-
Notifications
You must be signed in to change notification settings - Fork 64
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Assertion failure when running vanilla train_nav.py #5
Comments
Me too ! |
I use this function and replace default fn here to get rid of bad data. But I found that the peak performance is reached at the end of 2nd epoch(epoch 1 with d_D_50=1.36). After that, performance drops to e.g, epoch 6, d_D_50=1.22 |
@ZhuFengdaaa, thanks so much for that! By d_D_50, you mean \Delta d_50, right? If that's the case, 1.36 is not matching with the values reported in the paper (2.38). Did you manage to get the reported values? |
Hi~I meet the same error. Do you solve the problem? |
@ZhuFengdaaa , thanks so much for that. I follow your change, but I still get this error. Could you please give me some advice? Thanks! |
Found some mismatches in eqa_v1.json which leads to an AssertionError.
house.json
|
I met the same assertion errors here too. Any suggestions? |
Surprise! Try SUNCG_v0! |
I think my dataset version is v0. I don't know what actually went wrong. |
I'm running into an assertion error when running the train_nav.py
I'm able to train VQA models, but when I run train_nav.py (after setting the -target_obj_conn_map_dir to the appropriate path on my system), the code starts training on the first epoch, reaching about 2% before failing an assertion in training/data.py.
The issued command matches the github example:
I sort of assume I have failed to download or move some specific file, but as far as I can tell everything checks out, so I thought it could be a bug from the changes you guys have been making recently.
Let me know what you think; I'd really appreciate it so we can get the system running!
The text was updated successfully, but these errors were encountered: