search

ToTheBeginning / PuLID

[NeurIPS 2024] Official code for PuLID: Pure and Lightning ID Customization via Contrastive Alignment
Apache License 2.0
2.53k stars 176 forks source link

Why when I run app_flux.py, it killed #95

Open Zhuangvictor0 opened 1 month ago

Zhuangvictor0 commented 1 month ago

my environment: A100, GRAM 40G, VRAM 72G, it should not be the cache problem

image
ToTheBeginning commented 1 month ago

Due to the limited information contained in this error message, we are unable to assist you in troubleshooting.

ingtellect commented 4 weeks ago

For me, I had to allocate more ram for wsl.

my enviromnet: intel 13900k, 4090, 64GB RAM, windows 10 wsl ubuntu 22.04

.wslconfig

[wsl2]

# Limits VM memory to use no more than 48 GB, defaults to 50% of ram
memory=48GB

# Sets the VM to use 8 virtual processors
processors=8

# Sets the amount of swap storage space to 8GB, default is 25% of available RAM
swap=16GB

and run

python app_flux.py --offload --fp8
sysonlai commented 1 week ago

For me, I had to allocate more ram for wsl.

my enviromnet: intel 13900k, 4090, 64GB RAM, windows 10 wsl ubuntu 22.04

.wslconfig

[wsl2]

# Limits VM memory to use no more than 48 GB, defaults to 50% of ram
memory=48GB

# Sets the VM to use 8 virtual processors
processors=8

# Sets the amount of swap storage space to 8GB, default is 25% of available RAM
swap=16GB

and run

python app_flux.py --offload --fp8

With the same error in my WSL, it's working fine now when i use this setting~