question about scripts/run.sh
cwwjyh opened this issue · comments
Hello, this is a good work!
I ran "sh scripts/run.sh input/g.png exp/examples/name", but encountered an issue as below:
When running: python utils/ldm_utils/main.py -t --data_root exp/examples/name/png/ --logdir exp/examples/name/ldm/ --reg_data_root data/dreambooth_data/class_man_images/ --bg_root data/dreambooth_data/bg_images/ --class_word man --no-test --gpus 2,3
Can you give me some advice? Thank you very much!
Hello, this is a good work! I ran "sh scripts/run.sh input/g.png exp/examples/name", but encountered an issue as below:
When running: python utils/ldm_utils/main.py -t --data_root exp/examples/name/png/ --logdir exp/examples/name/ldm/ --reg_data_root data/dreambooth_data/class_man_images/ --bg_root data/dreambooth_data/bg_images/ --class_word man --no-test --gpus 2,3
Can you give me some advice? Thank you very much!
Hi, i have solved this problem
@cwwjyh How did you fix it?
@cwwjyh How did you fix it?
Because the GPU setting is wrong, it can run after modification.
@cwwjyh
Thanks. As per the example, sh scripts/run.sh input/examples/name.img exp/examples/name
....
where is the folder and files located ?
Are, do I need to provide it myself ?
And what is in your g.png ? A human pose ?
@cwwjyh How did you fix it?
Because the GPU setting is wrong, it can run after modification.
Hello, may I ask how much memory each of your GPUs has? I encountered Out-Of-Memory (OOM) errors when running the code on a server with 4 GPUs, each having 24GB of VRAM. Thanks.
@cwwjyh How did you fix it?
Because the GPU setting is wrong, it can run after modification.
Hello, may I ask how much memory each of your GPUs has? I encountered Out-Of-Memory (OOM) errors when running the code on a server with 4 GPUs, each having 24GB of VRAM. Thanks.
I use 32GB of VRAM and reset batch_size=6 in test.yaml. If you use 24GB, you can try to set batch_size=1.
@cwwjyh How did you fix it?
Because the GPU setting is wrong, it can run after modification.
Hello, may I ask how much memory each of your GPUs has? I encountered Out-Of-Memory (OOM) errors when running the code on a server with 4 GPUs, each having 24GB of VRAM. Thanks.
I use 32GB of VRAM and reset batch_size=6 in test.yaml. If you use 24GB, you can try to set batch_size=1.
Hello, I would like to double-check one thing. Since I couldn't find the test.yaml file, are you referring to the v1-finetune.yaml file?
@cwwjyh How did you fix it?
Because the GPU setting is wrong, it can run after modification.
Hello, may I ask how much memory each of your GPUs has? I encountered Out-Of-Memory (OOM) errors when running the code on a server with 4 GPUs, each having 24GB of VRAM. Thanks.
I use 32GB of VRAM and reset batch_size=6 in test.yaml. If you use 24GB, you can try to set batch_size=1.
Hello, I would like to double-check one thing. Since I couldn't find the test.yaml file, are you referring to the v1-finetune.yaml file?
![]()
@cwwjyh How did you fix it?
Because the GPU setting is wrong, it can run after modification.
Hello, may I ask how much memory each of your GPUs has? I encountered Out-Of-Memory (OOM) errors when running the code on a server with 4 GPUs, each having 24GB of VRAM. Thanks.
I use 32GB of VRAM and reset batch_size=6 in test.yaml. If you use 24GB, you can try to set batch_size=1.
Hello, I would like to double-check one thing. Since I couldn't find the test.yaml file, are you referring to the v1-finetune.yaml file?
![]()
Haha, perhaps you made a mistake. This work is from TeCH, not HumanGaussian.
@cwwjyh How did you fix it?
Because the GPU setting is wrong, it can run after modification.
Hello, may I ask how much memory each of your GPUs has? I encountered Out-Of-Memory (OOM) errors when running the code on a server with 4 GPUs, each having 24GB of VRAM. Thanks.
I use 32GB of VRAM and reset batch_size=6 in test.yaml. If you use 24GB, you can try to set batch_size=1.
haha, sorry. I use 2GPU 32GB of VRAM in TeCH project. Then it can work.
@cwwjyh How did you fix it?
Because the GPU setting is wrong, it can run after modification.
Hello, may I ask how much memory each of your GPUs has? I encountered Out-Of-Memory (OOM) errors when running the code on a server with 4 GPUs, each having 24GB of VRAM. Thanks.
I use 32GB of VRAM and reset batch_size=6 in test.yaml. If you use 24GB, you can try to set batch_size=1.
haha, sorry. I use 2GPU 32GB of VRAM in TeCH project. Then it can work.
By the way, may I ask if the final output file 000_texture.obj
contains color? My 000_texture.obj
file is the same as 000_geometry.obj
, both lacking color. When I proceed with the optional code below, the program becomes stuck.
python cores/main.py --config configs/tech_texture_export.yaml --exp_dir $EXP_DIR --sub_name $SUBJECT_NAME --test
And I'm not sure if running the final script generates an OBJ file with color.
@glorioushonor
I get an error when I run this line, so I'm not sure what the end result will be.
@glorioushonor
I get an error when I run this line, so I'm not sure what the end result will be.
What problem have you encountered? Maybe I can help you.
@glorioushonor
I encounter this problem. The command in the red box doesn't seem to run, it then just runs the next line, resulting in no g_geometry.obj file.
@glorioushonor
I encounter this problem. The command in the red box doesn't seem to run, it then just runs the next line, resulting in no g_geometry.obj file.
I suggest commenting out the commands after this line of code in the script first to obtain the error message and locate the issue.