cerebras.modelzoo.data.multimodal.llava.LlavaHDF5MapDataProcessor.LlavaHDF5MapDataProcessorConfig#

class cerebras.modelzoo.data.multimodal.llava.LlavaHDF5MapDataProcessor.LlavaHDF5MapDataProcessorConfig(*args, **kwargs)[source]#

Bases: cerebras.modelzoo.data.common.h5_map_dataset.dataset.MultiModalHDF5DatasetConfig, cerebras.modelzoo.config.data_config.DataConfig

Methods

`check_for_deprecated_fields`
`check_literal_discriminator_field`
`check_mutual_exclusivity`
`copy`
`get_orig_class`
`get_orig_class_args`
`model_copy`
`model_post_init`
`post_init`

Attributes

`batch_size`	The batch size
`bos_token_id`
`data_dir`	The path to the HDF5 files.
`data_subset`	An optional specification to only consider a subset of the full dataset, useful for sequence length scheduling and multi-epoch testing.
`dataset_map_fn`
`discriminator`
`discriminator_value`
`drop_last`	Similar to the PyTorch drop_last setting except that samples that when set to True, samples that would have been dropped at the end of one epoch are yielded at the start of the next epoch so that there is no data loss.
`image_data_size`	The final C x H x W shape of the image.
`img_data_dir`	The path to the directory containing the images.
`max_sequence_length`	The sequence length of samples produced by the dataloader.
`mixed_precision`
`mixture`	An optional specification of multiple datasets to mix over to create one single weighted combination.
`model_config`
`num_samples`	The number of samples to shuffle over (if shuffling is enabled).
`num_workers`	The number of PyTorch processes used in the dataloader.
`pad_last`	Flag to enable padding of the last batch so that the last batch has the same batch size as the rest of the batches.
`persistent_workers`	Whether or not to keep workers persistent between epochs.
`pos_token_id`
`prefetch_factor`	The number of batches to prefetch in the dataloader.
`shuffle`	Whether or not to shuffle the dataset.
`shuffle_seed`	The seed used for deterministic shuffling.
`sort_files`	Whether or not the reader should sort the input files.
`transforms`	A specification of the torchvision transforms.
`use_vsl`	Flag to enable variable sequence length training.
`use_worker_cache`	Whether or not to copy data to storage that is directly attached to each individual worker node.
`vocab_size`
`data_processor`

num_workers = 0#: The number of PyTorch processes used in the dataloader.

prefetch_factor = 10#: The number of batches to prefetch in the dataloader.

persistent_workers = True#: Whether or not to keep workers persistent between epochs.

cerebras.modelzoo.data.multimodal.llava.LlavaHDF5MapDataProcessor.LlavaHDF5MapDataProcessor

cerebras.modelzoo.data.multimodal.llava.MultimodalSimpleHDF5MapDataProcessor