credit.models

Submodules

• credit.models.base_model
• credit.models.camulator
• credit.models.checkpoint
• credit.models.crossformer
• credit.models.debugger_model
• credit.models.dscale_wrf
• credit.models.fuxi
• credit.models.graph
• credit.models.reset
• credit.models.swin
• credit.models.swin_wrf
• credit.models.unet
• credit.models.unet_attention_modules
• credit.models.unet_diffusion
• credit.models.unet_downscaling
• credit.models.wxformer

Attributes

logger
_MODEL_REGISTRY
_CLASS_SOURCES

Functions

__getattr__(name)
register_model(model_type[, message])	Decorator that adds an external PyTorch model class to the model registry.
_load_model_entry(model_type)	Lazily import and return (model_class, log_message) for a registered model type.
load_fsdp_or_checkpoint_policy(conf)
load_custom_model_modules(conf)	Import every file listed under custom_models in the config.
load_model(conf[, load_weights, model_name])
load_model_name(conf, model_name[, load_weights])

Package Contents

credit.models.logger

credit.models._MODEL_REGISTRY

credit.models._CLASS_SOURCES

credit.models.__getattr__(name)

credit.models.register_model(model_type, message=None)

Decorator that adds an external PyTorch model class to the model registry.

Parameters:

• model_type – Key used in the config model.type field.

• message – Optional log message shown when the model is loaded.

Example:

@register_model("my_model", "Loading my custom model ...")
class MyModel(torch.nn.Module):
    ...

credit.models._load_model_entry(model_type)

Lazily import and return (model_class, log_message) for a registered model type.

credit.models.load_fsdp_or_checkpoint_policy(conf)

credit.models.load_custom_model_modules(conf)

Import every file listed under custom_models in the config.

Each file is executed as a standalone module. The expected use-case is that each file contains one or more classes decorated with @register_model, so the import triggers registration as a side-effect.

Parameters:

conf (dict) – Top-level config dict. If custom_models is absent or empty this function is a no-op.

Raises:

FileNotFoundError – If a listed path does not exist on disk.

credit.models.load_model(conf, load_weights=False, model_name=False)

credit.models.load_model_name(conf, model_name, load_weights=False)