noether.core.schemas.models.upt¶

Classes¶

Configuration for a UPT model.

Module Contents¶

class noether.core.schemas.models.upt.UPTConfig(/, **data)¶

Bases: noether.core.schemas.models.base.ModelBaseConfig, noether.core.schemas.mixins.InjectSharedFieldFromParentMixin

Configuration for a UPT model.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

model_config¶: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

num_heads: int = None¶: Number of attention heads in the model.

hidden_dim: int = None¶: Hidden dimension of the model.

mlp_expansion_factor: int = None¶: Expansion factor for the MLP of the FF layers.

approximator_depth: int = None¶: Number of approximator layers.

use_rope: bool = None¶

supernode_pooling_config: Annotated[noether.core.schemas.modules.SupernodePoolingConfig, noether.core.schemas.mixins.Shared]¶

approximator_config: Annotated[noether.core.schemas.modules.blocks.TransformerBlockConfig, noether.core.schemas.mixins.Shared]¶

decoder_config: Annotated[noether.core.schemas.modules.DeepPerceiverDecoderConfig, noether.core.schemas.mixins.Shared]¶

bias_layers: bool = None¶

data_specs: noether.core.schemas.dataset.AeroDataSpecs¶

linear_output_projection_config()¶

Return type:: noether.core.schemas.modules.layers.LinearProjectionConfig

rope_frequency_config()¶

Return type:: noether.core.schemas.modules.layers.RopeFrequencyConfig

validate_rope_usage()¶

Ensure that if use_rope is True in the main config, it is also True in the approximator_config.

Return type:: UPTConfig

update_supernode_pooling_config()¶

Inject shared fields into supernode_pooling_config.

Return type:: UPTConfig

pos_embedding_config()¶

Return type:: noether.core.schemas.modules.layers.ContinuousSincosEmbeddingConfig

validate_parameters()¶

Validate validity of parameters across the model and its submodules.

Ensures that: 1. hidden_dim is divisible by num_heads in parent and all submodules with num_heads 2. hidden_dim is consistent across parent and all submodules

Return type:: UPTConfig