CUDA out of memory. Tried to allocate 32.00 MiB (GPU 0; 3.00 GiB total capacity; 1.83 GiB already allocated; 27.55 MiB free; 1.94 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 Fun fact: In the olden times, PyTorch would print “Buy more RAM” along with the error message, but then things got all serious. What is strange, is that the EXACT same code ran fine the first time. When I tried to run the same code with slightly different hyperparams (doesn't affect the model, things like early-stop patience) it breaks during the first few batches of the first epoch. Even when I try to run the same hyperparams as my first experiment, it breaks

try SET COMMANDLINE_ARGS= --lowvram --precision full --no-half or EXPORT COMMANDLINE_ARGS= --lowvram --precision full --no-halfIn file txt2img.py, function load_model_from_config(...), change from: model.cuda() to model.cuda().half() File "/home/emarquer/miniconda3/envs/pytorch/lib/python3.6/runpy.py", line 193, in _run_module_as_main

Make sure to add model.to(torch.float16) in load_model_from_config function, just before model.cuda() is called. CUDA out of memory. Tried to allocate 32.00 MiB (GPU 0; 3.00 GiB total capacity; 1.87 GiB already allocated; 13.55 MiB free; 1.95 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 RuntimeError: CUDA out of memory. Tried to allocate 11.88 MiB (GPU 4; 15.75 GiB total capacity; 10.50 GiB already allocated; 1.88 MiB free; 3.03 GiB cached)File "/content/gdrive/My Drive/Colab Notebooks/STANet-withpth/models/CDFA_model.py", line 90, in forward

loading half-model sucessfull fix error at txt2img. But when I try do the same at img2img I got new error: The answer to the question how much kB in 8 MB is usually 8000 kB, but depending on the vendor of the RAM, hard disk, the software producer or the CPU manufacturer for example, MB could also mean 1024 * 1024 B = 1024 2 bytes. Even a mixed use 1000 * 1024 B cannot completely ruled out. Unless indicated differently, go with 8 MB equal 8000 kB. CUDA out of memory. Tried to allocate 232.00 MiB (GPU 0; 3.00 GiB total capacity; 1.61 GiB already allocated; 119.55 MiB free; 1.85 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 Megabyte per second is a unit of data transfer rate which is equal to 8 × 106 bit/s, or 106 bytes per second. The symbol for Megabyte per second are MB/s, and MBps.RuntimeError: CUDA out of memory. Tried to allocate 3.12 GiB (GPU 0; 24.00 GiB total capacity; 2.06 GiB already allocated; 19.66 GiB free; 2.31 GiB reserved in total by PyTorch)”