Skip to content
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.

Sign up for GitHub

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

Update GETTING_STARTED.md #1008

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Update GETTING_STARTED.md #1008

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion readme/GETTING_STARTED.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -83,7 +83,7 @@ By default, pytorch evenly splits the total batch size to each GPUs.
`--master_batch` allows using different batchsize for the master GPU, which usually costs more memory than other GPUs.
If it encounters GPU memory out, using slightly less batch size (e.g., `112` of `128`) with the same learning is fine.

If the training is terminated before finishing, you can use the same commond with `--resume` to resume training. It will found the lastest model with the same `exp_id`.
If the training is terminated before finishing, you can use the same commond with `--resume` to resume training. It will find the lastest model with the same `exp_id`.

Our HourglassNet model is finetuned from the pretrained [ExtremeNet model](https://drive.google.com/file/d/1omiOUjWCrFbTJREypuZaODu0bOlF_7Fg/view?usp=sharing) (from the [ExtremeNet repo](https://github.com/xingyizhou/ExtremeNet)).
You will need to download the model, run `python convert_hourglass_weight.py` to convert the model format, and load the model for training (see the [script](../experiments/ctdet_coco_hg.sh)).