CUDA_SETUP: WARNING! libcudart.so not found in any environmental path.

[duskfallcrew]

===================================BUG REPORT===================================
Welcome to bitsandbytes. For bug reports, please submit your error trace to: https://github.com/TimDettmers/bitsandbytes/issues
For effortless bug reporting copy-paste your error into this form: https://docs.google.com/forms/d/e/1FAIpQLScPB8emS3Thkp66nvqwmjTEgxp8Y9ufuWTzFyr9kJ5AoI47dQ/viewform?usp=sf_link

CUDA_SETUP: WARNING! libcudart.so not found in any environmental path. Searching /usr/local/cuda/lib64...
CUDA SETUP: CUDA runtime path found: /usr/local/cuda/lib64/libcudart.so
CUDA SETUP: Highest compute capability among GPUs detected: 7.5
CUDA SETUP: Detected CUDA version 118
CUDA SETUP: Loading binary /usr/local/lib/python3.8/dist-packages/bitsandbytes/libbitsandbytes_cuda118.so...
use 8-bit AdamW optimizer | {}
override steps. steps for 20 epochs is / 指定エポックまでのステップ数: 2000
running training / 学習開始
num train images * repeats / 学習画像の数×繰り返し回数: 400
num reg images / 正則化画像の数: 0
num batches per epoch / 1epochのバッチ数: 100
num epochs / epoch数: 20
batch size per device / バッチサイズ: 4
total train batch size (with parallel & distributed & accumulation) / 総バッチサイズ（並列学習、勾配合計含む）: 4
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: 2000

Weirdly after this it let me still train it?

[FinaBro69]

go to (Training parameters) you can see is (AdamW8bit change to AdamW)
and (Gradient checkpointing, Memory efficient attention) is two also tick it

Hope it helps you

[saikaboty]

go to (Training parameters) you can see is (AdamW8bit change to AdamW) and (Gradient checkpointing, Memory efficient attention) is two also tick it

Hope it helps you

Thanks a lot, that helped ... I though it was the CUDA version or something like this ...
And changing it to AdamW still benefit from the acceleration of having RTX 3000 or 4000 series ?

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.