I just installed and started using DeepFaceLab_DirectX12_build_05_04_2022 yesterday and things went smoothly until about 8 hours into the training. I got this message and I have no idea what it means or how to solve it. Nothing happens when I press a key to continue. I found a thread on another forum where a guy wrote that it can be fixed by doing a tensorflow update (in the project). He wrote: Simply go to the folder “DeepFake\ _internal\python-3.6.8”, open the terminal, and enter “python -m pip install tensorflow”. I assume that the “terminal” means to open the python application and type in “python -m pip install tensorflow” and press enter. I tried doing this but nothing happens or I am doing something wrong. I am obviously a beginner at both DFL, Python, and coding so would really appreciate some help from the pros. This is the message:
2022-09-14 00:41:00.440554: F tensorflow/core/common_runtime/dml/dml_upload_heap.cc:56] HRESULT failed with 0x887a0005: chunk->resource->Map(0, nullptr, &upload_heap_data)
The terminal in this case is opened by running python_console.bat inside /_internal.
I don’t know about that specific error. I assume your card supports Directx12 and that the other builds will not work for your card.