video_captioning_from_summarizer_mapper¶

Mapper to generate video captions by summarizing several kinds of generated texts (captions from video/audio/frames, tags from audio/frames, ...)

通过总结多种生成的文本（来自视频/音频/帧的字幕、来自音频/帧的标签等）来生成视频字幕的映射器。

Type 算子类型: mapper

Tags 标签: cpu, hf, multimodal

🔧 Parameter Configuration 参数配置¶

name 参数名	type 类型	default 默认值	desc 说明
`hf_summarizer`	<class 'str'>	`None`	the summarizer model used to summarize texts
`trust_remote_code`	<class 'bool'>	`False`
`consider_video_caption_from_video`	<class 'bool'>	`True`	whether to consider the video
`consider_video_caption_from_audio`	<class 'bool'>	`True`	whether to consider the video
`consider_video_caption_from_frames`	<class 'bool'>	`True`	whether to consider the
`consider_video_tags_from_audio`	<class 'bool'>	`True`	whether to consider the video
`consider_video_tags_from_frames`	<class 'bool'>	`True`	whether to consider the video
`vid_cap_from_vid_args`	typing.Optional[typing.Dict]	`None`	the arg dict for video captioning from
`vid_cap_from_frm_args`	typing.Optional[typing.Dict]	`None`	the arg dict for video captioning from
`vid_tag_from_aud_args`	typing.Optional[typing.Dict]	`None`	the arg dict for video tagging from audio
`vid_tag_from_frm_args`	typing.Optional[typing.Dict]	`None`	the arg dict for video tagging from
`keep_tag_num`	typing.Annotated[int, Gt(gt=0)]	`5`	max number N of tags from sampled frames to keep.
`keep_original_sample`	<class 'bool'>	`True`	whether to keep the original sample. If
`args`		`''`	extra args
`kwargs`		`''`	extra args

not available 暂无