You need to agree to share your Hugging Face information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

lfm2a-fr `s12-unique-sft`, transformers layout

baptistefrancois1/lfm2a-fr (revision s12-unique-sft: LFM2.5-Audio-1.5B fully fine-tuned for a French and English voice assistant, SIWIS voice in French) converted to the native transformers layout with convert_lfm2_audio_to_hf of PR #48249 (commit 4ac58b3). No additional training. Weights, config, processor and the bundled audio detokenizer are included.

Loads with that branch of transformers:

from transformers import AutoProcessor, Lfm2AudioForConditionalGeneration
model = Lfm2AudioForConditionalGeneration.from_pretrained("Rcarvalo/lfm2a-fr-s12-hf")
processor = AutoProcessor.from_pretrained("Rcarvalo/lfm2a-fr-s12-hf")

Used as the teacher of the distillation into a 230M backbone in the Unsloth tutorial (finetuning_liquidai/tutorials/unsloth-lfm2.5-audio). License of the weights: LFM Open License v1.0 (see LICENSE).

Downloads last month: 47

Safetensors

Model size

1B params

Tensor type

BF16

Model tree for Rcarvalo/lfm2a-fr-s12-hf

Base model

baptistefrancois1/lfm2a-fr

Finetuned

(1)

this model

You need to agree to share your Hugging Face information to access this model

lfm2a-fr s12-unique-sft, transformers layout

Model tree for Rcarvalo/lfm2a-fr-s12-hf

lfm2a-fr `s12-unique-sft`, transformers layout