这个配置 为什么还是跑崩 哪里设置不对吗

186

星级打分

• 1
• 2
• 3
• 4
• 5

平均分:NAN  参与人数:0  我的评分:未评

Error:Traceback (most recent call last):  File "D:\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\mainscripts\Trainer.py", line 173, in trainerThread    model_save()  File "D:\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\mainscripts\Trainer.py", line 68, in model_save    model.save()  File "D:\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\models\ModelBase.py", line 393, in save    self.onSave()  File "D:\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\models\Model_SAEHD\Model.py", line 759, in onSave    model.save_weights ( self.get_strpath_storage_for_file(filename) )  File "D:\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\core\leras\layers\Saveable.py", line 61, in save_weights    d_dumped = pickle.dumps (d, 4)MemoryError

ken2099

186 发表于 2023-7-3 02:19
Error:Traceback (most recent call last):  File "D:\DeepFaceLab_NVIDIA_RTX3000_series\_internal\Dee ...

C盘的系统托管虚拟内存关掉

186

有没大佬遇到过 解决过 跑一下就崩了

seancai110

gpu快跑满了，显存不够很正常啊。你ae_dim是不是设太大了，512，好夸张

186

seancai110 发表于 2023-7-3 01:48
gpu快跑满了，显存不够很正常啊。你ae_dim是不是设太大了，512，好夸张

我是用的滚石大大的迪丽热巴万能模型跑的 都跑了100万次了 今天手贱重新设置了一下虚拟内存 就开始崩了 然后我把虚拟内存改到512G  一样崩

186

davio 发表于 2023-7-3 02:15
报错信息发一下

Error:Traceback (most recent call last):  File "D:\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\mainscripts\Trainer.py", line 173, in trainerThread    model_save()  File "D:\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\mainscripts\Trainer.py", line 68, in model_save    model.save()  File "D:\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\models\ModelBase.py", line 393, in save    self.onSave()  File "D:\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\models\Model_SAEHD\Model.py", line 759, in onSave    model.save_weights ( self.get_strpath_storage_for_file(filename) )  File "D:\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\core\leras\layers\Saveable.py", line 61, in save_weights    d_dumped = pickle.dumps (d, 4)MemoryError

lknet

对,c的托管去掉

denvor

不过这个模型的 ae_dims 是有点夸张。 512啊。