[Alex-Wengg]

Reverts #1083 (4e8b59d). Draft: merge only if the crash can't be confirmed.

Why

The maintainers couldn't reproduce what #1083 fixes. The crash needs a process built with the macOS 27 SDK and run on macOS 26 or earlier, where Core AI is missing. Then objc_copyClassList completes the metadata for LFM3AudioCoreAIBackbone and LFM3AudioCoreAIDepthDecoder and segfaults.

• The maintainers' only Mac runs macOS 27.0, where Core AI is present.
• CI's macos-15 runners build with an older SDK, which doesn't compile the Core AI path. Build and Test passed on main both before and after fix(lfm3-audio): keep Core AI types out of class metadata on older systems #1083.

What was verified for #1083 is that it changes nothing on macOS 27: LFM3-Audio --backend coreai gives byte-identical TTS and identical ASR text and token counts.

Risk of merging this

The PR author's crash trace and before/after run on macOS 26.7 are specific: swift test dies with signal 11 in swiftpm-xctest-helper, and downstream apps' XCTest runs crash at class enumeration. Merging this brings that crash back for anyone building with Xcode 27 and testing on macOS 26. Prefer reproducing on a macOS 26 machine or VM, and keeping #1083 if it reproduces.

🤖 Generated with Claude Code

[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	18.80x	>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	11.792	21.1	Fetching diarization models
Model Compile	5.054	9.1	CoreML compilation
Audio Load	0.071	0.1	Loading audio file
Segmentation	16.736	30.0	Detecting speech regions
Embedding	27.893	50.0	Extracting speaker voices
Clustering	11.157	20.0	Grouping same speakers
Total	55.820	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 • 55.8s diarization time • Test runtime: 4m 3s • 11/01/2026, 02:32 PM EST}

[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	8.09x	Real-time factor (higher = faster)
Total Audio	470.6s	Total audio duration processed
Total Time	60.1s	Total processing time

Streaming Metrics

Metric	Value	Description
Avg Chunk Time	0.060s	Average chunk processing time
Max Chunk Time	0.120s	Maximum chunk processing time
EOU Detections	0	Total End-of-Utterance detections

_{Test runtime: 1m28s • 11/01/2026, 02:35 PM EST}

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

[github-actions]

Supertonic3 Smoke Test ✅

Check	Result
Build	✅
Model download (incl. VectorEstimatorVariants/ int4 buckets)	✅
Model load	✅
Synthesis pipeline (--ve-variant int4)	✅
Output WAV	✅ (368.5 KB)

_{Runtime: 0m26s}

_{Note: CI VMs lack a physical Neural Engine; the ANE-bucketed VectorEstimator falls back to CPU here. This validates download + variant resolution + synthesis, not ANE residency/perf.}

[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%	3.08x	✅
test-other	1.19%	0.00%	3.44x	✅

Parakeet v2 (English-optimized)

Dataset	WER Avg	WER Med	RTFx	Status
test-clean	0.80%	0.00%	4.72x	✅
test-other	1.00%	0.00%	3.23x	✅

Streaming (v3)

Metric	Value	Description
WER	0.00%	Word Error Rate in streaming mode
RTFx	0.57x	Streaming real-time factor
Avg Chunk Time	1.618s	Average time to process each chunk
Max Chunk Time	2.013s	Maximum chunk processing time
First Token	1.966s	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	0.55x	Streaming real-time factor
Avg Chunk Time	1.618s	Average time to process each chunk
Max Chunk Time	2.121s	Maximum chunk processing time
First Token	1.712s	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: 9m16s • 11/01/2026, 03:52 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}