LeanQuant commited on
Commit
93a5634
·
verified ·
1 Parent(s): 466f022

Add files using upload-large-folder tool

Browse files
Files changed (1) hide show
  1. README.md +6 -5
README.md CHANGED
@@ -113,13 +113,14 @@ We apply Huffman coding to the exponent bits of BFloat16 model weights, which ar
113
 
114
  4. To run without CPU offloading (40GB VRAM required):
115
  ```bash
116
- python t2v.py
117
  ```
118
 
119
- To run with CPU offloading (22.5GB VRAM required):
120
- ```bash
121
- python t2v.py --cpu_offload
122
- ```
 
123
 
124
  ### 📄 Learn More
125
 
 
113
 
114
  4. To run without CPU offloading (40GB VRAM required):
115
  ```bash
116
+ PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True python t2v.py
117
  ```
118
 
119
+ To run with CPU offloading (22.5GB VRAM required):
120
+ ```bash
121
+ PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True python t2v.py --cpu_offload
122
+ ```
123
+ > Setting `PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True` is strongly recommended to prevent out-of-memory errors caused by GPU memory fragmentation.
124
 
125
  ### 📄 Learn More
126