Skip to content

Mega-ASR-bf16

mlx-community/Mega-ASR-bf16
bf16 Open on Hugging Face
Parameters
2.0B
Weights (est.)
≈ 4.1 GB
Mac memory (est.)
8 GB+

Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.

Details

Base model
Qwen/Qwen3-ASR-1.7B (finetune)
Vendor
Qwen
Task
Speech recognition
Quantisation
bf16
Licence
apache-2.0
Downloads
78 in the last 30 days · 673 all time · 1 likes
Published
2026-05-22 · updated 2026-05-23 (5mo ago)
Languages
en, zh
Tags
mega_asr, speech-to-text, robust-asr, qwen3-asr

Other MLX builds of Qwen/Qwen3-ASR-1.7B

Build Quantisation Weights Mac memory Downloads
Mega-ASR-6bit 6-bit ≈ 2.0 GB 8 GB+ 26
Mega-ASR-8bit 8-bit ≈ 2.5 GB 8 GB+ 49
Mega-ASR-bf16 (this one) bf16 ≈ 4.1 GB 8 GB+ 78