You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I think there exists the type conflict of jax numpy and python's float() function.
In evojax/task/slimevolley.py, jax.numpy.float32 is used, and this is the tax scaler constructor of type float 32. But, Jax represents scalars as zero dimensional arrays, which causes the conflict with python's float() function.
When I ran examples/train_slimevolley.py file, I got this error after training the model.
TypeError: Only scalar arrays can be converted to Python scalars; got arr.ndim=1
The text was updated successfully, but these errors were encountered:
I think there exists the type conflict of jax numpy and python's float() function.
In evojax/task/slimevolley.py, jax.numpy.float32 is used, and this is the tax scaler constructor of type float 32. But, Jax represents scalars as zero dimensional arrays, which causes the conflict with python's float() function.
When I ran examples/train_slimevolley.py file, I got this error after training the model.
TypeError: Only scalar arrays can be converted to Python scalars; got arr.ndim=1
The text was updated successfully, but these errors were encountered: