mirror of
https://github.com/graphdeco-inria/gaussian-splatting
synced 2024-11-22 08:18:17 +00:00
Merge branch 'main' into develop
This commit is contained in:
commit
0166108c12
90
README.md
90
README.md
@ -97,67 +97,67 @@ python train.py -s <path to dataset>
|
||||
```
|
||||
|
||||
<details>
|
||||
<summary style="font-size: 16px;">Command Line Arguments for train.py</summary>
|
||||
<summary><span style="font-weight: bold;">Command Line Arguments for train.py</span></summary>
|
||||
|
||||
### --source_path / -s
|
||||
#### --source_path / -s
|
||||
Path to the source directory containing a COLMAP or Synthetic NeRF data set.
|
||||
### --model_path / -m
|
||||
#### --model_path / -m
|
||||
Path where the trained model should be stored (```output/<random>``` by default).
|
||||
### --images / -i
|
||||
#### --images / -i
|
||||
Alternative subdirectory for COLMAP images (```images``` by default).
|
||||
### --eval
|
||||
#### --eval
|
||||
Add this flag to use a MipNeRF360-style training/test split for evaluation.
|
||||
### --resolution / -r
|
||||
#### --resolution / -r
|
||||
Changes the resolution of the loaded images before training. If provided ```1, 2, 4``` or ```8```, uses original, 1/2, 1/4 or 1/8 resolution, respectively. For all other values, rescales the width to the given number while maintaining image aspect. ```1``` by default.
|
||||
### --white_background / -w
|
||||
#### --white_background / -w
|
||||
Add this flag to use white background instead of black (default), e.g., for evaluation of NeRF Synthetic dataset.
|
||||
### --sh_degree
|
||||
#### --sh_degree
|
||||
Order of spherical harmonics to be used (no larger than 3). ```3``` by default.
|
||||
### --convert_SHs_python
|
||||
#### --convert_SHs_python
|
||||
Flag to make pipeline compute forward and backward of SHs with PyTorch instead of ours.
|
||||
### --convert_cov3D_python
|
||||
#### --convert_cov3D_python
|
||||
Flag to make pipeline compute forward and backward of the 3D covariance with PyTorch instead of ours.
|
||||
### --iterations
|
||||
#### --iterations
|
||||
Number of total iterations to train for, ```30_000``` by default.
|
||||
### --feature_lr
|
||||
#### --feature_lr
|
||||
Spherical harmonics features learning rate, ```0.0025``` by default.
|
||||
### --opacity_lr
|
||||
#### --opacity_lr
|
||||
Opacity learning rate, ```0.05``` by default.
|
||||
### --scaling_lr
|
||||
#### --scaling_lr
|
||||
Scaling learning rate, ```0.001``` by default.
|
||||
### --rotation_lr
|
||||
#### --rotation_lr
|
||||
Rotation learning rate, ```0.001``` by default.
|
||||
### --position_lr_max_steps
|
||||
#### --position_lr_max_steps
|
||||
Number of steps (from 0) where position learning rate goes from ```initial``` to ```final```. ```30_000``` by default.
|
||||
### --position_lr_init
|
||||
#### --position_lr_init
|
||||
Initial 3D position learning rate, ```0.00016``` by default.
|
||||
### --position_lr_final
|
||||
#### --position_lr_final
|
||||
Final 3D position learning rate, ```0.0000016``` by default.
|
||||
### --position_lr_delay_mult
|
||||
#### --position_lr_delay_mult
|
||||
Position learning rate multiplier (cf. Plenoxels), ```0.01``` by default.
|
||||
### --densify_from_iter
|
||||
#### --densify_from_iter
|
||||
Iteration where densification starts, ```500``` by default.
|
||||
### --densify_until_iter
|
||||
#### --densify_until_iter
|
||||
Iteration where densification stops, ```15_000``` by default.
|
||||
### --densify_grad_threshold
|
||||
#### --densify_grad_threshold
|
||||
Limit that decides if points should be densified based on 2D position gradient, ```0.0002``` by default.
|
||||
### --densification_interal
|
||||
#### --densification_interal
|
||||
How frequently to densify, ```100``` (every 100 iterations) by default.
|
||||
### --opacity_reset_interval
|
||||
#### --opacity_reset_interval
|
||||
How frequently to reset opacity, ```3_000``` by default.
|
||||
### --lambda_dssim
|
||||
#### --lambda_dssim
|
||||
Influence of SSIM on total loss from 0 to 1, ```0.2``` by default.
|
||||
### --percent_dense
|
||||
#### --percent_dense
|
||||
Percentage of scene extent (0--1) a point must exceed to be forcibly densified, ```0.1``` by default.
|
||||
### --ip
|
||||
#### --ip
|
||||
IP to start GUI server on, ```127.0.0.1``` by default.
|
||||
### --port
|
||||
#### --port
|
||||
Port to use for GUI server, ```6009``` by default.
|
||||
### --test_iterations
|
||||
#### --test_iterations
|
||||
Space-separated iterations at which the training script computes L1 and PSNR over test set, ```7000 30000``` by default.
|
||||
### --save_iterations
|
||||
#### --save_iterations
|
||||
Space-separated iterations at which the training script saves the Gaussian model, ```7000 30000 <iterations>``` by default.
|
||||
### --quiet
|
||||
#### --quiet
|
||||
Flag to omit any text written to standard out pipe.
|
||||
|
||||
</details>
|
||||
@ -175,40 +175,40 @@ python metrics.py -m <path to trained model> # Compute error metrics on renderin
|
||||
```
|
||||
|
||||
<details>
|
||||
<summary style="font-size: 16px;">Command Line Arguments for render.py</summary>
|
||||
<summary><span style="font-weight: bold;">Command Line Arguments for render.py</span></summary>
|
||||
|
||||
### --model_path / -m
|
||||
#### --model_path / -m
|
||||
Path where the trained model should be stored (```output/<random>``` by default).
|
||||
### --skip_train
|
||||
#### --skip_train
|
||||
Flag to skip rendering the training set.
|
||||
### --skip_test
|
||||
#### --skip_test
|
||||
Flag to skip rendering the test set.
|
||||
### --quiet
|
||||
#### --quiet
|
||||
Flag to omit any text written to standard out pipe.
|
||||
|
||||
**The below parameters will be read automatically from the model path, based on what was used for training. However, you may override them by providing them explicitly on the command line.**
|
||||
|
||||
### --source_path / -s
|
||||
#### --source_path / -s
|
||||
Path to the source directory containing a COLMAP or Synthetic NeRF data set.
|
||||
### --images / -i
|
||||
#### --images / -i
|
||||
Alternative subdirectory for COLMAP images (```images``` by default).
|
||||
### --eval
|
||||
#### --eval
|
||||
Add this flag to use a MipNeRF360-style training/test split for evaluation.
|
||||
### --resolution / -r
|
||||
#### --resolution / -r
|
||||
Changes the resolution of the loaded images before training. If provided ```1, 2, 4``` or ```8```, uses original, 1/2, 1/4 or 1/8 resolution, respectively. For all other values, rescales the width to the given number while maintaining image aspect. ```1``` by default.
|
||||
### --white_background / -w
|
||||
#### --white_background / -w
|
||||
Add this flag to use white background instead of black (default), e.g., for evaluation of NeRF Synthetic dataset.
|
||||
### --convert_SHs_python
|
||||
#### --convert_SHs_python
|
||||
Flag to make pipeline render with computed SHs from PyTorch instead of ours.
|
||||
### --convert_cov3D_python
|
||||
#### --convert_cov3D_python
|
||||
Flag to make pipeline render with computed 3D covariance from PyTorch instead of ours.
|
||||
|
||||
</details>
|
||||
|
||||
<details>
|
||||
<summary style="font-size: 16px;">Command Line Arguments for metrics.py</summary>
|
||||
<summary><span style="font-weight: bold;">Command Line Arguments for metrics.py</span></summary>
|
||||
|
||||
### --model_paths / -m
|
||||
#### --model_paths / -m
|
||||
Space-separated list of model paths for which metrics should be computed.
|
||||
</details>
|
||||
<br>
|
||||
|
@ -1 +1 @@
|
||||
Subproject commit 3c35b80b65fdfca4ba580a6d7d11e9a8ede29f8a
|
||||
Subproject commit bd01fcfc13052b8d1afb7005716f4d76a7685d78
|
@ -1 +1 @@
|
||||
Subproject commit 4aedd8226f7257935891049f5a378b0e21d0aa37
|
||||
Subproject commit 3a07ac2e39b9ba7043ffc8bb98397c3ba6e2532d
|
@ -1 +1 @@
|
||||
Subproject commit 10b4e40c4e73330ceb1adcb3bae4a494e7d3b9c1
|
||||
Subproject commit 76eff30b1533718fe652799729203821b7f54d73
|
Loading…
Reference in New Issue
Block a user