Training & Optimization

New Training Job
3
Active Jobs
12
Completed
2.3h
Avg Duration
1.42
Best Sharpe

RL Training

Train reinforcement learning agents with PPO

Parameter Optimization

Optimize strategy parameters with Optuna

Hybrid Training

Combine DSL signals with RL agents

Filter by type:

Training Jobs

PPO EURUSD Strategy

RLrunningStarted: 1/15/2024, 5:30:00 AM
View Details
Progress65%
Episodes
1,300
Reward
0.420
Sharpe
1.18
Loss
0.0230
Duration: 14951h 9m
ETA: 9:30:00 AM

RSI Parameter Optimization

OPTIMIZATIONcompletedStarted: 1/14/2024, 4:00:00 AM
View Details
Progress100%
Episodes
2,000
Reward
0.380
Sharpe
1.45
Loss
0.0150
Duration: 14976h 39m

Hybrid MACD+RL

HYBRIDqueuedStarted: 1/15/2024, 6:00:00 AM
View Details
Progress0%
Duration: 14950h 39m

Training Configuration

RL Parameters

  • Algorithm: PPO (Proximal Policy Optimization)
  • Learning Rate: 3e-4 (adaptive)
  • Batch Size: 64
  • Gamma: 0.99
  • Early Stopping: Patience on OOS Sharpe

Anti-Overfitting

  • Validation: Walk-forward + CPCV
  • Risk Score: Composite overfit detection
  • Robustness: Stress testing enabled
  • Leakage: Static analysis on all strategies
  • Seeds: Fixed for reproducibility