Test with real demo - Githubissues

vayneli commented 2 weeks ago

Thanks for your work. I would like to consult about tests with real data. After using process_quan.py to convert realdata, the test refiner does not seem to run correctly and an error will be reported in TestDataset. What changes need to be made? Looking forward to your reply, thanks again.

fhan235 commented 2 weeks ago

Hi, thanks for your interest in our work. The output of process_quan.py will be something like .pt. Usually, it doesn't need to make any changes, you can just specify the location of .pt and run the inference code like this python test_refiner.py --cfg config_decoder/refiner_realdemo.yaml To better address the issue, could you please provide more information about the error? Let me know if you need any further adjustments!

vayneli commented 2 weeks ago

Thanks for your reply, after getting the output pt file of process_quan, I run the inference code: python test_refiner.py --cfg config_decoder/refiner_realdemo.yaml,but I get the error like this: No implementation of “rotation _local_full_gt_list in TestDataset code below: for i in all_info: self.motions.append(i["rotation_local_full_gt_list"]),i also found that the output pt file of process_quan is indeed not include the key, so I would like to confirm whether I need to modify myself

Hi vayneli! The output of process_quan.py will be something like .pt. Usually, it doesn't need to make any changes, you can just specify the location of .pt and run the inference code like this python test_refiner.py --cfg config_decoder/refiner_realdemo.yaml To better address the issue, could you please provide more information about the error? Let me know if you need any further adjustments!

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

fhan235 commented 2 weeks ago

Hi, very sorry about that, previously, it was indeed not necessary to make any changes. I just checked the code and it seems that the wrong file version might have been uploaded. If you need it urgently, you can temporarily modify the TestDataset according to the output of process_quan.py. I will fix this issue as soon as possible.

vayneli commented 2 weeks ago

Okay, I get it, very appreciated for your kind reply.

Hi, very sorry about that, previously, it was indeed not necessary to make any changes. I just checked the code and it seems that the wrong file version might have been uploaded. If you need it urgently, you can temporarily modify the TestDataset according to the output of process_quan.py. I will fix this issue as soon as possible.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

fhan235 commented 2 weeks ago

Hi, to make inference with real data more convenient, I add a new file named inference_realdemo.py. You can now simply run python inference_realdemo.py to obtain the full body motion results in the form of a .pkl file. Based on my testing, it is currently producing the correct results.

If you encounter any issues while using it, please feel free to reach out

vayneli commented 2 weeks ago

Okay，I will download and try it, thanks for your kind replay again

Hi, to make inference with real data more convenient, I add a new file named inference_realdemo.py. You can now simply run python inference_realdemo.py to obtain the full body motion results in the form of a .pkl file. Based on my testing, it is currently producing the correct results.

image.png (view on web)image.png (view on web)

If you encounter any issues while using it, please feel free to reach out

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

Wenchao-M / SAGE

Test with real demo #4