美女全套:SRC(2.7W) DST(6.3W) DFM Model(812W LOSS<0.15,可做底丹,切换SRC一秒出脸) 简介 专享一:美女专丹model文件夹下所有模型文件(LOSS已降至1.5) 专享二:美女直播DFM文件(迭代次数812W,LOSS<1.5) 专项三:DST文件(切好脸,已应用万能遮罩): DST人脸63000张,角度全、光影丰富,像素768*768,全部应用了镇坛之宝万能遮罩,文件大小8.74G! 专享四:SRC文件(切好脸,已应用万能遮罩): DFL脚本赛选约5000张最优人脸SRC(分辨率512*512及1024*1024都有),备用22000 张(分辨率512*512及1024*1024都有) 下载后的使用方法: 1、作为专丹可以继续训练: 1)常规训练:DST人脸文件放入DST的aligned文件夹,SRC人脸文件放入SRC的aligned文件夹,model文件放入model文件夹,点击6) 训练重量级模型 train SAEHD,按照7步法调整参数优化; 2)专题特训:挑选带上/下牙的DST人脸文件放入DST的aligned文件夹,挑选带上/下牙的SRC人脸文件放入SRC的aligned文件夹,model文件放入model文件夹,点击6) 训练重量级模型 train SAEHD,按照7步法调整参数优化(见截图); 2、DFM放入Deepfacelive使用; 3、SRC(切好脸,已应用万能遮罩)人脸文件放入你的其他分辨率模型进行训练 4、DST(切好脸,已应用万能遮罩)人脸文件作为其他模型训练的素材之一
================== Model Summary ===================
== ==
== Model name: new_SAEHD ==
== ==
== Current iteration: 8128862 ==
== ==
==---------------- Model Options -----------------==
== ==
== resolution: 224 ==
== face_type: wf ==
== models_opt_on_gpu: True ==
== archi: liae-udt ==
== ae_dims: 512 ==
== e_dims: 64 ==
== d_dims: 64 ==
== d_mask_dims: 32 ==
== masked_training: True ==
== eyes_mouth_prio: False ==
== uniform_yaw: False ==
== blur_out_mask: True ==
== adabelief: True ==
== lr_dropout: n ==
== random_warp: False ==
== random_hsv_power: 0.0 ==
== true_face_power: 0.0 ==
== face_style_power: 0.0 ==
== bg_style_power: 0.0 ==
== ct_mode: rct ==
== clipgrad: True ==
== pretrain: False ==
== autobackup_hour: 0 ==
== write_preview_history: False ==
== target_iter: 0 ==
== random_src_flip: False ==
== random_dst_flip: False ==
== batch_size: 8 ==
== gan_power: 0.0 ==
== gan_patch_size: 32 ==
== gan_dims: 16 ==
== ==
==------------------ Running On ------------------==
== ==
== Device index: 0 ==
== Name: NVIDIA GeForce RTX 4090 ==
== VRAM: 20.85GB ==
== ==
====================================================
|