HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks

cohere-transcribe-arabic-07-2026-MLX-Mixed-2bit3bit4bit

MarkChen1214/cohere-transcribe-arabic-07-2026-MLX-Mixed-2bit3bit4bit

kind
mlx
quant
mlx
family
command
modality
automatic-speech-recognition
root model
CohereLabs/cohere-transcribe-arabic-07-2026
provider
MarkChen1214
license
mit
language
ar, en
context
—
library
mlx
vRAM (GB)
—
size on disk
—
downloads
240
likes
1
created
—
last modified
2026-07-18T16:07:53+00:00 (2026)
huggingface
https://huggingface.co/MarkChen1214/cohere-transcribe-arabic-07-2026-MLX-Mixed-2bit3bit4bit
tags
mlxsafetensorscohere_asrautomatic-speech-recognitionspeech-recognitiontranscriptionaudioapplemacoson-devicequantizedmixed-precisiondynamic-quantizationarabicarabic-asrarabic-dialectcode-switching2bit3bit4bitarenbase_model:CohereLabs/cohere-transcribe-arabic-07-2026base_model:quantized:CohereLabs/cohere-transcribe-arabic-07-2026

Benchmark scores (4)

benchmarkcategoryscorecontextconfevidence
WERspeech-asr5.09low| FLEURS ar_eg test | 428 | 5.09% | 19.0x |
WER FLEURSspeech-asr5.09FLEURS ar_eg testmedium| FLEURS ar_eg test | 428 | 5.09% | 19.0x |
WER LibriSpeechspeech-asr2.55LibriSpeech test-othermedium| LibriSpeech test-other | 2,939 | 2.55% | 18.9x |
WER LibriSpeechspeech-asr1.14LibriSpeech test-cleanmedium| LibriSpeech test-clean | 2,620 | 1.14% | 17.9x |

Discussions (0)

none scraped.

Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.