Spaces:
Build error
Build error
File size: 1,582 Bytes
3366701 1f99294 deef182 15f8f33 3366701 467ad58 deef182 660e224 3366701 d0dcc2c 3366701 1f99294 d90d4fc 1f99294 3366701 660e224 3366701 ee173ac 3366701 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 |
import torch
from typing import Any
from transformers.agents.tools import Tool
from transformers.utils import is_accelerate_available
from diffusers import DiffusionPipeline
if is_accelerate_available():
from accelerate import PartialState
TEXT_TO_VIDEO_DESCRIPTION = (
"This is a tool that creates a video according to a text description. "
"It takes an optional input `seconds` which will be the duration of the video. "
"The default is of two seconds. The tool outputs a video object."
)
class TextToVideoTool(Tool):
default_checkpoint = "damo-vilab/text-to-video-ms-1.7b"
description = TEXT_TO_VIDEO_DESCRIPTION
name = "video_generator"
inputs = {"prompt": {"type": str, "description": "contains the image description"}}
output_type = Any
def __init__(self, device=None, **hub_kwargs) -> None:
if not is_accelerate_available():
raise ImportError("Accelerate should be installed in order to use tools.")
super().__init__()
self.device = device
self.pipeline = None
self.hub_kwargs = hub_kwargs
def setup(self):
if self.device is None:
self.device = PartialState().default_device
self.pipeline = DiffusionPipeline.from_pretrained(
self.default_checkpoint, variant="fp16"
)
self.pipeline.to(self.device)
self.is_initialized = True
def __call__(self, prompt, seconds=2):
if not self.is_initialized:
self.setup()
return self.pipeline(prompt, num_frames=8 * seconds).frames
|