ModelAufgelistetExtern

Qwen3 ForcedAligner 0.6B BF16

Full-precision (bf16→float16) conversion of Qwen/Qwen3-ForcedAligner-0.6B for Apple Silicon inference via MLX.

Quellehuggingface.co/aufklarer/Qwen3-ForcedAligner-0.6B-bf16/tree/bc14e9b89e34586b0fe86903d8171187a57adb1d

ArtText modelBasismodellQwen/Qwen3-ForcedAligner-0.6BVersionmodelBBewertung

Modellquelle

Art: Text model
Dateigröße: 1,71 GB
Basismodell: Qwen/Qwen3-ForcedAligner-0.6B
Version: vmodel
Aktualisiert: 26. Sept. 2026
Downloads: 20
Quelle: Hugging Face

Quellenbeschreibung

Qwen3-ForcedAligner-0.6B — MLX bf16

Full-precision (bf16→float16) conversion of Qwen/Qwen3-ForcedAligner-0.6B for Apple Silicon inference via MLX.

No quantization applied — maximum accuracy for word-level timestamp prediction.

Model Details

Detail	Value
Audio encoder	24 layers, 1024 dim, 16 heads, float16
Text decoder	28 layers, 1024 hidden, 16Q/8KV heads, float16
Classify head	Linear(1024, 5000), float16
Timestamp resolution	80ms per class
Total size	~1.8 GB

Usage

let aligner = try await Qwen3ForcedAligner.fromPretrained(
    modelId: "aufklarer/Qwen3-ForcedAligner-0.6B-bf16"
)
let aligned = aligner.align(
    audio: samples, text: "Hello world", sampleRate: 24000
)

Variants

Variant	Size	Model ID
4-bit	~979 MB	aufklarer/Qwen3-ForcedAligner-0.6B-4bit
8-bit	~1.4 GB	aufklarer/Qwen3-ForcedAligner-0.6B-8bit
bf16	~1.8 GB	aufklarer/Qwen3-ForcedAligner-0.6B-bf16

Links

Swift library: soniqo/speech-swift
Base model: Qwen/Qwen3-ForcedAligner-0.6B

Quellen

1 Quelle

Hugging FacePrimärVersion bc14e9b89e34586b0fe86903d8171187a57adb1d

huggingface.co/aufklarer/Qwen3-ForcedAligner-0.6B-bf16/tree/bc14e9b89e34586b0fe86903d8171187a57adb1d

Verifiziert 29. Aug.

Modellartefakte

1 Artefakt

model.safetensors

safetensors · 1,71 GB · SHA-256 9b4c0a01fb59…121f · Hugging Face

Download file

Quellenauszüge

2 Auszüge

Readme29. Aug.huggingface.co/aufklarer/Qwen3-ForcedAligner-0.6B-bf16

--- license: apache-2.0 language: - en - zh - ja - ko - de - fr - es - it - ru tags: - mlx - forced-alignment - speech - qwen3 - timestamps - bf16 library_name: mlx base_model: Qwen/Qwen3-ForcedAligner-0.6B pipeline_tag: audio-classification --- # Qwen3-ForcedAligner-0.6B — MLX bf16 Full-precision (bf16→float16) conversion of [Qwen/Qwen3-ForcedAligner-0.6B](https://huggingface.co/Qwen/Qwen3-ForcedAligner-0.6B) for Apple Silicon inference via MLX. No quantization applied — maximum accuracy for word-level timestamp prediction. ## Model Details | Detail | Value | |--------|-------| | Audio encoder | 24 layers, 1024 dim, 16 heads, float16 | | Text decoder | 28 layers, 1024 hidden, 16Q/8KV heads, float16 | | Classify head | Linear(1024, 5000), float16 | | Timestamp resolution | 80ms per class | | Total size | ~1.8 GB | ## Usage ```swift let aligner = try await Qwen3ForcedAligner.fromPretrained( modelId: "aufklarer/Qwen3-ForcedAligner-0.6B-bf16" ) let aligned = aligner.align( audio: samples, text: "Hello world", sampleRate: 24000 ) ``` ## Variants | Variant | Size | Model ID | |---------|------|----------| | 4-bit | ~979 MB | [aufklarer/Qwen3-ForcedAligner-0.6B-4bit](https://huggingface.co/aufklarer/Qwen3-ForcedAligner-0.6B-4bit) | | 8-bit | ~1.4 GB | [aufklarer/Qwen3-ForcedAligner-0.6B-8bit](https://huggingface.co/aufklarer/Qwen3-ForcedAligner-0.6B-8bit) | | **bf16** | **~1.8 GB** | **aufklarer/Qwen3-ForcedAligner-0.6B-bf16** | ## Links - **Swift library**: [soniqo/speech-swift](https://github.com/soniqo/speech-swift) - **Base model**: [Qwen/Qwen3-ForcedAligner-0.6B](https://huggingface.co/Qwen/Qwen3-ForcedAligner-0.6B) --- --- - **Guide**: [soniqo.audio/guides/align](https://soniqo.audio/guides/align) - **Docs**: [soniqo.audio](https://soniqo.audio) - **GitHub**: [soniqo/speech-swift](https://github.com/soniqo/speech-swift)

Note29. Aug.huggingface.co/aufklarer/Qwen3-ForcedAligner-0.6B-bf16

Source context: 20 downloads · 0 likes · Pipeline audio-classification · Library mlx · Repo aufklarer/Qwen3-ForcedAligner-0.6B-bf16