Implementation of Dreambooth (https://arxiv.org/abs/2208.12242) by way of Textual Inversion (https://arxiv.org/abs/2208.01618) for Stable Diffusion (https://arxiv.org/abs/2112.10752). Tweaks focused on training faces, objects, and styles.
Under ubuntu 22.04, cudatoolkit 11.3 and 11.7 were throwing errors.
cudatoolkit=11.3 :
"ImportError: /home/user/anaconda3/envs/db-joepenna2/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so:
undefined symbol: cudaGraphInstantiateWithFlags, version libcudart.so.11.0"
cudatoolkit=11.7:
"torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 23.65 GiB total capacity; 22.44 GiB already allocated; 10.75 MiB free; 22.77 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF"
No problems with 11.8. Also tested under Windows 11.
-Added conda-forge channel (for cudatoolkit 11.8 availability) -Changed cudatoolkit=11.8
Under ubuntu 22.04, cudatoolkit 11.3 and 11.7 were throwing errors.
cudatoolkit=11.3: "ImportError: /home/user/anaconda3/envs/db-joepenna2/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so: undefined symbol: cudaGraphInstantiateWithFlags, version libcudart.so.11.0"cudatoolkit=11.7: "torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 23.65 GiB total capacity; 22.44 GiB already allocated; 10.75 MiB free; 22.77 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF"No problems with 11.8. Also tested under Windows 11.
Updated
Open In Colablink.