You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+4-6Lines changed: 4 additions & 6 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -22,7 +22,8 @@ Experience the CogVideoX-5B model online at <a href="https://huggingface.co/spac
22
22
23
23
## Project Updates
24
24
25
-
- 🔥🔥 News: ```2024/11/08```: We have released the CogVideoX1.5 model. CogVideoX1.5 is an upgraded version of the open-source model CogVideoX.
25
+
- 🔥🔥 **News**: ```2024/11/14```: We released the `CogVideoX1.5` model in the diffusers version. Only minor parameter adjustments are needed to continue using previous code.
26
+
- 🔥 News: ```2024/11/08```: We have released the CogVideoX1.5 model. CogVideoX1.5 is an upgraded version of the open-source model CogVideoX.
26
27
The CogVideoX1.5-5B series supports 10-second videos with higher resolution, and CogVideoX1.5-5B-I2V supports video generation at any resolution.
27
28
The SAT code has already been updated, while the diffusers version is still under adaptation. Download the SAT version code [here](https://huggingface.co/THUDM/CogVideoX1.5-5B-SAT).
28
29
- 🔥 **News**: ```2024/10/13```: A more cost-effective fine-tuning framework for `CogVideoX-5B` that works with a single
@@ -279,15 +280,12 @@ pipe.vae.enable_tiling()
279
280
used to quantize the text encoder, transformer, and VAE modules to reduce the memory requirements of CogVideoX. This
280
281
allows the model to run on free T4 Colabs or GPUs with smaller memory! Also, note that TorchAO quantization is fully
281
282
compatible with `torch.compile`, which can significantly improve inference speed. FP8 precision must be used on
282
-
devices with NVIDIA H100 and above, requiring source installation of `torch`, `torchao`, `diffusers`, and `accelerate`
283
-
Python packages. CUDA 12.4 is recommended.
283
+
devices with NVIDIA H100 and above, requiring source installation of `torch`, `torchao` Python packages. CUDA 12.4 is recommended.
284
284
+ The inference speed tests also used the above memory optimization scheme. Without memory optimization, inference speed
285
285
increases by about 10%. Only the `diffusers` version of the model supports quantization.
286
286
+ The model only supports English input; other languages can be translated into English for use via large model
287
287
refinement.
288
-
+ The memory usage of model fine-tuning is tested in an `8 * H100` environment, and the program automatically
289
-
uses `Zero 2` optimization. If a specific number of GPUs is marked in the table, that number or more GPUs must be used
0 commit comments