https://github.com/kwea123/ngp_pl

Instant-ngp in pytorch+cuda trained with pytorch-lightning (high quality with high speed, with only few lines of legible code)

ngp_pl

Advertisement: stay tuned with my channel, I will upload cuda tutorials recently, and do a stream about this implementation!

Instant-ngp (only NeRF) in pytorch+cuda trained with pytorch-lightning (high quality with high speed). This repo aims at providing a concise pytorch interface to facilitate future research, and am grateful if you can share it (and a citation is highly appreciated)!

progress.mp4

lego_trainval_5min_PNSR35.73.mp4

• Official CUDA implementation
• torch-ngp another pytorch implementation that I highly referenced.

💻 Installation

This implementation has strict requirements due to dependencies on other libraries, if you encounter installation problem due to hardware/software mismatch, I'm afraid there is no intention to support different platforms (you are welcomed to contribute).

Hardware

• OS: Ubuntu 20.04
• NVIDIA GPU with Compute Compatibility >= 75 and memory > 6GB (Tested with RTX 2080 Ti), CUDA 11.3 (might work with older version)

Software

• Clone this repo by git clone https://github.com/kwea123/ngp_pl

• Python>=3.8 (installation via anaconda is recommended, use conda create -n ngp_pl python=3.8 to create a conda environment and activate it by conda activate ngp_pl)

• Python libraries

  • Install pytorch by pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cu113
  • Install tinycudann following their instruction (compilation and pytorch extension)
  • Install apex following their instruction
  • Install core requirements by pip install -r requirements.txt

• Cuda extension: Upgrade pip to >= 22.1 and run pip install models/csrc/

📚 Data preparation

Download preprocessed datasets from NSVF.

🔑 Training

Quickstart: python train.py --root_dir <path/to/lego> --exp_name Lego

It will train the lego scene for 20k steps (each step with 8192 rays), and perform one testing at the end. The training process should finish within about 5 minutes (saving testing image is slow, add --no_save_test to disable). Testing PSNR will be shown at the end.

If your GPU has larger memory, you can try increasing batch_size (and lr) and reducing num_epochs (e.g. --batch_size 16384 --lr 1.2e-2 --num_epochs 10). In my experiments, this further reduces the training time by 10~25s while maintaining the same quality.

More options can be found in opt.py.

🔎 Testing

Use test.ipynb to generate images. Lego pretrained model is available here

Comparison with torch-ngp and the paper

I compared the quality (average testing PSNR on Synthetic-NeRF) and the inference speed (on Lego scene) v.s. the concurrent work torch-ngp (default settings) and the paper, all trained for about 5 minutes:

Method	avg PSNR	FPS
torch-ngp	31.46	18.2
mine	32.38	36.2
instant-ngp paper	33.18	60

As for quality, mine is slightly better than torch-ngp, but the result might fluctuate across different runs.

As for speed, mine is faster than torch-ngp, but is still only half fast as instant-ngp. Speed is dependent on the scene (if most of the scene is empty, speed will be faster).

^{Left: torch-ngp. Right: mine.}

More details are in the following section.

Benchmarks

To run benchmarks, use the scripts under benchmarking.

Followings are my results (qualitative results here):

Synthetic-NeRF

	Mic	Ficus	Chair	Hotdog	Materials	Drums	Ship	Lego	AVG
PSNR	35.23	33.64	34.78	36.76	28.77	25.61	29.57	34.69	32.38
FPS	40.81	34.02	49.80	25.06	20.08	37.77	15.77	36.20	32.44

Synthetic-NSVF

	Wineholder	Steamtrain	Toad	Robot	Bike	Palace	Spaceship	Lifestyle	AVG
PSNR	31.06	35.65	34.49	36.23	36.99	36.36	35.48	33.96	35.03
FPS	47.07	75.17	50.42	64.87	66.88	28.62	35.55	22.84	48.93

Tanks and Temples

	Ignatius	Truck	Barn	Caterpillar	Family	AVG
PSNR	28.90	28.21	28.92	26.30	33.77	29.22
*FPS	10.04	7.99	16.14	10.91	6.16	10.25

*Evaluated on test-traj

BlendedMVS

	*Jade	*Fountain	Character	Statues	AVG
PSNR	25.69	26.91	30.16	26.93	27.42
**FPS	26.02	21.24	35.99	19.22	25.61

*I manually switch the background from black to white, so the number isn't directly comparable to that in the papers.

**Evaluated on test-traj

TODO

• support custom dataset