DeepFace 常见错误问题汇总（不断更新）

DeepFace 常见报OOM 错误的解决方法
https://tieba.baidu.com/p/7805083557

请问dfl能在云gpu服务器上玩吗，想租个腾讯云gpu来炼丹，自己搞感觉太费电了

为什么训练越来越差呀，AMD集成显卡。

第一步就是系统找不到指定路径

大佬，我需不需要将练出来的视频截一些图放进人像里再继续练，毕竟练出来的图有好多新表情动作之类的，还是说保存模型里已经包含了这些

大佬们，问一下，合成的时候进度一直是百分之0是怎么回事啊？

Running DeepFaceLive.
Exception in thread Thread-2:
Traceback (most recent call last):
File "threading.py", line 932, in _bootstrap_inner
File "threading.py", line 870, in run
File "D:\DeepFaceLive_NVIDIA\_internal\DeepFaceLive\xlib\mp\csw\CSWBase.py", line 316, in <lambda>
threading.Thread(target=lambda: self._process.start(), daemon=True).start()
File "multiprocessing\process.py", line 121, in start
File "multiprocessing\context.py", line 224, in _Popen
File "multiprocessing\context.py", line 326, in _Popen
ImportError: cannot import name 'Popen' from 'multiprocessing.popen_spawn_win32' (D:\DeepFaceLive_NVIDIA\_internal\python\python38.zip\multiprocessing\popen_spawn_win32.pyc)
FaceMerger error: exception: access violation writing 0x0000000000000188 Traceback (most recent call last):
File "D:\DeepFaceLive_NVIDIA\_internal\DeepFaceLive\xlib\mp\csw\CSWBase.py", line 472, in _start_proc
self.on_start(*worker_start_args, **worker_start_kwargs)
File "D:\DeepFaceLive_NVIDIA\_internal\DeepFaceLive\apps\DeepFaceLive\backend\FaceMerger.py", line 61, in on_start
cs.device.set_choices( ['CPU'] + lib_cl.get_available_devices_info(), none_choice_name='@misc.menu_select')
File "D:\DeepFaceLive_NVIDIA\_internal\DeepFaceLive\xlib\avecl\_internal\backend\Device.py", line 462, in get_available_devices_info
for device_index, device_id in enumerate(_get_opencl_device_ids()):
File "D:\DeepFaceLive_NVIDIA\_internal\DeepFaceLive\xlib\avecl\_internal\backend\Device.py", line 421, in _get_opencl_device_ids
if CL.clGetPlatformIDs(0, None, num_platforms) != CL.CLERROR.SUCCESS or \
File "D:\DeepFaceLive_NVIDIA\_internal\DeepFaceLive\xlib\avecl\_internal\backend\OpenCL\OpenCL.py", line 27, in wrapper
return dll_func(*args)
OSError: exception: access violation writing 0x0000000000000188
运行错误 那位大佬帮我看看