Voxtral-Mini-3B-2507-AWQ4-GEN

Home/Models/juanginer/Voxtral-Mini-3B-2507-AWQ4-GEN

juanginer/Voxtral-Mini-3B-2507-AWQ4-GEN

kind: awq
quant: awq
family: misc
modality: automatic-speech-recognition
root model: mistralai/Voxtral-Mini-3B-2507
provider: juanginer
license: apache-2.0
language: en, es, fr, de, it, pt, pl, nl
context: —
library: vllm
vRAM (GB): —
size on disk: —
downloads: 495
likes: 1
created: —
last modified: 2026-07-20T00:35:02+00:00 (2026)
huggingface: https://huggingface.co/juanginer/Voxtral-Mini-3B-2507-AWQ4-GEN
tags: vllmsafetensorsvoxtralvllm-compatibleautomatic-speech-recognitionspeechaudioawqw4a16quantizationllm-compressorenesfrdeitptplnlbase_model:mistralai/Voxtral-Mini-3B-2507base_model:quantized:mistralai/Voxtral-Mini-3B-2507license:apache-2.0compressed-tensorsregion:us

benchmark	category	score	context	conf	evidence
WER Common Voice	speech-asr	15.0		medium	`Common Voice 15`
WER	speech-asr	9.24	WER nl	medium	`\| WER nl \| 9.24 \| 7.95 \| 11.29 \| 10.59 \|`
WER	speech-asr	8.01	WER fr	medium	`\| WER fr \| 8.01 \| 10.59 \| 6.70 \| 9.33 \|`
WER	speech-asr	7.16	WER pt	medium	`\| WER pt \| 7.16 \| 7.21 \| 6.14 \| 6.08 \|`
WER	speech-asr	7.13	WER avg-6	medium	`\| WER avg-6 \| 7.13 \| 7.43 \| 7.07 \| 7.61 \|`
WER	speech-asr	6.65	WER en (no MLS)	medium	`\| WER en (no MLS) \| 6.65 \| 7.00 \| 6.96 \| 7.01 \|`
WER	speech-asr	6.5	WER it	medium	`\| WER it \| 6.50 \| 7.02 \| 7.01 \| 6.94 \|`
WER	speech-asr	6.06	WER de	medium	`\| WER de \| 6.06 \| 7.13 \| 6.66 \| 8.16 \|`
WER	speech-asr	5.82	WER es	medium	`\| WER es \| 5.82 \| 4.68 \| 4.60 \| 4.55 \|`
WER	speech-asr	4.69	WER es-AR	medium	`\| WER es-AR \| 4.69 \| 5.00 \| 5.11 \| 5.04 \|`

none scraped.