A class that keeps all conversation history.

Track a series of values and provide access to smoothed values over a window or the global series average.

The model can behave as an encoder (with only self-attention) as well as a decoder, in which case a layer of cross-attention is added between

Construct the embeddings from word and position embeddings.

r"""Dataset for chaining multiple :class:`DataPipeline` s. This class is useful to assemble different existing dataset streams. The chaining

This is a preliminary implementation to centralize and validate the configuration. May be altered in the future. A helper class to valid

Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).

A wrapper to convert DataLoader as an infinite iterator. Modified from: https://github.com/open-mmlab/mmcv/blob/master/mmcv/runner/i

Subclass torch's LayerNorm to handle fp16.

A utility class to create string representation of available model architectures and types. >>> from minigpt4.models import model_zoo >>

A simple wrapper for iterating over multiple iterators. Args: loaders (List[Loader]): List of Iterator loaders. ratios (List

Image to Patch Embedding

Modified from https://github.com/ChenRocks/UNITER. overlap compute and cuda data transfer (copied and then modified from nvidia apex)

Vision Transformer with support for patch or hybrid CNN input stage

Base class for models.

An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained models.

MiniGPT-4 model

Base class for MiniGPT-4 and MiniGPT-v2

MiniGPT-v2 model

A runner class to train and evaluate a model given a task and datasets. The runner uses pytorch distributed data parallel by default. Future

Different separator style.