Colab notebooks do not work

[qo4on]

System Info / 系統信息

Your Colab notebooks do not work on T4 GPU.
Your session crashed after using all available RAM.
CogVideoX-5B-T2V-Int8-Colab.ipynb
CogVideoX-5B-I2V-Colab.ipynb

But this returns GPU OutOfMemoryError.
CogVideoX-5B-T2V-Colab.ipynb

100%
 50/50 [1:20:09<00:00, 103.47s/it]
---------------------------------------------------------------------------
OutOfMemoryError                          Traceback (most recent call last)
[<ipython-input-11-f4c854f12377>](https://localhost:8080/#) in <cell line: 0>()
----> 1 video = pipe(prompt=prompt, guidance_scale=6, use_dynamic_cfg=True, num_inference_steps=50).frames[0]

17 frames
[/usr/local/lib/python3.11/dist-packages/diffusers/models/autoencoders/autoencoder_kl_cogvideox.py](https://localhost:8080/#) in forward(self, f, zq, conv_cache)
    193 
    194         norm_f = self.norm_layer(f)
--> 195         new_f = norm_f * conv_y + conv_b
    196         return new_f, new_conv_cache
    197 

OutOfMemoryError: CUDA out of memory. Tried to allocate 676.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 196.12 MiB is free. Process 26021 has 14.55 GiB memory in use. Of the allocated memory 14.13 GiB is allocated by PyTorch, and 295.13 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation.  See documentation for Memory Management  (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)

Information / 问题信息

• The official example scripts / 官方的示例脚本
• My own modified scripts / 我自己修改的脚本和任务

Reproduction / 复现过程

Run notebooks in Colab.

Expected behavior / 期待表现

Without errors.

[zRzRzRzRzRzRzR]

This part needs to be retained instead of commented out.