Inference

[sumantha-NTS]

Hi,
I want to perform inference for human parsing using this model. What are the inputs required apart from the input image...?

I have already added the input image in JPEGImages directory. While inferencing I am getting assertion error at below line
File "/content/drive/MyDrive/POCs/InvPT/data/pascal_context.py", line 165, in init
assert os.path.isfile(_edge)
AssertionError

[prismformore]

@sumantha-NTS Thank you for your interest in our work. This model is very strong so it should perform well for multi-task learning in the real world. So far this project doesn't contain code for inference in the wild but I will consider adding it if more people are interested in it in future.

It is also easy for you to write the inference code based on main.py and utils/test_utils.py. The file-saving code is here: https://github.com/prismformore/InvPT/blob/3b70fcc5a4f7053a7e32a9f85da5dda670c18fba/evaluation/evaluate_utils.py#L69

Thank you.

[sumantha-NTS]

Hi,
I have created script to run inference for human parsing using this model which is attached below.
inference.py

input image:

output image:

As you can observe, i am getting black output...
Can you provide suggestion to resolve this..?

[prismformore]

@sumantha-NTS Hi, can you show your code?

[sumantha-NTS]

@sumantha-NTS Hi, can you show your code?

Already attached code above i.e., inference.py

[prismformore]

@sumantha-NTS Please check this commit https://github.com/prismformore/InvPT/blob/main/inference.py.

I implement multi-task prediction inference code for any image input using the PASCAL-Context checkpoint. I borrow a few lines of code from yours. Please check whether this is helpful to you.

An example:
CUDA_VISIBLE_DEVICES=0 python inference.py --image_path=IMAGE_PATH --ckp_path=CKP_PATH --save_dir=SAVE_DIR

[prismformore]

Some generated results.

[sumantha-NTS]

Thank you..