run_pair --num-candidates 5 --work-dir test --target-count 16 --workers 20 --p-nonlinear 0.0 --max-params 1 --min-params 0 --min-vars 1 --max-vars 6 --max-assign 3 ...
In the current training implementation, each training batch only executes a single forward pass of the model. Specifically, in train_batch(), the code calls train_state.model(...) once, obtains the ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results