Open jiriyu98 opened 11 months ago
Based on type annotation and according to tradition, I think this behavior is unexpected.
jnp.logical_and won't convert reward dtype from bool to float. Simply add *1.0 to convert it.
*1.0
Based on type annotation and according to tradition, I think this behavior is unexpected.
jnp.logical_and won't convert reward dtype from bool to float. Simply add
*1.0
to convert it.