fastvideo.pipelines.preprocess.preprocess_pipeline_i2v

Contents

`fastvideo.pipelines.preprocess.preprocess_pipeline_i2v`#

I2V Data Preprocessing pipeline implementation.

This module contains an implementation of the I2V Data Preprocessing pipeline using the modular pipeline architecture.

Module Contents#

Classes#

PreprocessPipeline_I2V

I2V preprocessing pipeline implementation.

Data#

API#

fastvideo.pipelines.preprocess.preprocess_pipeline_i2v.EntryClass#: None

class fastvideo.pipelines.preprocess.preprocess_pipeline_i2v.PreprocessPipeline_I2V(model_path: str, fastvideo_args: fastvideo.fastvideo_args.FastVideoArgs | fastvideo.fastvideo_args.TrainingArgs, required_config_modules: list[str] | None = None, loaded_modules: dict[str, torch.nn.Module] | None = None)[source]#

Bases: fastvideo.pipelines.preprocess.preprocess_pipeline_base.BasePreprocessPipeline

I2V preprocessing pipeline implementation.

Initialization

Initialize the pipeline. After init, the pipeline should be ready to use. The pipeline should be stateless and not hold any batch state.

create_pipeline_stages(fastvideo_args: fastvideo.fastvideo_args.FastVideoArgs)[source]#

create_record(video_name: str, vae_latent: numpy.ndarray, text_embedding: numpy.ndarray, valid_data: dict[str, Any], idx: int, extra_features: dict[str, Any] | None = None) → dict[str, Any][source]#: Create a record for the Parquet dataset with CLIP features.

get_extra_features(valid_data: dict[str, Any], fastvideo_args: fastvideo.fastvideo_args.FastVideoArgs) → dict[str, Any][source]#

get_schema_fields() → list[str][source]#: Get the schema fields for I2V pipeline.