-
when running `pretrain.py` with 1 or 4 GPUs and the DDPStrategy as described in the docs, I get the following error
```bash
"PyTorch/1.12.0-foss-2022a-CUDA-11.7.0/lib/python3.10/.../torch/distribu…
-
make a sqlite database with roughly the following schema
- one table with person characteristics that are fixed over time, ie gender, birth year, municipality of birth, ....
```sql
create tabl…
-
After a 3-4 hours of research, iterabledataset seems to be the right way to train on datasets that are too huge to load entirely in-memory.
The key point is a gpu will only load one batch at a tim…
-
### To include
- new reports
- summary statistics for fake data
- which are the relevant files
- statistics for numeric variables: mean, sd, N missing, min, max, covariance matrix
- fre…
-
# Broken Links
**This is an automated issue created in a weekly continuous integration job 🤖.**
This issue contains a list of links external to The Turing Way that appeared to be broken when the che…
-
#### Is your feature request related to a problem? Please describe.
In the metadata form in Yoda right now, you can select License and Access conditions independently of each other. However, I do not…
-
-
-
related to https://github.com/IQSS/dataverse/issues/3436
-
*All information shared will be publicly available. It will also be displayed on the map unless otherwise indicated*
**Name of repository:**
RODBUK
**Homepage URL of repository:**
https://rodb…