granite-speech-4.1-2b-gguf

Home/Models/handy-computer/granite-speech-4.1-2b-gguf

handy-computer/granite-speech-4.1-2b-gguf

kind: gguf
quant: gguf (16-bit)
family: misc
modality: automatic-speech-recognition
root model: ibm-granite/granite-speech-4.1-2b
provider: handy-computer
license: apache-2.0
language: en, fr, de, es, pt, ja
context: 128k
library: transcribe.cpp
vRAM (GB): —
size on disk: —
downloads: 41,096
likes: 0
created: —
last modified: 2026-06-29T05:08:33+00:00 (2026)
huggingface: https://huggingface.co/handy-computer/granite-speech-4.1-2b-gguf
tags: transcribe.cppggufasrspeech-to-textgranitegranite-speechaudio-llmmultilingualautomatic-speech-recognitionenfrdeesptjaarxiv:2310.02943arxiv:2505.08699arxiv:2603.11243arxiv:2604.12398arxiv:2604.22817arxiv:2604.11269arxiv:2603.08397base_model:ibm-granite/granite-speech-4.1-2bbase_model:quantized:ibm-granite/granite-speech-4.1-2b

benchmark	category	score	context	conf	evidence
WER AMI	speech-asr	100.0	AMI English	medium	`\| AMI English \| ASR \| 100 \| https://huggingface.co/datasets/edinburghcstr/ami \|`
WER VoxPopuli	speech-asr	24.86	VoxPopuli	medium	`\| VoxPopuli \| 24.86 \| 95.35 \|`
WER Earnings-22	speech-asr	22.87	Earnings-22	medium	`\| Earnings-22 \| 22.87 \| 95.19 \|`
WER	speech-asr	1.31		low	`\| BF16 \| [granite-speech-4.1-2b-BF16.gguf](https://huggingface.co/handy-computer/granite-speech-4.1-2b-gguf/resolve/main/granite-speech-4.1-2b-BF16.gguf) \| 4.63 GB \| 1.31% \|`

none scraped.