Computes and stores the average and current value

ResidualBlock introduced in: https://arxiv.org/abs/1512.03385 recommended architecture: http://torch.ch/blog/2016/02/04/resnets.html

Initialize the RMSNorm normalization layer. Args: dim (int): The dimension of the input tensor. eps (float, optional): A sma

Graph Attention Layer (GAT) as described in the paper `"Graph Attention Networks" <https://arxiv.org/pdf/1710.10903.pdf>`. This operatio

Graph Convolutional Layer described in "Semi-Supervised Classification with Graph Convolutional Networks". Given an input feature re

UpsampleConvLayer Upsamples the input and then does a convolution. This method gives better results compared to ConvTranspose2d. ref: http

Multi-head attention module. Args: model_args (ModelArgs): Model configuration arguments. Attributes: n_kv_heads (int):

an unassuming Transformer block

Assemble two parts as an nn.Module and define pipelining logic

FeedForward module Args: dim (int): Input dimension. hidden_dim (int): Hidden dimension of the feedforward layer. mu

Graph Attention Network (GAT) as described in the paper `"Graph Attention Networks" <https://arxiv.org/pdf/1710.10903.pdf>`. Consists of a 2-

Graph Convolutional Network (GCN) as described in the paper `"Semi-Supervised Classification with Graph Convolutional Networks" <https://arx

GPT Language Model

r""" The model consists of a sparse part and a dense part. 1) The dense part is an nn.Linear module that is replicated across all trainers usi

ModulePathTracer is an FX tracer that--for each operation--also records the qualified name of the Module from which the operation originated.

A multi-head masked self-attention layer with a projection at the end.

implements both actor and critic in one model

r""" Borrowing the ``Policy`` class from the Reinforcement Learning example. Copying the code to make these two examples independent. See

r""" Borrowing the ``Policy`` class from the Reinforcement Learning example. Copying the code to make these two examples independent. See

r"""Inject some information about the relative or absolute position of the tokens in the sequence. The positional encodings have the same dime

Container module with an encoder, a recurrent module, and a decoder.

Siamese network for image similarity estimation. The network is composed of two identical networks, one for each input. The o

MLP based model

MLP based model

TransformerBlock Module Args: layer_id (int): Identifier for the layer. model_args (ModelArgs): Model configuration argument

Container module with an encoder, a recurrent or transformer module, and a decoder.

r""" Decoding layers of the RNNModel

r""" Encoding layers of the RNNModel

r""" An observer has exclusive access to its own environment. Each observer captures the state from its environment, and send the state to the

r""" An observer has exclusive access to its own environment. Each observer captures the state from its environment, and send the state to the