Standardize model loading API across all ASR managers

[Alex-Wengg]

Summary

Standardizes the model loading API across all ASR managers to reduce developer cognitive load and improve consistency. This addresses issue #457 comment #4203327648.

Problem

Each ASR manager had different model loading APIs:

• AsrManager: configure(models:) ❌
• SlidingWindowAsrManager: startStreaming(models:, source:) ❌
• StreamingEouAsrManager: loadModels(modelDir:) with inconsistent overloads ❌
• StreamingNemotronAsrManager: loadModels(modelDir:) ❌

This created developer confusion and increased documentation burden.

Solution

Unified API pattern across all managers:

// All managers now use consistent naming:
manager.loadModels(from: URL)                    // Load from local directory
manager.loadModels(_ models: PreloadedModels)    // Use pre-loaded models
manager.loadModels(to: URL?, progressHandler:)   // Download and load (optional)

Changes

• AsrManager: Added loadModels(_:), deprecated configure(models:)
• SlidingWindowAsrManager: Separated model loading from streaming activation, added CoreML import
• StreamingEouAsrManager: Standardized to loadModels(from:)
• StreamingNemotronAsrManager: Standardized to loadModels(from:) with download support
• CLI: Updated 9 command files to use new APIs

Verification

Ran full benchmark suite (8 models × 100 files) to verify zero regression:

Model	Baseline	Current	Delta	Status
Parakeet TDT v3 (0.6B)	2.6%	2.64%	+0.04%	✅
Parakeet TDT v2 (0.6B)	3.8%	3.79%	-0.01%	✅
CTC-TDT 110M	3.6%	3.56%	-0.04%	✅
CTC Earnings	16.54%	16.55%	+0.01%	✅
EOU 320ms (120M)	7.11%	7.11%	0.00%	✅
Nemotron 1120ms (0.6B)	1.99%	1.99%	0.00%	✅
TDT Japanese (0.6B)	6.11%	6.11%	0.00%	✅
CTC Chinese (0.6B)	8.37%	8.37%	0.00%	✅

✓ No WER/CER regressions (all within 0.3% of baseline)

Benefits

• ✅ Reduced cognitive load - single pattern across all managers
• ✅ Cleaner separation of concerns - model loading vs. streaming activation
• ✅ Consistent prepositions - all use from: for loading from directory
• ✅ Zero performance impact - validated with comprehensive benchmarks
• ✅ Backward compatibility - deprecated APIs still work with migration warnings

---

[github-actions]

VAD Benchmark Results

Performance Comparison

Dataset	Accuracy	Precision	Recall	F1-Score	RTFx	Files
MUSAN	92.0%	86.2%	100.0%	92.6%	382.4x faster	50
VOiCES	92.0%	86.2%	100.0%	92.6%	487.6x faster	50

Dataset Details

• MUSAN: Music, Speech, and Noise dataset - standard VAD evaluation
• VOiCES: Voices Obscured in Complex Environmental Settings - tests robustness in real-world conditions

✅: Average F1-Score above 70%

[github-actions]

Sortformer High-Latency Benchmark Results

ES2004a Performance (30.4s latency config)

Metric	Value	Target	Status
DER	33.4%	<35%	✅
Miss Rate	24.4%	-	-
False Alarm	0.2%	-	-
Speaker Error	8.8%	-	-
RTFx	12.6x	>1.0x	✅
Speakers	4/4	-	-

_{Sortformer High-Latency • ES2004a • Runtime: 2m 53s • 2026-04-08T18:00:42.673Z}

[github-actions]

PocketTTS Smoke Test ✅

Check	Result
Build	✅
Model download	✅
Model load	✅
Synthesis pipeline	✅
Output WAV	✅ (202.5 KB)

_{Runtime: 0m38s}

_{Note: PocketTTS uses CoreML MLState (macOS 15) KV cache + Mimi streaming state. CI VM lacks physical GPU — audio quality and performance may differ from Apple Silicon.}

[github-actions]

Kokoro TTS Smoke Test ✅

Check	Result
Build	✅
Model download	✅
Model load	✅
Synthesis pipeline	✅
Output WAV	✅ (634.8 KB)

_{Runtime: 0m33s}

_{Note: Kokoro TTS uses CoreML flow matching + Vocos vocoder. CI VM lacks physical ANE — performance may differ from Apple Silicon.}

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 4 additional findings.

[github-actions]

Parakeet EOU Benchmark Results ✅

Status: Benchmark passed
Chunk Size: 320ms
Files Tested: 100/100

Performance Metrics

Metric	Value	Description
WER (Avg)	7.03%	Average Word Error Rate
WER (Med)	4.17%	Median Word Error Rate
RTFx	6.92x	Real-time factor (higher = faster)
Total Audio	470.6s	Total audio duration processed
Total Time	71.2s	Total processing time

Streaming Metrics

Metric	Value	Description
Avg Chunk Time	0.071s	Average chunk processing time
Max Chunk Time	0.142s	Maximum chunk processing time
EOU Detections	0	Total End-of-Utterance detections

_{Test runtime: 1m18s • 04/08/2026, 01:58 PM EST}

_{RTFx = Real-Time Factor (higher is better) • Processing includes: Model inference, audio preprocessing, state management, and file I/O}

[github-actions]

Qwen3-ASR int8 Smoke Test ✅

Check	Result
Build	✅
Model download	✅
Model load	✅
Transcription pipeline	✅
Decoder size	571 MB (vs 1.1 GB f32)

Performance Metrics

Metric	CI Value	Expected on Apple Silicon
Median RTFx	0.04x	~2.5x
Overall RTFx	0.04x	~2.5x

_{Runtime: 4m54s}

_{Note: CI VM lacks physical GPU — CoreML MLState (macOS 15) KV cache produces degraded results on virtualized runners. On Apple Silicon: ~1.3% WER / 2.5x RTFx.}

[github-actions]

ASR Benchmark Results ✅

Status: All benchmarks passed

Parakeet v3 (multilingual)

Dataset	WER Avg	WER Med	RTFx	Status
test-clean	0.57%	0.00%	4.60x	✅
test-other	1.35%	0.00%	2.92x	✅

Parakeet v2 (English-optimized)

Dataset	WER Avg	WER Med	RTFx	Status
test-clean	0.80%	0.00%	4.07x	✅
test-other	1.22%	0.00%	2.68x	✅

Streaming (v3)

Metric	Value	Description
WER	0.00%	Word Error Rate in streaming mode
RTFx	0.47x	Streaming real-time factor
Avg Chunk Time	1.880s	Average time to process each chunk
Max Chunk Time	2.504s	Maximum chunk processing time
First Token	2.194s	Latency to first transcription token
Total Chunks	31	Number of chunks processed

Streaming (v2)

Metric	Value	Description
WER	0.00%	Word Error Rate in streaming mode
RTFx	0.44x	Streaming real-time factor
Avg Chunk Time	1.965s	Average time to process each chunk
Max Chunk Time	3.243s	Maximum chunk processing time
First Token	2.044s	Latency to first transcription token
Total Chunks	31	Number of chunks processed

_{Streaming tests use 5 files with 0.5s chunks to simulate real-time audio streaming}

_{25 files per dataset • Test runtime: 6m58s • 04/08/2026, 02:14 PM EST}

_{RTFx = Real-Time Factor (higher is better) • Calculated as: Total audio duration ÷ Total processing time
Processing time includes: Model inference on Apple Neural Engine, audio preprocessing, state resets between files, token-to-text conversion, and file I/O
Example: RTFx of 2.0x means 10 seconds of audio processed in 5 seconds (2x faster than real-time)}

Expected RTFx Performance on Physical M1 Hardware:

• M1 Mac: ~28x (clean), ~25x (other)
• CI shows ~0.5-3x due to virtualization limitations

_{Testing methodology follows HuggingFace Open ASR Leaderboard}

[github-actions]

Offline VBx Pipeline Results

Speaker Diarization Performance (VBx Batch Mode)

Optimal clustering with Hungarian algorithm for maximum accuracy

Metric	Value	Target	Status	Description
DER	14.5%	<20%	✅	Diarization Error Rate (lower is better)
RTFx	3.35x	>1.0x	✅	Real-Time Factor (higher is faster)

Offline VBx Pipeline Timing Breakdown

Time spent in each stage of batch diarization

Stage	Time (s)	%	Description
Model Download	15.032	4.8	Fetching diarization models
Model Compile	6.442	2.1	CoreML compilation
Audio Load	0.104	0.0	Loading audio file
Segmentation	28.918	9.2	VAD + speech detection
Embedding	311.852	99.6	Speaker embedding extraction
Clustering (VBx)	0.984	0.3	Hungarian algorithm + VBx clustering
Total	313.106	100	Full VBx pipeline

Speaker Diarization Research Comparison

Offline VBx achieves competitive accuracy with batch processing

Method	DER	Mode	Description
FluidAudio (Offline)	14.5%	VBx Batch	On-device CoreML with optimal clustering
FluidAudio (Streaming)	17.7%	Chunk-based	First-occurrence speaker mapping
Research baseline	18-30%	Various	Standard dataset performance

Pipeline Details:

• Mode: Offline VBx with Hungarian algorithm for optimal speaker-to-cluster assignment
• Segmentation: VAD-based voice activity detection
• Embeddings: WeSpeaker-compatible speaker embeddings
• Clustering: PowerSet with VBx refinement
• Accuracy: Higher than streaming due to optimal post-hoc mapping

_{🎯 Offline VBx Test • AMI Corpus ES2004a • 1049.0s meeting audio • 341.8s processing • Test runtime: 5m 44s • 04/08/2026, 02:09 PM EST}

[github-actions]

Speaker Diarization Benchmark Results

Speaker Diarization Performance

Evaluating "who spoke when" detection accuracy

Metric	Value	Target	Status	Description
DER	15.1%	<30%	✅	Diarization Error Rate (lower is better)
JER	24.9%	<25%	✅	Jaccard Error Rate
RTFx	20.54x	>1.0x	✅	Real-Time Factor (higher is faster)

Diarization Pipeline Timing Breakdown

Time spent in each stage of speaker diarization

Stage	Time (s)	%	Description
Model Download	10.430	20.4	Fetching diarization models
Model Compile	4.470	8.7	CoreML compilation
Audio Load	0.114	0.2	Loading audio file
Segmentation	15.321	30.0	Detecting speech regions
Embedding	25.535	50.0	Extracting speaker voices
Clustering	10.214	20.0	Grouping same speakers
Total	51.096	100	Full pipeline

Speaker Diarization Research Comparison

Research baselines typically achieve 18-30% DER on standard datasets

Method	DER	Notes
FluidAudio	15.1%	On-device CoreML
Research baseline	18-30%	Standard dataset performance

Note: RTFx shown above is from GitHub Actions runner. On Apple Silicon with ANE:

• M2 MacBook Air (2022): Runs at 150 RTFx real-time
• Performance scales with Apple Neural Engine capabilities

_{🎯 Speaker Diarization Test • AMI Corpus ES2004a • 1049.0s meeting audio • 51.1s diarization time • Test runtime: 2m 14s • 04/08/2026, 02:04 PM EST}