Cogvideox I2V Locobot

🚀 CogVideoX LoRA - BelGio13/cogvideoX-I2V-locobot

这是用于图像到视频转换的CogVideoX LoRA模型，能够基于输入图像生成高质量的视频。

🚀 快速开始

你可以按照以下步骤使用该模型：

1. 下载模型的 *.safetensors LoRA 文件，可在 Files & versions 标签页 进行下载。
2. 使用以下代码示例，结合 🧨 diffusers 库 来运行模型。

import torch
from diffusers import CogVideoXImageToVideoPipeline
from diffusers.utils import load_image, export_to_video

pipe = CogVideoXImageToVideoPipeline.from_pretrained("THUDM/CogVideoX-5b-I2V", torch_dtype=torch.bfloat16).to("cuda")
pipe.load_lora_weights("BelGio13/cogvideoX-I2V-locobot", weight_name="pytorch_lora_weights.safetensors", adapter_name="cogvideox-i2v-lora")

# The LoRA adapter weights are determined by what was used for training.
# In this case, we assume `--lora_alpha` is 32 and `--rank` is 64.
# It can be made lower or higher from what was used in training to decrease or amplify the effect
# of the LoRA upto a tolerance, beyond which one might notice no effect at all or overflows.
pipe.set_adapters("cogvideox-i2v-lora", [32 / 64])

image = load_image("/path/to/image")
video = pipe(image=image, "", guidance_scale=6, use_dynamic_cfg=True).frames[0]
export_to_video(video, "output.mp4", fps=8)

更多细节，包括 LoRA 的加权、合并和融合等操作，请查看 diffusers 中加载 LoRA 的文档。

✨ 主要特性

• 这些是 BelGio13/cogvideoX-I2V-locobot 的 LoRA 权重，适用于 THUDM/CogVideoX-5b-I2V 基础模型。
• 权重使用 CogVideoX Diffusers 训练器 进行训练。
• 文本编码器未启用 LoRA。

📦 安装指南

暂未提供具体安装步骤，可参考上述快速开始部分结合 diffusers 库使用。

💻 使用示例

基础用法

import torch
from diffusers import CogVideoXImageToVideoPipeline
from diffusers.utils import load_image, export_to_video

pipe = CogVideoXImageToVideoPipeline.from_pretrained("THUDM/CogVideoX-5b-I2V", torch_dtype=torch.bfloat16).to("cuda")
pipe.load_lora_weights("BelGio13/cogvideoX-I2V-locobot", weight_name="pytorch_lora_weights.safetensors", adapter_name="cogvideox-i2v-lora")

# The LoRA adapter weights are determined by what was used for training.
# In this case, we assume `--lora_alpha` is 32 and `--rank` is 64.
# It can be made lower or higher from what was used in training to decrease or amplify the effect
# of the LoRA upto a tolerance, beyond which one might notice no effect at all or overflows.
pipe.set_adapters("cogvideox-i2v-lora", [32 / 64])

image = load_image("/path/to/image")
video = pipe(image=image, "", guidance_scale=6, use_dynamic_cfg=True).frames[0]
export_to_video(video, "output.mp4", fps=8)

高级用法

暂未提供高级用法示例。

📚 详细文档

模型描述

这些是 BelGio13/cogvideoX-I2V-locobot 的 LoRA 权重，用于 THUDM/CogVideoX-5b-I2V 模型。权重使用 CogVideoX Diffusers 训练器 进行训练，且文本编码器未启用 LoRA。

预期用途与限制

如何使用

# TODO: add an example code snippet for running this diffusion pipeline

限制和偏差

[TODO: provide examples of latent issues and potential remediations]

训练细节

[TODO: describe the data used to train the model]

📄 许可证

请遵守 此处 描述的许可条款。

📦 模型信息