16GB is not enough?

[wpl427]

torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 22.00 MiB (GPU 0; 14.61 GiB total capacity; 13.30 GiB already allocated; 9.19 MiB free; 13.78 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

What is the minimum video memory??

[wpl427]

Or how to reduce video memory usage by modifying the configuration?

[zideliu]

you can try this open_clip.create_model_and_transforms('ViT-bigG-14', 'laion2b_s39b_b160k',precision='fp16') in train_t2i_custom_v2.py

[wpl427]

before: prompt_model,, = open_clip.create_model_and_transforms('ViT-bigG-14', 'laion2b_s39b_b160k') after: prompt_model,, = open_clip.create_model_and_transforms('ViT-bigG-14', 'laion2b_s39b_b160k',precision='fp16') error: (faceswap) [root@prod-emr-gpu01 StyleDrop-PyTorch]# accelerate launch --num_processes 8 --mixed_precision fp16 train_t2i_custom_v2.py --config=configs/custom.py 2023-08-21 13:16:44.463755: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations. To enable the following instructions: AVX2 AVX512F FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags. 2023-08-21 13:16:45.329695: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT The following values were not passed to accelerate launch and had defaults used instead: --num_machines was set to a value of 1 --num_cpu_threads_per_process was set to 8 to improve out-of-box performance To avoid this warning pass in values for each of the problematic parameters or run accelerate config. 2023-08-21 13:16:49.788896: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT 2023-08-21 13:16:52.551 | INFO | main:train:63 - Process 0 using device: cuda I0821 13:16:52.551910 140480274581312 factory.py:158] Loaded ViT-bigG-14 model config. 2023-08-21 13:16:52.578 | DEBUG | open_clip.transformer:init:314 - xattn in transformer of CLIP is True 2023-08-21 13:17:09.847 | DEBUG | open_clip.transformer:init:314 - xattn in transformer of CLIP is True I0821 13:17:20.080452 140480274581312 factory.py:206] Loading pretrained ViT-bigG-14 weights (laion2b_s39b_b160k). Traceback (most recent call last): File "/data/miniconda3/envs/faceswap/bin/accelerate", line 8, in sys.exit(main()) File "/data/miniconda3/envs/faceswap/lib/python3.9/site-packages/accelerate/commands/accelerate_cli.py", line 43, in main args.func(args) File "/data/miniconda3/envs/faceswap/lib/python3.9/site-packages/accelerate/commands/launch.py", line 837, in launch_command simple_launcher(args) File "/data/miniconda3/envs/faceswap/lib/python3.9/site-packages/accelerate/commands/launch.py", line 354, in simple_launcher raise subprocess.CalledProcessError(returncode=process.returncode, cmd=cmd) subprocess.CalledProcessError: Command '['/data/miniconda3/envs/faceswap/bin/python', 'train_t2i_custom_v2.py', '--config=configs/custom.py']' died with <Signals.SIGKILL: 9>.