| On this page | |
| Since | 21.0 |
Overview ¶
This node is a wrapper around the OIDN training script used to train an OIDN
denoising filter model on preprocessed training and validation datasets. Generally,
you should use this node with 
Preprocess OIDN.
Parameters ¶
Training ¶
Input Directory
The input directory that contains the preprocessed dataset.
Training Dataset
Name of the training dataset folder.
Validation Dataset
Name of the validation dataset folder.
Output Directory
The directory to output training results.
Output Name
The desired name for the training result.
Epochs
The number of training epochs.
Validation Rate
The interval the model is evaluated with the validation dataset.
Checkpoint Rate
The interval the model checkpoints are created to save training progress.
Loss Function
The loss function to use.
MS-SSIM Weights
When Loss Function is MS-SSIM or L1 MS-SSIM, the MS-SSIM scale weights.
Learning Rate
When on, specifies a custom initial learning rate. By default, the
initial learning rate used in the one cycle learning rate scheduler is set
to 25. Otherwise, an initial learning rate of Max Learning Rate/Initial Learning Rate
is used.
Max Learning Rate
Wwhen on, specifies a custom maximum learning rate. By default, the max
learning rate is set to 3.125e-6 * Batch Size.
Learning Rate Warmup
The percentage of the learning rate schedule cycle spent increasing the learning rate (warm-up).
Precision
The precision to perform training with. By default, training is performed with mixed precision (FP16 and FP32), allowing for faster training and less memory usage.
Tile Size
The size of the cropped image tiles.
Batch Size
The total batch size of all devices. The batch size should be divisible by the number of devices.
Worker Threads
The number of data loader worker threads per device.
Seed
When on, specifies the initial seed value for the PyTorch random number generator.
Filter
Filter to use when training.
RT
Generic ray tracing denoising filter. Suitable for denoising images rendered with Monte Carlo ray tracing methods like unidirectional and bidirectional path tracing.
RTLightmap
Variant of RT filter optimized for denoising HDR and normalized directional lightmaps and does not support LDR images.
The choice of filter determines the HDR/LDR transfer function used.
Input Features
Set of input features of the dataset to preprocess for training.
hdr
color (high dynamic range) with file extension .hdr.exr
ldr
color (low dynamic range) with file extension .ldr.exr
sh1
color (normalized L1 spherical harmonics) with file extensions
.sh1x.exr, .sh1y.exr, and .sh1z.exr
alb
albedo with file extension .alb.exr
nrm
shading normal with file extension .nrm.exr
All input features are assumed to be noisy, including the auxilliary features (e.g. albedo, normal). The auxiliary feature images are optional inputs which usually improve denoising quality and preserve more details.
Deterministic
Make computations deterministic, resulting in slower performance.
Export Model File
When on, the trained model is exported to a .tza or
.zip file when training finishes.
Export Directory
The directory that contains exported model files.
Export File Name
The file name to export the model. The file name must end in either .tza
or zip. The extension will determine how the model is exported.
Export Epoch
When on, overrides the epoch the model is exported. When off, the model from the final training epoch is exported.
Environment ¶
Device ID
Specified device to use if there are multiple devices of the same kind available (e.g. multiple GPUs).
Device Count
Number of devices to use for data-parallel execution for faster performance.
Environment Path
The path to the virtual environment on disk.
Python Bin
Determines which Python executable to use when creating the virtual
environment and installing packages. Also the version of Python
that will be associated with the venv.
Custom Python Bin
When Python Bin is set to Custom, determines the path
to the Python interpreter to use when creating the virtual environment.
Use Symlinks
When on, the virtual environment is created with symlinks to the source Python binaries if possible. This is recommened when creating a virtual environment using Houdini’s Python interpreter.
Use Pip Cache
When enabled, pip will attempt to use cached packages on the local system instead of downloading them every time. This can improve the installation times when repeatedly installing the same Python package in different virtual environments.
Schedulers ¶
TOP Scheduler Override
This parameter overrides the TOP scheduler for this node.
Schedule When
When enabled, this parameter can be used to specify an expression that determines which work items from the node should be scheduled. If the expression returns zero for a given work item, that work item will immediately be marked as cooked instead of being queued with a scheduler. If the expression returns a non-zero value, the work item is scheduled normally.
Work Item Label
Determines how the node should label its work items. This parameter allows you to assign non-unique label strings to your work items which are then used to identify the work items in the attribute panel, task bar, and scheduler job names.
Use Default Label
The work items in this node will use the default label from the TOP network, or have no label if the default is unset.
Inherit From Upstream Item
The work items inherit their labels from their parent work items.
Custom Expression
The work item label is set to the Label Expression custom expression which is evaluated for each item.
Node Defines Label
The work item label is defined in the node’s internal logic.
Label Expression
When on, this parameter specifies a custom label for work items created by this node. The parameter can be an expression that includes references to work item attributes or built-in properties. For example, $OS: @pdg_frame will set the label of each work item based on its frame value.
Work Item Priority
This parameter determines how the current scheduler prioritizes the work items in this node.
Inherit From Upstream Item
The work items inherit their priority from their parent items. If a work item has no parent, its priority is set to 0.
Custom Expression
The work item priority is set to the value of Priority Expression.
Node Defines Priority
The work item priority is set based on the node’s own internal priority calculations.
This option is only available on the 
Python Processor TOP, 
ROP Fetch TOP, and ROP Output TOP nodes. These nodes define their own prioritization schemes that are implemented in their node logic.
Priority Expression
This parameter specifies an expression for work item priority. The expression is evaluated for each work item in the node.
This parameter is only available when Work Item Priority is set to Custom Expression.
| See also |
