fastvideo.pipelines.stages.encoding
#
Encoding stage for diffusion pipelines.
Module Contents#
Classes#
Stage for encoding pixel space representations into latent space. |
Data#
API#
- class fastvideo.pipelines.stages.encoding.EncodingStage(vae: fastvideo.models.vaes.common.ParallelTiledVAE)[source]#
Bases:
fastvideo.pipelines.stages.base.PipelineStage
Stage for encoding pixel space representations into latent space.
This stage handles the encoding of pixel-space video/images into latent representations for further processing in the diffusion pipeline.
Initialization
- forward(batch: fastvideo.pipelines.pipeline_batch_info.ForwardBatch, fastvideo_args: fastvideo.fastvideo_args.FastVideoArgs) fastvideo.pipelines.pipeline_batch_info.ForwardBatch [source]#
Encode pixel space representations into latent space.
- Parameters:
batch β The current batch information.
fastvideo_args β The inference arguments.
- Returns:
The batch with encoded latents.
- verify_input(batch: fastvideo.pipelines.pipeline_batch_info.ForwardBatch, fastvideo_args: fastvideo.fastvideo_args.FastVideoArgs) fastvideo.pipelines.stages.validators.VerificationResult [source]#
Verify encoding stage inputs.
- verify_output(batch: fastvideo.pipelines.pipeline_batch_info.ForwardBatch, fastvideo_args: fastvideo.fastvideo_args.FastVideoArgs) fastvideo.pipelines.stages.validators.VerificationResult [source]#
Verify encoding stage outputs.