with version v9 I am getting 9.47 s/It (on a rtx 3090) I get some warnings like WARNING: Failed to find Windows SDK. is this normal? enable full bf16 training.
running training / 学習開始
num examples / サンプル数: 50
num batches per epoch / 1epochのバッチ数: 50
num epochs / epoch数: 200
batch size per device / バッチサイズ: 1
gradient accumulation steps / 勾配を合計するステップ数 = 1
total optimization steps / 学習ステップ数: 10000
steps: 0%| | 0/10000 [00:00<?, ?it/s]
epoch 1/200
2024-10-25 19:28:16 INFO epoch is incremented. current_epoch: 0, epoch: 1 train_util.py:715
D:\IA\kohya\kohya_ss\venv\lib\site-packages\torch\autograd\graph.py:825: UserWarning: cuDNN SDPA backward got grad_output.strides() != output.strides(), attempting to materialize a grad_output with matching strides... (Triggered internally at C:\actions-runner_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\cudnn\MHA.cpp:676.)
return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass
steps: 0%|▎ | 50/10000 [07:53<26:09:50, 9.47s/it, avr_loss=0.399]
epoch 2/200
