fishy.models.deep¶

Deep learning models for spectral analysis.

Submodules

`augformer`	AugFormer: Multi-View Self-Attention over Augmented Spectral Views.
`augformer_v2`	AugFormerV2: Enhanced version of AugFormer with advancements from Parameter Golf.
`cnn`	Simple Convolutional Neural Network (CNN) model for spectral classification.
`dense`	Dense Neural Network model for spectral classification.
`ensemble`	Multi-Scale/Resolution Ensemble of Transformers.
`gatedmlp`	GatedMLP: A pure gated MLP baseline for REIMS spectral data.
`gmoe`	Sparsely-Gated Mixture of Experts (GMOE) for spectral data.
`hybrid`	Hybrid CNN-LSTM model for spectral classification.
`kan`	Kolmogorov-Arnold Network (KAN) model for spectral classification.
`lewm`	LEWM: LeWorldModel-inspired Gated MLP.
`lstm`	Bidirectional Gated Recurrent Unit (GRU) model for spectral classification.
`mamba`	Mamba model for spectral classification.
`moe`	Mixture of Experts (MoE) model for spectral data.
`nextformer`	NextFormer model for spectral analysis.
`ode`	Neural Ordinary Differential Equation (ODE) model for spectral classification.
`ordinal`
`performer`	Performer model for spectral classification.
`rcnn`	Residual Convolutional Neural Network (RCNN) for spectral classification.
`rwkv`	Receptance Weighted Key-Value (RWKV) model for spectral classification.
`srmlp`	SRMLP: Sign-Random Gated MLP.
`tcn`	Temporal Convolutional Network (TCN) model for spectral classification.
`transformer`	Transformer model for time series classification.
`ttt_mixin`
`vae`	Variational Autoencoder (VAE) for spectral classification.
`wavenet`	WaveNet model for spectral classification.