6530| 75
|
LIAE-UD_128_WF_69w迭代_4万亚洲人脸_小显存可用 |
4.8万 亚洲人脸数据(512分辨率) 迭代100w 的预训练模型。
人脸数据是我从网上爬的(已切好),见我另一个帖子:https://www.dfldata.xyz/forum.php?mod=viewthread&tid=4526&extra= 我用的batch_size是4(占用显存4.64G),减小batch后 4G卡也可以用 参数如下: =============== Model Summary =============== == == == Model name: pretrained_SAEHD == == == == Current iteration: 688068 == == == ==------------- Model Options -------------== == == == resolution: 128 == == face_type: wf == == models_opt_on_gpu: True == == archi: liae-ud == == ae_dims: 256 == == e_dims: 64 == == d_dims: 64 == == d_mask_dims: 22 == == masked_training: True == == eyes_mouth_prio: True == == uniform_yaw: True == == blur_out_mask: False == == adabelief: True == == lr_dropout: y == == random_warp: True == == true_face_power: 0.0 == == face_style_power: 0.0 == == bg_style_power: 0.0 == == ct_mode: rct == == clipgrad: True == == pretrain: False == == autobackup_hour: 5 == == write_preview_history: False == == target_iter: 0 == == random_src_flip: False == == random_dst_flip: False == == batch_size: 4 == == gan_power: 0.0 == == gan_patch_size: 16 == == gan_dims: 16 == == == ==-------------- Running On ---------------== == == == Device index: 0 == == Name: GeForce RTX 2060 == == VRAM: 4.64GB == == == 目前是4个batch size,占用了显存4.64G,如果减小batch size,就可以在4G甚至2G上使用了。 使用的是数据 4w+的亚洲人脸数据,链接:https://www.dfldata.xyz/forum.php?mod=viewthread&tid=4526&extra= 一个印度兄弟说他用不了百度网盘,让我传到mega网盘上,我也顺便传分享一个到百度网盘。 度盘链接:https://pan.baidu.com/s/1g_KMbIjV2nuFEdPZMZ9R8Q
购买主题
已有 33 人购买
本主题需向作者支付 10 灵石 才能浏览
| |
发表于 2021-12-11 18:47:30
|
显示全部楼层
提示: 作者被禁止或删除 内容自动屏蔽
| |
发表于 2021-12-11 18:52:15
|
显示全部楼层
| ||
发表于 2021-12-12 10:31:02
|
显示全部楼层
| ||
|Archiver|手机版|deepfacelab中文网 |网站地图
GMT+8, 2024-11-22 03:44 , Processed in 0.160831 second(s), 48 queries .
Powered by Discuz! X3.4
Copyright © 2001-2020, Tencent Cloud.