestamp_output
from nemo.collections.asr.parts.utils.timestamp_utils import process_timestamp_outputs

# Enable debug logging with DEBUG_ASR=1
DEBUG_ASR = os.environ.get("DEBUG_ASR", "0") == "1"


def _hash_audio(audio: np.ndarray) -> str:
    """Get short hash of audio array for debugging."""
    if audio is None or len(audio) == 0:
        return "empty"
    return hashlib.md5(audio.tobytes()).hexdigest()[:8]


# DEFAULT_MODEL = "./nemotron-speech-streaming-en-0.6b/nemotron-speech-streaming-en-0.6b.nemo"
# DEFAULT_MODEL = "results/NeMo_Ja_FastConformer_Streaming/checkpoints/NeMo_Ja_FastConformer_Streaming.nemo"
DEFAULT_MODEL = "results/NeMo_Ja_FastConformer_Transducer_RNNT_EOU/checkpoints/NeMo_Ja_FastConformer_Transducer_RNNT_EOU.nemo"

# Right context options for att_context_size=[70, X]
RIGHT_CONTEXT_OPTIONS = {
    0: "~80ms ultra-low latency",
    1: "~160ms low latency (recommended)",
    6: "~560ms balanced",
    13: "~1.12s highest accuracy",
}


@dataclass
class ASRSession:
    """Per-connection session state with caches for true incremental streaming."""

    id: str
    websocket: Any

    # Accumulated audio buffer (all audio received so far)
    accumulated_audio: Optional[np.ndarray] = None

    # Number of mel frames already emitted to encoder
    emitted_frames: int = 0

    # Encoder cache state
    cache_last_channel: Optional[torch.Tensor] = None
    cache_last_time: Optional[torch.Tensor] = None
    cache_last_channel_len: Optional[torch.Tensor] = None

    # Decoder state
    previous_hypotheses: Any = None
    pred_out_stream: Any = None

    # Current transcription (model's cumulative output)
    current_text: str = ""
    
    # Current timestamps
    current_timestamps: Optional[dict] = None

    # Last text emitted to client on hard reset (for server-side deduplication)
    # We only send the delta (new portion) to avoid downstream duplication
    last_emitted_text: str = ""

    # Audio overlap buffer for mid-utterance reset continuity
    # This preserves the last N ms of audio to provide encoder left-context
    # when a new segment starts after a reset
    overlap_buffer: Optional[np.ndarray] = None


class ASRServer:
    """WebSocket server for streaming ASR with true incremental processing."""

    def __init__(
        self,
        model: str,
        host: str = "0.0.0.0",
        port: int = 8080,
        right_context: int = 1,
    ):
        self.model_name_or_path = model
        self.host = host
        self.port = port
        self.right_context = right_context
        self.model = None
        self.sample_rate = 16000

        # Inference lock
        self.inference_lock = asyncio.Lock()

        # Active sessions
        self.sessions: dict[str, ASRSession] = {}

        # Model loaded flag for health check
        self.model_loaded = False

        # Streaming parameters (calculated from model config)
        self.shift_frames = None
        self.pre_encode_cache_size = None
        self.hop_samples = None

        # Audio overlap for mid-utterance reset continuity (calculated in load_model)
        self.overlap_samples = None

    def load_model(self):
        """Load the NeMo ASR model with streaming configuration."""
        import nemo.collections.asr as nemo_asr
        from omegaconf import OmegaConf

        # Detect if model is a local .nemo file or HuggingFace model name
        is_local_file = (
            self.model_name_or_path.endswith('.nemo') or
            os.path.exists(self.model_name_or_path)
        )

        if is_local_file:
            logger.info(f"Loading model from local file: {self.model_name_or_path}")
            self.model = nemo_asr.models.ASRModel.restore_from(
                self.model_name_or_path, map_location='cpu'
            )
        else:
            logger.info(f"Loading model from HuggingFace: {self.model_name_or_path}")
            self.model = nemo_asr.models.ASRModel.from_pretrained(
                self.model_name_or_path, map_location='cpu'
            )
        self.model = self.model.cuda()

        # Configure attention context for streaming
        logger.info(f"Setting att_context_size=[70, {self.right_context}] ({RIGHT_CONTEXT_OPTIONS.get(self.right_context, 'custom')})")
        if hasattr(self.model.encoder, "set_default_att_context_size"):
            self.model.encoder.set_default_att_context_size([70, self.right_context])

        # Configure greedy decoding (required for Blackwell GPU)
        logger.info("Configuring greedy decoding for Blackwell compatibility and enabling timestamps...")
        
        # Check model type to set preserve_alignments
        preserve_alignments = False
        if hasattr(self.model, 'joint'):  # RNNT model
            preserve_alignments = True
            
        decoding_cfg_dict = {
            'strategy': 'greedy',
            'greedy': {
                'max_symbols': 10,
                'loop_labels': False,
                'use_cuda_graph_decoder': False,
            },
            'compute_timestamps': True
        }
        
        if preserve_alignments:
            decoding_cfg_dict['preserve_alignments'] = True
            
        self.model.change_decoding_strategy(
            decoding_cfg=OmegaConf.create(decoding_cfg_dict)
        )
        
        # Force enable timestamps
        if hasattr(self.model, 'decoding'):
            if hasattr(self.model.decoding, 'compute_timestamps'):
                self.model.decoding.compute_timestamps = True
            if hasattr(self.model.decoding, 'preserve_alignments'):
                self.model.decoding.preserve_alignments = preserve_alignments
            if hasattr(self.model.decoding, 'ctc_decoder') and hasattr(self.model.decoding.ctc_decoder, 'compute_timestamps'):
                self.model.decoding.ctc_decoder.compute_timestamps = True
                self.model.decoding.ctc_decoder.return_hypotheses = True
                
        # Force RNNT decoder settings if present
        if hasattr(self.model, 'joint'):
            if hasattr(self.model.decoding, 'rnnt_decoder_predictions_tensor'):
                if hasattr(self.model.decoding, 'compute_timestamps'):
                    # We MUST set this to False during the stream step, 
                    # otherwise the internal `rnnt_decoder_predictions_tensor` will try to compute 
                    # timestamps on partial chunks and fail with length mismatch ValueError.
                    # We will manually compute the timestamps later in `_process_chunk` / `_process_final_chunk`
                    self.model.decoding.compute_timestamps = False
                if hasattr(self.model.decoding, 'preserve_alignments'):
                    self.model.decoding.preserve_alignments = True
                if hasattr(self.model.decoding, 'return_hypotheses'):
                    self.model.decoding.return_hypotheses = True
                
        self.model.eval()

        # Disable dither for deterministic preprocessing
        self.model.preprocessor.featurizer.dither = 0.0

        # Get streaming config
        scfg = self.model.encoder.streaming_cfg
        logger.info(f"Streaming config: chunk_size={scfg.chunk_size}, shift_size={scfg.shift_size}")

        # Calculate parameters
        preprocessor_cfg = self.model.cfg.preprocessor
        hop_length_sec = preprocessor_cfg.get('window_stride', 0.01)
        self.hop_samples = int(hop_length_sec * self.sample_rate)

        # shift_size[1] = 16 frames for 160ms chunks
        self.shift_frames = scfg.shift_size[1] if isinstance(scfg.shift_size, list) else scfg.shift_size

        # pre_encode_cache_size[1] = 9 frames
        pre_cache = scfg.pre_encode_cache_size
        self.pre_encode_cache_size = pre_cache[1] if isinstance(pre_cache, list) else pre_cache

        # drop_extra_pre_encoded for non-first chunks
        self.drop_extra = scfg.drop_extra_pre_encoded

        # Calculate silence padding for final chunk:
        # - right_context chunks for encoder lookahead
        # - 1 additional chunk for decoder finalization
        # With right_context=1, this is (1+1)*160ms = 320ms
        self.final_padding_frames = (self.right_context + 1) * self.shift_frames
        padding_ms = self.final_padding_frames * hop_length_sec * 1000

        # Calculate audio overlap for mid-utterance reset continuity
        # Use pre_encode_cache_size frames = 90ms of left-context
        # This allows the encoder to have proper context when starting a new segment
        self.overlap_samples = self.pre_encode_cache_size * self.hop_samples
        overlap_ms = self.overlap_samples * 1000 / self.sample_rate

        shift_ms = self.shift_frames * hop_length_sec * 1000
        logger.info(f"Model loaded: {type(self.model).__name__}")
        logger.info(f"Shift size: {shift_ms:.0f}ms ({self.shift_frames} frames)")
        logger.info(f"Pre-encode cache: {self.pre_encode_cache_size} frames")
        logger.info(f"Final chunk padding: {padding_ms:.0f}ms ({self.final_padding_frames} frames)")
        logger.info(f"Audio overlap for resets: {overlap_ms:.0f}ms ({self.overlap_samples} samples)")

        # Warmup inference to ensure model is fully loaded on GPU
        # This prevents GPU memory issues when LLM starts later
        self._warmup()

    def _warmup(self):
        """Run warmup inference using streaming API to claim GPU memory."""
        import time

        logger.info("Running warmup inference (streaming API) to claim GPU memory...")
        start = time.perf_counter()

        # Generate 1 second of silence plus padding for warmup
        warmup_samples = self.sample_rate + (self.final_padding_frames * self.hop_samples)
        warmup_audio = np.zeros(warmup_samples, dtype=np.float32)

        # Run streaming inference to force all CUDA kernels to compile
        with torch.inference_mode():
            audio_tensor = torch.from_numpy(warmup_audio).unsqueeze(0).cuda()
            audio_len = torch.tensor([len(warmup_audio)], device='cuda')

            # Preprocess
            mel, mel_len = self.model.preprocessor(input_signal=audio_tensor, length=audio_len)

            # Get initial cache
            cache = self.model.encoder.get_initial_cache_state(batch_size=1)

            # Run streaming step (processes entire mel as one chunk)
            _ = self.model.conformer_stream_step(
                processed_signal=mel,
                processed_signal_length=mel_len,
                cache_last_channel=cache[0],
                cache_last_time=cache[1],
                cache_last_channel_len=cache[2],
                keep_all_outputs=True,
                previous_hypotheses=None,
                previous_pred_out=None,
                drop_extra_pre_encoded=0,
                return_transcription=True,
            )

        elapsed = (time.perf_counter() - start) * 1000
        logger.info(f"Warmup complete in {elapsed:.0f}ms - GPU memory claimed")

    def _init_session(self, session: ASRSession):
        """Initialize a fresh session."""
        # Initialize encoder cache
        cache = self.model.encoder.get_initial_cache_state(batch_size=1)
        session.cache_last_channel = cache[0]
        session.cache_last_time = cache[1]
        session.cache_last_channel_len = cache[2]

        # Reset audio buffer and frame counter
        if session.overlap_buffer is not None and len(session.overlap_buffer) > 0:
            session.accumulated_audio = session.overlap_buffer.copy()
            overlap_ms = len(session.overlap_buffer) * 1000 / self.sample_rate
            logger.debug(
                f"Session {session.id}: prepending {len(session.overlap_buffer)} samples "
                f"({overlap_ms:.0f}ms) of overlap audio"
            )
            session.overlap_buffer = None  # Clear after use
        else:
            session.accumulated_audio = np.array([], dtype=np.float32)

        session.emitted_frames = 0

        # Reset decoder state
        session.previous_hypotheses = None
        session.pred_out_stream = None
        session.current_text = ""
        session.current_timestamps = None

    async def websocket_handler(self, request: web.Request) -> web.WebSocketResponse:
        """Handle a WebSocket client connection."""
        import uuid

        ws = web.WebSocketResponse(max_msg_size=10 * 1024 * 1024)
        await ws.prepare(request)

        session_id = str(uuid.uuid4())[:8]
        session = ASRSession(id=session_id, websocket=ws)
        self.sessions[session_id] = session

        logger.info(f"Client {session_id} connected")

        try:
            async with self.inference_lock:
                await asyncio.get_event_loop().run_in_executor(
                    None, self._init_session, session
                )

            await ws.send_str(json.dumps({"type": "ready"}))
            logger.debug(f"Client {session_id}: sent ready")

            async for msg in ws:
                if msg.type == WSMsgType.BINARY:
                    await self._handle_audio(session, msg.data)
                elif msg.type == WSMsgType.TEXT:
                    try:
                        data = json.loads(msg.data)
                        msg_type = data.get("type")

                        if msg_type == "reset" or msg_type == "end":
                            finalize = data.get("finalize", True)
                            await self._reset_session(session, finalize=finalize)
                        else:
                            logger.warning(f"Client {session_id}: unknown message type: {msg_type}")

                    except json.JSONDecodeError:
                        logger.warning(f"Client {session_id}: invalid JSON")
                elif msg.type == WSMsgType.ERROR:
                    logger.error(f"Client {session_id} WebSocket error: {ws.exception()}")
                    break

            logger.info(f"Client {session_id} disconnected")

        except Exception as e:
            logger.error(f"Client {session_id} error: {e}")
            import traceback
            logger.error(traceback.format_exc())
            try:
                await ws.send_str(json.dumps({
                    "type": "error",
                    "message": str(e)
                }))
            except:
                pass
        finally:
            if session_id in self.sessions:
                del self.sessions[session_id]

        return ws

    async def _handle_audio(self, session: ASRSession, audio_bytes: bytes):
        """Accumulate audio and process when enough frames available."""
        audio_np = np.frombuffer(audio_bytes, dtype=np.int16).astype(np.float32) / 32768.0

        if DEBUG_ASR:
            chunk_hash = hashlib.md5(audio_bytes).hexdigest()[:8]
            logger.debug(f"Session {session.id}: recv chunk {len(audio_bytes)}B hash={chunk_hash}")

        session.accumulated_audio = np.concatenate([session.accumulated_audio, audio_np])

        # Process if we have enough audio for new frames
        min_audio_for_chunk = (session.emitted_frames + self.shift_frames + 1) * self.hop_samples

        while len(session.accumulated_audio) >= min_audio_for_chunk:
            async with self.inference_lock:
                result = await asyncio.get_event_loop().run_in_executor(
                    None, self._process_chunk, session
                )

            if result is not None:
                text, timestamps = result
                if text is not None and text != session.current_text:
                    session.current_text = text
                    session.current_timestamps = timestamps
                    logger.debug(f"Session {session.id} interim: {text[-50:] if len(text) > 50 else text}")
                    
                    formatted_timestamps = []
                    if timestamps:
                        if isinstance(timestamps, dict):
                            for key, val in timestamps.items():
                                if key != 'timestep':
                                    formatted_timestamps.append({key: normalize_timestamp_output(val)})
                        elif isinstance(timestamps, list):
                            # It might be a list of dictionaries if returned directly from Hypothesis
                            formatted_timestamps = timestamps
                                
                    await session.websocket.send_str(json.dumps({
                        "type": "transcript",
                        "text": text,
                        "timestamps": formatted_timestamps if formatted_timestamps else None,
                        "is_final": False
                    }))

            # Update minimum for next iteration
            min_audio_for_chunk = (session.emitted_frames + self.shift_frames + 1) * self.hop_samples

    def _decode_stream_output(self, session, pred_out_stream):
        """Manually decode model outputs to retrieve timestamps."""
        # For RNNT models
        if hasattr(self.model, 'joint'):
            decoding = self.model.decoding
            transcribed_texts = []
            for preds_idx, preds_concat in enumerate(pred_out_stream):
                # We need to reshape for RNNT decoder which expects [B, D] or [B, T, D]
                # preds_concat is usually [T, D] from streaming step
                if preds_concat.dim() == 2:
                    preds_tensor = preds_concat.unsqueeze(0)  # [1, T, D]
                else:
                    preds_tensor = preds_concat
                    
                encoded_len = torch.tensor([preds_tensor.size(1)], device=preds_tensor.device)
                
                # We must use decoding() directly instead of rnnt_decoder_predictions_tensor
                # so that hypothesis is correctly initialized with alignments before calling compute_rnnt_timestamps
                hypotheses_list = decoding(
                    encoder_output=preds_tensor, 
                    encoded_lengths=encoded_len, 
                    partial_hypotheses=session.previous_hypotheses
                )
                
                # decoding() returns a tuple where [0] is a list of hypotheses
                hypotheses_list = hypotheses_list[0]
                
                if isinstance(hypotheses_list[0], list):
                    transcribed_texts.append(hypotheses_list[0][0])
                else:
                    transcribed_texts.append(hypotheses_list[0])
        # For CTC models
        else:
            if hasattr(self.model, 'ctc_decoder'):
                decoding = self.model.ctc_decoding
            else:
                decoding = self.model.decoding
            
            transcribed_texts = []
            for preds_idx, preds_concat in enumerate(pred_out_stream):
                encoded_len = torch.tensor([len(preds_concat)], device=preds_concat.device)
                decoded_out = decoding.ctc_decoder_predictions_tensor(
                    decoder_outputs=preds_concat.unsqueeze(0),
                    decoder_lengths=encoded_len,
                    return_hypotheses=True,
                )
                if isinstance(decoded_out[0], list):
                    transcribed_texts.append(decoded_out[0][0])
                else:
                    transcribed_texts.append(decoded_out[0])
                
        # process timestamps
        if hasattr(self.model.cfg, 'preprocessor'):
            window_stride = self.model.cfg.preprocessor.get('window_stride', 0.01)
        else:
            window_stride = 0.01
            
        subsampling_factor = 1
        if hasattr(self.model, 'encoder') and hasattr(self.model.encoder, 'subsampling_factor'):
            subsampling_factor = self.model.encoder.subsampling_factor
        elif hasattr(self.model, 'encoder') and hasattr(self.model.encoder, 'conv_subsampling_factor'):
            subsampling_factor = self.model.encoder.conv_subsampling_factor
            
        # RNNT model returns a tuple of lists containing hypotheses when using decoding() directly
        # We need to process the timestamps if they haven't been computed inside decoding()
        if hasattr(self.model, 'joint'):
            import copy
            timestamp_type = 'all'
            if hasattr(decoding, 'cfg'):
                timestamp_type = decoding.cfg.get('rnnt_timestamp_type', 'all')
                
            for i in range(len(transcribed_texts)):
                if hasattr(transcribed_texts[i], 'timestamp') and not transcribed_texts[i].timestamp:
                    # Before computing timestamps, ensure the Hypothesis text contains the temporary storage 
                    # format required by `compute_rnnt_timestamps` 
                    if hasattr(transcribed_texts[i], 'y_sequence'):
                        prediction = transcribed_texts[i].y_sequence
                        if type(prediction) != list:
                            prediction = prediction.tolist()
                            
                        # Remove any blank and possibly big blank tokens from prediction
                        if decoding.big_blank_durations is not None and decoding.big_blank_durations != []:  # multi-blank RNNT
                            num_extra_outputs = len(decoding.big_blank_durations)
                            prediction = [p for p in prediction if p < decoding.blank_id - num_extra_outputs]
                        elif hasattr(decoding, '_is_tdt') and decoding._is_tdt:  # TDT model.
                            prediction = [p for p in prediction if p < decoding.blank_id]
                        else:  # standard RNN-T
                            prediction = [p for p in prediction if p != decoding.blank_id]
                            
                        alignments = copy.deepcopy(transcribed_texts[i].alignments)
                        token_repetitions = [1] * len(alignments)
                        
                        # Update hypothesis text to hold the tuple (prediction, alignments, token_repetitions)
                        transcribed_texts[i].text = (prediction, alignments, token_repetitions)
                        
                        # Now compute the timestamps 
                        transcribed_texts[i] = decoding.compute_rnnt_timestamps(transcribed_texts[i], timestamp_type)
        
        process_timestamp_outputs(transcribed_texts, subsampling_factor=subsampling_factor, window_stride=window_stride)
        return transcribed_texts

    def _process_chunk(self, session: ASRSession) -> Optional[Tuple[str, Optional[dict]]]:
        """Process accumulated audio, extract new mel frames, run streaming inference."""
        try:
            # Preprocess ALL accumulated audio
            audio_tensor = torch.from_numpy(session.accumulated_audio).unsqueeze(0).cuda()
            audio_len = torch.tensor([len(session.accumulated_audio)], device='cuda')

            with torch.inference_mode():
                mel, mel_len = self.model.preprocessor(
                    input_signal=audio_tensor,
                    length=audio_len
                )

                # Available frames (excluding last edge frame)
                available_frames = mel.shape[-1] - 1
                new_frame_count = available_frames - session.emitted_frames

                if new_frame_count < self.shift_frames:
                    return session.current_text, session.current_timestamps  # Not enough new frames

                # Extract chunk with pre-encode cache
                if session.emitted_frames == 0:
                    chunk_start = 0
                    chunk_end = self.shift_frames
                    drop_extra = 0
                else:
                    chunk_start = session.emitted_frames - self.pre_encode_cache_size
                    chunk_end = session.emitted_frames + self.shift_frames
                    drop_extra = self.drop_extra

                chunk_mel = mel[:, :, chunk_start:chunk_end]
                chunk_len = torch.tensor([chunk_mel.shape[-1]], device='cuda')

                # Run streaming inference
                (
                    session.pred_out_stream,
                    transcribed_texts,
                    session.cache_last_channel,
                    session.cache_last_time,
                    session.cache_last_channel_len,
                    session.previous_hypotheses,
                ) = self.model.conformer_stream_step(
                    processed_signal=chunk_mel,
                    processed_signal_length=chunk_len,
                    cache_last_channel=session.cache_last_channel,
                    cache_last_time=session.cache_last_time,
                    cache_last_channel_len=session.cache_last_channel_len,
                    keep_all_outputs=False,
                    previous_hypotheses=session.previous_hypotheses,
                    previous_pred_out=session.pred_out_stream,
                    drop_extra_pre_encoded=drop_extra,
                    return_transcription=True, 
                )

                # Update emitted frame count
                session.emitted_frames += self.shift_frames

                # For RNNT models, conformer_stream_step already returns the decoded hypotheses in `transcribed_texts`
                # (which is assigned to `best_hyp` inside the method). So we do not need to call _decode_stream_output manually.
                # For CTC models, if we want full hypotheses with timestamps, we still need to decode manually.
                if hasattr(self.model, 'joint'):
                    # The hypotheses are directly returned in `transcribed_texts`
                    pass
                else:
                    transcribed_texts = self._decode_stream_output(session, session.pred_out_stream)
                
                if transcribed_texts and transcribed_texts[0]:
                    hyp = transcribed_texts[0]
                    text = hyp.text if hasattr(hyp, 'text') else str(hyp)
                    timestamps = hyp.timestamp if hasattr(hyp, 'timestamp') else None
                    return text, timestamps

                return session.current_text, session.current_timestamps

        except Exception as e:
            logger.error(f"Session {session.id} chunk processing error: {e}")
            import traceback
            logger.error(traceback.format_exc())
            return None

    async def _reset_session(self, session: ASRSession, finalize: bool = True):
        """Handle reset with soft or hard finalization."""
        import time

        # Log audio state at reset for diagnostics
        audio_samples = len(session.accumulated_audio) if session.accumulated_audio is not None else 0
        audio_duration_ms = (audio_samples * 1000) // self.sample_rate
        logger.debug(
            f"Session {session.id} {'hard' if finalize else 'soft'} reset: "
            f"accumulated={audio_samples} samples ({audio_duration_ms}ms), "
            f"emitted={session.emitted_frames} frames"
        )

        if not finalize:
            text = session.current_text
            timestamps = session.current_timestamps
            
            formatted_timestamps = []
            if timestamps:
                if isinstance(timestamps, dict):
                    for key, val in timestamps.items():
                        if key != 'timestep':
                            formatted_timestamps.append({key: normalize_timestamp_output(val)})
                elif isinstance(timestamps, list):
                    formatted_timestamps = timestamps

            await session.websocket.send_str(json.dumps({
                "type": "transcript",
                "text": text,
                "timestamps": formatted_timestamps if formatted_timestamps else None,
                "is_final": True,
                "finalize": False
            }))

            logger.debug(f"Session {session.id} soft reset: '{text[-50:] if len(text) > 50 else text}'")
            return

        # HARD RESET: Full finalization with padding
        original_audio_length = len(session.accumulated_audio) if session.accumulated_audio is not None else 0

        if original_audio_length > 0:
            padding_samples = self.final_padding_frames * self.hop_samples
            silence_padding = np.zeros(padding_samples, dtype=np.float32)
            session.accumulated_audio = np.concatenate([session.accumulated_audio, silence_padding])

        # Process all remaining audio with keep_all_outputs=True
        final_text = session.current_text
        final_timestamps = session.current_timestamps
        if session.accumulated_audio is not None and len(session.accumulated_audio) > 0:
            start_time = time.perf_counter()
            async with self.inference_lock:
                result = await asyncio.get_event_loop().run_in_executor(
                    None, self._process_final_chunk, session
                )
                if result is not None:
                    final_text, final_timestamps = result
                    session.current_text = final_text
                    session.current_timestamps = final_timestamps
            elapsed_ms = (time.perf_counter() - start_time) * 1000
            logger.debug(f"Session {session.id} final chunk processed in {elapsed_ms:.1f}ms: '{final_text[-50:] if len(final_text) > 50 else final_text}'")

        # Server-side deduplication: only send the delta (new portion)
        if final_text.startswith(session.last_emitted_text):
            delta_text = final_text[len(session.last_emitted_text):].lstrip()
        else:
            delta_text = final_text

        session.last_emitted_text = final_text
        
        formatted_timestamps = []
        if final_timestamps:
            if isinstance(final_timestamps, dict):
                for key, val in final_timestamps.items():
                    if key != 'timestep':
                        formatted_timestamps.append({key: normalize_timestamp_output(val)})
            elif isinstance(final_timestamps, list):
                formatted_timestamps = final_timestamps

        # Send only the delta to client
        await session.websocket.send_str(json.dumps({
            "type": "transcript",
            "text": delta_text,
            "timestamps": formatted_timestamps if formatted_timestamps else None,
            "is_final": True,
            "finalize": True
        }))

        session.last_emitted_text = ""
        session.overlap_buffer = None
        self._init_session(session)

    def _process_final_chunk(self, session: ASRSession) -> Optional[Tuple[str, Optional[dict]]]:
        """Process all remaining audio with keep_all_outputs=True."""
        try:
            if len(session.accumulated_audio) == 0:
                return session.current_text, session.current_timestamps

            # Preprocess ALL accumulated audio
            audio_tensor = torch.from_numpy(session.accumulated_audio).unsqueeze(0).cuda()
            audio_len = torch.tensor([len(session.accumulated_audio)], device='cuda')

            with torch.inference_mode():
                mel, mel_len = self.model.preprocessor(
                    input_signal=audio_tensor,
                    length=audio_len
                )

                # For final chunk, use ALL remaining frames (including edge)
                total_mel_frames = mel.shape[-1]
                remaining_frames = total_mel_frames - session.emitted_frames

                if remaining_frames <= 0:
                    return session.current_text, session.current_timestamps

                # Extract final chunk with pre-encode cache
                if session.emitted_frames == 0:
                    chunk_start = 0
                    drop_extra = 0
                else:
                    chunk_start = session.emitted_frames - self.pre_encode_cache_size
                    drop_extra = self.drop_extra

                chunk_mel = mel[:, :, chunk_start:]
                chunk_len = torch.tensor([chunk_mel.shape[-1]], device='cuda')

                (
                    session.pred_out_stream,
                    transcribed_texts,
                    session.cache_last_channel,
                    session.cache_last_time,
                    session.cache_last_channel_len,
                    session.previous_hypotheses,
                ) = self.model.conformer_stream_step(
                    processed_signal=chunk_mel,
                    processed_signal_length=chunk_len,
                    cache_last_channel=session.cache_last_channel,
                    cache_last_time=session.cache_last_time,
                    cache_last_channel_len=session.cache_last_channel_len,
                    keep_all_outputs=True,  # Final chunk - output all remaining
                    previous_hypotheses=session.previous_hypotheses,
                    previous_pred_out=session.pred_out_stream,
                    drop_extra_pre_encoded=drop_extra,
                    return_transcription=True,
                )

                if hasattr(self.model, 'joint'):
                    pass
                else:
                    transcribed_texts = self._decode_stream_output(session, session.pred_out_stream)
                
                if transcribed_texts and transcribed_texts[0]:
                    hyp = transcribed_texts[0]
                    text = hyp.text if hasattr(hyp, 'text') else str(hyp)
                    timestamps = hyp.timestamp if hasattr(hyp, 'timestamp') else None
                    return text, timestamps

                return session.current_text, session.current_timestamps

        except Exception as e:
            logger.error(f"Session {session.id} final chunk error: {e}")
            import traceback
            logger.error(traceback.format_exc())
            return None

    async def health_handler(self, request: web.Request) -> web.Response:
        """Health check endpoint."""
        return web.json_response({
            "status": "healthy" if self.model_loaded else "loading",
            "model_loaded": self.model_loaded,
        })

    async def start(self):
        """Start the HTTP + WebSocket server."""
        self.load_model()
        self.model_loaded = True

        logger.info(f"Starting streaming ASR server on ws://{self.host}:{self.port}")

        app = web.Application()
        app.router.add_get("/health", self.health_handler)
        app.router.add_get("/", self.websocket_handler)

        runner = web.AppRunner(app)
        await runner.setup()
        site = web.TCPSite(runner, self.host, self.port)
        await site.start()

        logger.info(f"ASR server listening on ws://{self.host}:{self.port}")
        logger.info(f"Health check available at http://{self.host}:{self.port}/health")
        await asyncio.Future()  # Run forever


def main():
    parser = argparse.ArgumentParser(description="Nemotron Streaming ASR WebSocket Server")
    parser.add_argument("--host", default="0.0.0.0", help="Host to bind to")
    parser.add_argument("--port", type=int, default=8080, help="Port to bind to")
    parser.add_argument(
        "--model",
        default=DEFAULT_MODEL,
        help="HuggingFace model name or path to local .nemo file"
    )
    parser.add_argument(
        "--right-context",
        type=int,
        default=1,
        choices=[0, 1, 6, 13],
        help="Right context frames: 0=80ms, 1=160ms, 6=560ms, 13=1.12s latency"
    )
    args = parser.parse_args()

    server = ASRServer(
        model=args.model,
        host=args.host,
        port=args.port,
        right_context=args.right_context,
    )

    asyncio.run(server.start())


if __name__ == "__main__":
    main()