fast-style-transfer
fast-style-transfer copied to clipboard
larger image results in ResourceExhaustError
Hi
I understand it could also be likely tensorflow tweaking, but with larger image and using tensorflow-gpu, I get ResourceExhaustedError. Can you kindly check?
This is even without training, but just using checkpoints. I tried reducing batch-size to 1, and still same issue. I also tried soft_config.gpu_options.per_process_gpu_memory_fraction to 0.1
, but that does not seem to have any effect
Command:
python evaluate.py --checkpoint ./rain-princess.ckpt --in-path my_photo_1.jpg --out-path ./output_image.jpg
Image size: 5344 x 3006 (4.1 MB)
Error: output_error.txt
According to the output, it appears that your GPU has 2.00GiB
of vram, and unfortunately, for large images being converted to numpy arrays, you would need significantly more memory.
I'm really sorry, but running a large image on GPU might not be an option for you; if you have more RAM on your system, try running on CPU
How much vram the GPU should have to typically run tensorflow in my system smoothly in GPU mode?
In most cases, a 2-4gb card is enough (like running a 1920x1080 image), but an image with the longer side being larger than 5000 might need 5-6gb to even be able to allocate enough space for tensorflow.