[WACV 2026] AuViRe: Audio-visual Speech Representation Reconstruction for Deepfake Temporal Localization (Model Checkpoints)

There is a newer version of the record available.

Published November 24, 2025 | Version v1

Model Open

Model checkpoints of the WACV 2026 paper "AuViRe: Audio-visual Speech Representation Reconstruction for Deepfake Temporal Localization".

Files

Name	Size
avdeepfake1m_b_avhubert_t_cnn_cnn_h_8_d_128_l_r1d1_w_15_o_subtraction_rl_r2d1u1s2_rm_av_aa_vv_f_True_conv_lr-_c_focal_diou_rec.pth md5:098cb9d0676276e3705a5ab5f57507ad	107.4 MB	Download
lavdf_b_avhubert_t_cnn_cnn_h_8_d_128_l_r2d2_w_15_o_subtraction_rl_r2d3u3s2_rm_av_aa_vv_f_True_conv_lr-_c_focal_diou_rec.pth md5:a694ab3fce5a1706f03d51d7f04e0261	145.5 MB	Download

European Commission
AI4TRUST - AI-based-technologies for trustworthy solutions against disinformation 101070190
European Commission
AI-CODE - AI-CODE - AI services for COntinuous trust in emerging Digital Environments 101135437

335

Views

340

Downloads

Show more details

DOI

Resource type

Model

Publisher

Zenodo

Conference

(WACV), 2026

License: Apache License 2.0

A permissive license whose main conditions require preservation of copyright and license notices. Contributors provide an express grant of patent rights. Licensed works, modifications, and larger works may be distributed under different terms and without source code. Read more