# The newly trained model is output into this file.Įxample #2 - Bank Lobby: view comparison in 24-bit HD, original photo CC-BY-SA 2. generator-start=5 -discriminator-start=0 -adversarial-start=5 \ perceptual-layer=conv5_2 -smoothness-weight=2e4 -adversary-weight=1e3 \ Python3.4 enhance.py -train "data/*.jpg " -model custom -scales=2 -epochs=250 \ # Train the model using an adversarial setup based on below. generator-blocks=4 -generator-filters=64
Python3.4 enhance.py -train "data/*.jpg " -model custom -scales=2 -epochs=50 \ # Pre-train the model using perceptual loss from paper below. # Remove the model file as don't want to reload the data to fine-tune it. 1.a) Enhancing ImagesĪ list of example command lines you can use with the pre-trained models provided in the GitHub releases: On the CPU, you can also set environment variable to OMP_NUM_THREADS=4, which is most useful when running the script multiple times in parallel. The default is to use -device=cpu, if you have NVIDIA card setup with CUDA already try -device=gpu0. Runtime depends on the neural network size.
CPU Rendering HQ - This will take roughly 20 to 60 seconds for 1080p output, however on most machines you can run 4-8 processes simultaneously given enough system RAM. GPU Rendering HQ - Assuming you have CUDA setup and enough on-board RAM to fit the image and neural network, generating 1080p output should complete in 5 seconds, or 2s per image if multiple at the same time. For the samples above, here are the performance results: The -device argument that lets you specify which GPU or CPU to use. The main script is called enhance.py, which you can run with Python 3.4+ once it's setup as below. That's only possible in Hollywood - but using deep learning as "Creative AI" works and it is just as cool! Here's how you can get started. It's not reconstructing your photo exactly as it would have been if it was HD. The catch? The neural network is hallucinating details based on its training from example images. You'll get even better results by increasing the number of neurons or training with a dataset similar to your low resolution image. Example #1 - Old Station: view comparison in 24-bit HD, original photo CC-BY-SA seen on TV! What if you could increase the resolution of your photos using technology from CSI laboratories? Thanks to deep learning and #NeuralEnhance, it's now possible to train a neural network to zoom in to your images at 2x or even 4x.
0 kommentar(er)
