PC Specs:
- GPU: RTX 4060 Ti (16GB VRAM)
- RAM: 16GB DDR4/DDR5
- CPU: Ryzen 5 5660x
- OS: Windows 11
Issue:
I'm trying to run the WAN 2.1 T2V-1.3B model using the following command:
python generate.py --task t2v-1.3B --size 832*480 --ckpt_dir ./Wan2.1-T2V-1.3B --offload_model True --t5_cpu --sample_shift 8 --sample_guide_scale 6 --prompt "Two anthropomorphic cats in comfy boxing gear and bright gloves fight intensely on a spotlighted stage."
used this model: Wan-AI/Wan2.1-T2V-1.3B
From my understanding, the model first loads into RAM before being offloaded to VRAM. However, it seems like my 16GB RAM doesn't seem to be sufficient, and the model fails to load properly.
I'm new to running these models, and according to the usage guide, it mentions that the model requires 8GB VRAM, but it doesn’t seem to be the case for me.
Questions:
- How much system RAM is required to run WAN 2.1 T2V-1.3B smoothly?
- Would using swap memory help in running this model?
- Are there any recommended optimizations (e.g., quantization, using GGUF, or smaller batch sizes) to make it fit within my setup?
Any guidance would be appreciated! Thanks in advance! 🙌
PC Specs:
Issue:
I'm trying to run the WAN 2.1 T2V-1.3B model using the following command:
used this model: Wan-AI/Wan2.1-T2V-1.3B
From my understanding, the model first loads into RAM before being offloaded to VRAM. However, it seems like my 16GB RAM doesn't seem to be sufficient, and the model fails to load properly.
I'm new to running these models, and according to the usage guide, it mentions that the model requires 8GB VRAM, but it doesn’t seem to be the case for me.
Questions:
Any guidance would be appreciated! Thanks in advance! 🙌