Implements a decoupling of parameters to allow freezing (or not) a subset of the embeddings. In practise, the regular `weight` can be trained

A class that keeps all conversation history.

Implements a decoupling of parameters to allow freezing (or not) a subset of the parameters. In practise, the regular `weight` can be trained

r""" This is the configuration class to store the configuration of a [`LlamaModel`]. It is used to instantiate an LLaMA model according to the

r""" This is the configuration class to store the configuration of a [`LlamaModel`]. It is used to instantiate an LLaMA model according to the

r""" TThis is the configuration class to store the configuration of a [`MistralModel`]. It is used to instantiate an Mistral model according t

r""" This is the configuration class to store the configuration of a [`MistralModel`]. It is used to instantiate an Mistral model according to

Collates examples containing text-only or text+image/video data. 1) Text sequences (input_ids, attention_mask, labels) are padded to the max

The are running variables that are used to tell when to increment GBS and when to stop doing that, they are never set directly in the config file,

r""" This is the configuration class to store the configuration of a [`IdeficsModel`]. It is used to instantiate an Idefics model according to

r""" This is the configuration class to store the configuration of a [`IdeficsModel`]. It is used to instantiate an Idefics model according to

Multi-headed attention from 'Attention Is All You Need' paper

Types of logging to use for the gradient and parameter statistics

Trainer object to monitor training and validation :- config -: json config file

This debug class helps detect and understand where the model starts getting very large or very small, and more importantly `nan` or `inf` act

Apply a bitmap-style transformation to an image. This transformation replaces all pixel values below a certain threshold with a specified va

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

Data Parameters

Apply dilation operation to an image. Dilation is a morphological operation that expands the white regions in a binary image. Args:

Apply erosion operation to an image. Erosion is a morphological operation that shrinks the white regions in a binary image. Args:

These are init variables that are used to set up the GBS ramp up protocol

General Hyperparameters

General Hyperparameters

General Hyperparameters

r""" This is the configuration class to store the configuration of a [`IdeficsModel`]. It is used to instantiate an Idefics model according to

Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`LlamaDecoderLayer`] Args: config: IdeficsConf

In context learning parameters

Multi-headed attention from 'Attention Is All You Need' paper. Modified to use sliding window attention: Longformer and "Generating Long Sequ

Mistral flash attention module. This module inherits from `MistralAttention` as the weights of the module stays untouched. The only required