1.6 KiB
ensure we use GPU if available to train faster. during training we need to have RL loop that looks at streaming data, and retrospective backtesting/training on predictions. sincr the start of the traing we're only loosing. implement robust penalty and analysis when closing a loosing trade and improve the reward function.
2025-03-10 12:11:28,651 - INFO - Initialized environment with 500 candles
C:\Users\popov\miniforge3\Lib\site-packages\torch\nn\modules\transformer.py:385: UserWarning: enable_nested_tensor is True, but self.use_nested_tensor is False because encoder_layer.self_attn.batch_first was not True(use batch_first for better inference performance)
warnings.warn(
main.py:1105: FutureWarning: torch.cuda.amp.GradScaler(args...) is deprecated. Please use torch.amp.GradScaler('cuda', args...) instead.
self.scaler = amp.GradScaler()
C:\Users\popov\miniforge3\Lib\site-packages\torch\amp\grad_scaler.py:132: UserWarning: torch.cuda.amp.GradScaler is enabled, but CUDA is not available. Disabling.
warnings.warn(
2025-03-10 12:11:30,927 - INFO - Starting training for 1000 episodes...
2025-03-10 12:11:30,927 - INFO - Starting training on device: cpu
2025-03-10 12:11:30,928 - ERROR - Training failed: 'TradingEnvironment' object has no attribute 'initialize_price_predictor'
2025-03-10 12:11:30,928 - INFO - Exchange connection closed
Backend tkagg is interactive backend. Turning interactive mode on.
2025-03-10 12:35:14,489 - INFO - Episode 34: Reward=232.41, Balance=$98.47, Win Rate=70.6%, Trades=17, Episode PnL=$-1.33, Total PnL=$-559.78, Max Drawdown=7.0%, Pred Accuracy=99.9%