Anony94
/

FLARE-checkpoints

Model card Files Files and versions

FLARE

Fairseq model checkpoints for the paper:

Learning Visual Speech Representations via Cross-Modal Distillation and Joint Face-Lip Modeling

Overview

This repository provides model checkpoints for two variants:

FLARE: pretrained and lightweight Transformer finetuned checkpoints
FLARE-LLM: LLM-based finetuned checkpoints

FLARE Checkpoints

Large

Pretrained:
pretrained_models/exp_22_2_large/checkpoints/checkpoint_52_800000.pt
Finetuned:
pretrained_models/exp_22_2_large/finetuned/checkpoints/checkpoint_last.pt

Base

Pretrained:
pretrained_models/exp_22_base/checkpoints/checkpoint_51_800000.pt
Finetuned:
pretrained_models/exp_22_base/finetuned/checkpoints/checkpoint_last.pt

FLARE-LLM Checkpoints

Large

Stage 3 checkpoint:
pretrained_models/lipreading_exp_2_large_face_lip_s2/checkpoints/checkpoint_4_60000.pt

Base

Stage 3 checkpoint:
pretrained_models/lipreading_exp_4_base_face_lip_s2/checkpoints/checkpoint_4_60000.pt

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Anony94/FLARE-checkpoints

Quantizations