LFM2-Audio-1.5B

Home/Models/LiquidAI/LFM2-Audio-1.5B

LiquidAI/LFM2-Audio-1.5B

kind: original
quant: bf16 (16-bit)
family: misc
modality: audio-to-audio
root model: LiquidAI/LFM2-1.2B
provider: LiquidAI
license: other
language: en
context: —
library: liquid-audio
vRAM (GB): —
size on disk: —
downloads: 253
likes: 358
created: —
last modified: 2026-03-27T16:11:29+00:00 (2026)
huggingface: https://huggingface.co/LiquidAI/LFM2-Audio-1.5B
tags: liquid-audiosafetensorsliquidlfm2audiolfm2-audiospeech-to-speechaudio-to-audioenarxiv:2511.23404base_model:LiquidAI/LFM2-1.2Bbase_model:finetune:LiquidAI/LFM2-1.2Blicense:otherregion:us

benchmark	category	score	context	conf	evidence
IFEval	instruction-following	98.85		high	`\| LFM2-Audio-1.5B \| 1.5B parameters \| 3.71 \| 3.49 \| 3.17 \| 30.56 \| 31.95 \| 44.40 \| 30.54 \| 98.85 \| 67.33 \| 56.78 \|`
BBH	reasoning	30.54		high	`\| LFM2-Audio-1.5B \| 1.5B parameters \| 3.71 \| 3.49 \| 3.17 \| 30.56 \| 31.95 \| 44.40 \| 30.54 \| 98.85 \| 67.33 \| 56.78 \|`
WER AMI	speech-asr	15.58	Open	high	`\| LFM2-Audio-1.5B \| 1.5B parameters \| Yes \| Yes \| 15.58 \| 10.67 \| 2.01 \| 4.39 \| 3.56 \| 7.24 \|`
WER GigaSpeech	speech-asr	10.67	Open	high	`\| LFM2-Audio-1.5B \| 1.5B parameters \| Yes \| Yes \| 15.58 \| 10.67 \| 2.01 \| 4.39 \| 3.56 \| 7.24 \|`
AlpacaEval	chat	3.71		high	`\| LFM2-Audio-1.5B \| 1.5B parameters \| 3.71 \| 3.49 \| 3.17 \| 30.56 \| 31.95 \| 44.40 \| 30.54 \| 98.85 \| 67.33 \| 56.78 \|`
WER TED-LIUM	speech-asr	3.56	Open	high	`\| LFM2-Audio-1.5B \| 1.5B parameters \| Yes \| Yes \| 15.58 \| 10.67 \| 2.01 \| 4.39 \| 3.56 \| 7.24 \|`
WER LibriSpeech	speech-asr	2.01	Open	high	`\| LFM2-Audio-1.5B \| 1.5B parameters \| Yes \| Yes \| 15.58 \| 10.67 \| 2.01 \| 4.39 \| 3.56 \| 7.24 \|`

none scraped.