Skip to content

Latest commit

 

History

History
162 lines (116 loc) · 8.63 KB

README.md

File metadata and controls

162 lines (116 loc) · 8.63 KB
Raw

MINER_pl

Unofficial implementation of MINER: Multiscale Implicit Neural Representations in pytorch-lightning.

Official implementation : https://github.com/vishwa91/miner

image

📖 Ref readings

⚠️ Main differences w.r.t. the original paper before continue:

  • In the pseudo code on page 8, where the author states Weight sharing for images, it means finer level networks are initialized with coarser level network weights. However, I did not find the correct way to implement this. Therefore, I initialize the network weights from scratch for all levels.
  • The paper says it uses sinusoidal activation (does he mean SIREN? I don't know), but I use gaussian activation (in hidden layers) with trainable parameters (per block) like my experiments in the other repo. In finer levels where the model predicts laplacian pyramids, I use sinusoidal activation x |-> sin(ax) with trainable parameters a (per block) as output layer (btw, this performs significantly better than simple tanh). Moreover, I precompute the maximum amplitude for laplacian residuals, and use it to scale the output, and I find it to be better than without scaling.
  • I experimented with a common trick in coordinate mlp: positional encoding and find that using it can increase training time/accuracy with the same number of parameters (by reducing 1 layer). This can be turned on/off by specifying the argument --use_pe. The optimal number of frequencies depends on the patch size, the larger patch sizes, the more number of frequencies you need and vice versa.
  • Some difference in the hyperparameters: the default learning rate is 3e-2 instead of 5e-4. Optimizer is RAdam instead of Adam. Block pruning happens when the loss is lower than 1e-4 (i.e. when PSNR>=40) for image and 5e-3 for occupancy rather than 2e-7.

💻 Installation

  • Run pip install -r requirements.txt.
  • Download the images from Acknowledgement or prepare your own images into a folder called images.
  • Download the meshes from Acknowledgement or prepare your own meshes into a folder called meshes.

🔑 Training

image

Pluto example:

python train.py \
    --task image --path images/pluto.png \
    --input_size 4096 4096 --patch_size 32 32 --batch_size 256 --n_scales 4 \
    --use_pe --n_layers 3 \
    --num_epochs 50 50 50 200 \
    --exp_name pluto4k_4scale

Tokyo station example:

python train.py \
    --task image --path images/tokyo-station.jpg \
    --input_size 6000 4000 --patch_size 25 25 --batch_size 192 --n_scales 5 \
    --use_pe --n_layers 3 \
    --num_epochs 50 50 50 50 150 \
    --exp_name tokyo6k_5scale
Image (size) Train time (s) GPU mem (MiB) #Params (M) PSNR
Pluto (4096x4096) 53 3171 9.16 42.14
Pluto (8192x8192) 106 6099 28.05 45.09
Tokyo station (6000x4000) 68 6819 35.4 42.48
Shibuya (7168x2560) 101 8967 17.73 37.78
Shibuya (14336x5120) 372 8847 75.42 39.32
Shibuya (28672x10240) 890 10255 277.37 41.93
Shibuya (28672x10240)* 1244 6277 98.7 37.59

*paper settings (6 scales, each network has 4 layer with 9 hidden units)

The original image will be resized to img_wh for reconstruction. You need to make sure img_wh divided by 2^(n_scales-1) (the resolution at the coarsest level) is still a multiple of patch_wh.

mesh

First, convert the mesh to N^3 occupancy grid by

python preprocess_mesh.py --N 512 --M 1 --T 1 --path <path/to/mesh>

This will create N^3 occupancy to be regressed by the neural network. For detailed options, please see preprocess_mesh.py. Typically, increase M or T if you find the resulting occupancy bad.

Next, start training (bunny example):

python train.py \
    --task mesh --path occupancy/bunny_512.npy \
    --input_size 512 --patch_size 16 --batch_size 512 --n_scales 4 \
    --use_pe --n_freq 5 --n_layers 2 --n_hidden 8 \
    --loss_thr 5e-3 --b_chunks 512 \
    --num_epochs 50 50 50 150 \
    --exp_name bunny512_4scale

For full options, please see here. Some important options:

  • If your GPU memory is not enough, try reducing batch_size.
  • By default it will not log intermediate images to tensorboard to save time. To visualize image reconstruction and active blocks, add --log_image argument.

You are recommended to monitor the training progress by

tensorboard --logdir logs

where you can see training curves and images.

🟥🟩🟦 Block decomposition

To reconstruct the image using trained model and to visualize block decomposition per scale like Fig. 4 in the paper, see image_test.ipynb or mesh_test.ipynb

Examples:

💡 Implementation tricks

  • Setting num_workers=0 in dataloader increased the speed a lot.
  • As suggested in training details on page 4, I implement parallel block inference by defining parameters of shape (n_blocks, n_in, n_out) and use @ operator (same as torch.bmm) for faster inference.
  • To perform block pruning efficiently, I create two copies of the same network, and continually train and prune one of them while copying the trained parameters to the target network (somehow like in reinforcement learning, e.g. DDPG). This allows the network as well as the optimizer to shrink, therefore achieve higher memory and speed performance.
  • In validation, I perform inference in chunks like NeRF, and pass each chunk to cpu to reduce GPU memory usage.

💝 Acknowledgement

❓ Further readings

During a stream, my audience suggested me to test on this image with random pixels:

random

The default 32x32 patch size doesn't work well, since the texture varies too quickly inside a patch. Decreasing to 16x16 and increasing network hidden units make the network converge right away to 43.91 dB under a minute. Surprisingly, with the other image reconstruction SOTA instant-ngp, the network is stuck at 17 dB no matter how long I train.

ngp-random

Is this a possible weakness of instant-ngp? What effect could it bring to real application? You are welcome to test other methods to reconstruct this image!