Open flynnamy opened 2 years ago
Hi @flynnamy , I am glad you are trying to use our model :) Could you post the command line you used?
Hi, I use command like this: python3 evaluate_multiview.py --resume=./4_views_mha64_gt.pth --device=0 @BrianG13
I run the repo on a clean machine and it works as expected. Can you please copy&paste here the entire logs that are printed on terminal?
Could you please log your conda list
output ?
I tried on 3 different machines and the results I am getting are as expected. I will update the code with an option to run over cpu later this week (Which GPU are you using?)
same problem using RTX3090 torch=1.8 cuda 11.1
@tbcvContributor did you created your conda env with the environment file?
Could you please log your conda list
output ?
@tbcvContributor did you created your conda env with the environment file? Could you please log your
conda list
output ? conda-list.txt
@tbcvContributor I see we are using different tensor versions and also different cuda drivers.
Yours: torch=1.8.0+cu111
Mine: pytorch=1.4.0=py3.6_cuda10.0.130_cudnn7.6.3_0
Maybe this is the root for the differences, in the next days I will try to run with your pkgs and see if this is the reason for different results
Could you try to create a conda list from the requirements file published in this repo and retry the experiment?
same question.
@Yamato-01 Could you please log your conda list
output ?
conda_list.txt here~
same problem
I'm having a similar problem evaluating the model.
+--------------+------------+---------------------+
| Action | MPJPE (mm) | Acc. Error (mm/s^2) |
+--------------+------------+---------------------+
| Directions | 59.38 | 0.83 |
| Discussion | 94.65 | 1.18 |
| Eating | 84.15 | 0.93 |
| Greeting | 82.13 | 1.38 |
| Phoning | 84.43 | 0.94 |
| Photo | 111.49 | 0.95 |
| Posing | 77.40 | 0.88 |
| Purchases | 105.69 | 1.66 |
| Sitting | 99.25 | 0.60 |
| SittingDown | 106.31 | 1.03 |
| Smoking | 89.77 | 0.83 |
| Waiting | 87.34 | 0.82 |
| WalkDog | 141.06 | 2.94 |
| WalkTogether | 97.24 | 2.66 |
| Walking | 92.90 | 2.57 |
| | | |
| Average | 94.80 | 1.35 |
+--------------+------------+---------------------+
Environment:
Did u guys installed the environment as specified?
same problem
conda list.txt
Did u guys installed the environment as specified?
I tried recreating the environment, both with pip
and conda
. I'm installing the exact same version of libraries (with the same CPython version). The conda
environment works fine and produces the expected results, but the pip
environment is outputting incorrect outputs.
Any guesses as to what might be the issue?
Hi,when I do evaluation by using your model but could not get your results. Have you checked the shared model ? @BrianG13![image](https://user-images.githubusercontent.com/20398890/144604100-72831cb5-8894-419a-92e7-f917f156b481.png)