image
#7
by
ajeetpatel
- opened
- README.md +0 -9
- assets/undistilled/A majestic lion strides across the golden savanna, its powerful frame glistening under the warm afte.gif +3 -0
- assets/undistilled/A superintelligent humanoid robot waking up. The robot has a sleek metallic body with futuristic des.gif +3 -0
- assets/undistilled/Man walking his dog in the woods on a hot sunny day/n.gif +3 -0
- assets/undistilled/fox in the forest close-up quickly turned its head to the left/n.gif +3 -0
README.md
CHANGED
|
@@ -47,12 +47,3 @@ We provide some qualitative comparison between FastHunyuan 6 step inference v.s.
|
|
| 47 |
|  |  |
|
| 48 |
|  |  |
|
| 49 |
|
| 50 |
-
## Memory requirements
|
| 51 |
-
|
| 52 |
-
Please check our github repo for details. https://github.com/hao-ai-lab/FastVideo
|
| 53 |
-
|
| 54 |
-
For inference, we can inference FastHunyuan on single RTX4090. We now support NF4 and LLM-INT8 quantized inference using BitsAndBytes for FastHunyuan. With NF4 quantization, inference can be performed on a single RTX 4090 GPU, requiring just 20GB of VRAM.
|
| 55 |
-
|
| 56 |
-
For Lora Finetune, minimum hardware requirement
|
| 57 |
-
- 40 GB GPU memory each for 2 GPUs with lora
|
| 58 |
-
- 30 GB GPU memory each for 2 GPUs with CPU offload and lora.
|
|
|
|
| 47 |
|  |  |
|
| 48 |
|  |  |
|
| 49 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
assets/undistilled/A majestic lion strides across the golden savanna, its powerful frame glistening under the warm afte.gif
ADDED
|
Git LFS Details
|
assets/undistilled/A superintelligent humanoid robot waking up. The robot has a sleek metallic body with futuristic des.gif
ADDED
|
Git LFS Details
|
assets/undistilled/Man walking his dog in the woods on a hot sunny day/n.gif
ADDED
|
assets/undistilled/fox in the forest close-up quickly turned its head to the left/n.gif
ADDED
|