DeepSDFStruct.deep_sdf.workspace

Experiment Workspace Management

This module provides utilities for managing DeepSDF experiment workspaces, including directory structures, file naming conventions, and model loading/saving.

Constants

The module defines standard subdirectory and file names for organizing experiment artifacts: - Model parameters and checkpoints - Optimizer states - Latent code vectors - Training logs and plots - Reconstructions and evaluations - Dataset samples and normalization parameters

Architecture Registry

ARCHITECTURES: dict

Maps architecture names to decoder classes, enabling dynamic model instantiation from configuration files.

Functions

load_experiment_specifications

Load experiment configuration from specs.json file.

load_trained_model

Load a trained decoder network from checkpoint.

load_latent_vectors

Load learned latent codes from checkpoint.

create_experiment_directory

Initialize directory structure for a new experiment.

The workspace utilities ensure consistent organization across experiments and simplify model loading for inference and continued training.

Functions

get_data_source_map_filename(data_dir)
get_default_device()
get_evaluation_dir(experiment_dir, checkpoint)
get_latent_codes_dir(experiment_dir[, ...])
get_model_params_dir(experiment_dir[, ...])
get_normalization_params_filename(data_dir, ...)
get_optimizer_params_dir(experiment_dir[, ...])
get_reconstructed_code_filename(...)
get_reconstructed_mesh_filename(...[, ...])
get_screenshots_dir(experiment_dir[, ...])
init_decoder(experiment_specs, device, ...)
load_experiment_specifications(...)
load_latent_vectors(experiment_directory, ...)
load_model_parameters(experiment_directory, ...)
load_optimizer(experiment_directory, ...)
load_trained_model(experiment_directory, ...)
print_model_specifications(experiment_directory)
save_experiment_summary(...)

Classes

ExperimentSummary

class DeepSDFStruct.deep_sdf.workspace.ExperimentSummary

Bases: TypedDict

data_dir: str

device: str

host_name: str

loss: float

num_epochs: int

timestamp: str

training_duration: str

version: str

DeepSDFStruct.deep_sdf.workspace.get_data_source_map_filename(data_dir)

DeepSDFStruct.deep_sdf.workspace.get_default_device()

DeepSDFStruct.deep_sdf.workspace.get_evaluation_dir(experiment_dir, checkpoint, create_if_nonexistent=False)

DeepSDFStruct.deep_sdf.workspace.get_latent_codes_dir(experiment_dir, create_if_nonexistent=False)

DeepSDFStruct.deep_sdf.workspace.get_model_params_dir(experiment_dir, create_if_nonexistent=False)

DeepSDFStruct.deep_sdf.workspace.get_normalization_params_filename(data_dir, dataset_name, class_name, instance_name)

DeepSDFStruct.deep_sdf.workspace.get_optimizer_params_dir(experiment_dir, create_if_nonexistent=False)

DeepSDFStruct.deep_sdf.workspace.get_reconstructed_code_filename(experiment_dir, epoch, dataset, class_name, instance_name)

DeepSDFStruct.deep_sdf.workspace.get_reconstructed_mesh_filename(experiment_dir, epoch, dataset, class_name, instance_name, create_dir=True, filetype='ply')

DeepSDFStruct.deep_sdf.workspace.get_screenshots_dir(experiment_dir, create_if_nonexistent=True)

DeepSDFStruct.deep_sdf.workspace.init_decoder(experiment_specs, device, data_parallel)

DeepSDFStruct.deep_sdf.workspace.load_experiment_specifications(experiment_directory)

DeepSDFStruct.deep_sdf.workspace.load_latent_vectors(experiment_directory, checkpoint, device)

DeepSDFStruct.deep_sdf.workspace.load_model_parameters(experiment_directory, checkpoint, decoder, device)

Parameters:

decoder (torch.nn.modules.module.Module)

DeepSDFStruct.deep_sdf.workspace.load_optimizer(experiment_directory, checkpoint, optimizer, device)

Parameters:

optimizer (torch.nn.modules.module.Module)

DeepSDFStruct.deep_sdf.workspace.load_trained_model(experiment_directory, checkpoint, device=None, data_parallel=False)

Parameters:

• experiment_directory (str)

• checkpoint (str)

DeepSDFStruct.deep_sdf.workspace.print_model_specifications(experiment_directory)

Parameters:

experiment_directory (str)

DeepSDFStruct.deep_sdf.workspace.save_experiment_summary(experiment_directory, summary)

Parameters:

• experiment_directory (str)

• summary (DeepSDFStruct.deep_sdf.workspace.ExperimentSummary)