training fixes
This commit is contained in:
@@ -84,7 +84,8 @@ impl Worker {
|
||||
// println!("{}", before.elapsed().as_millis());
|
||||
episode_buffer.push(results);
|
||||
|
||||
let temp = if ply_count < 30 { 1.0 } else { 0.2 };
|
||||
let temp = (1.0 - (ply_count as f32 / 200.0).min(1.0)).max(0.3);
|
||||
// let temp = if ply_count < 30 { 1.0 } else { 0.2 };
|
||||
|
||||
let adjusted = apply_temperature(&episode_buffer.last().unwrap().move_dist, temp);
|
||||
|
||||
@@ -125,9 +126,11 @@ impl Worker {
|
||||
_ => 0.0,
|
||||
};
|
||||
|
||||
let blended_target = 0.75 * outcome + 0.25 * result.value;
|
||||
|
||||
samples.push(TrainingSample::from_mcts_with_outcome(
|
||||
result.clone(),
|
||||
outcome,
|
||||
blended_target,
|
||||
));
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user