File size: 766 Bytes
22d888b | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 | """
RL BTC v4 — Offline Implicit Q-Learning for Bitcoin Trading.
Based on: "Offline Reinforcement Learning with Implicit Q-Learning" (Kostrikov et al., 2021)
https://hf.co/papers/2110.06169
Key innovation: learns from logged historical data without environment interaction.
Uses upper expectile value function to estimate the value of the best actions
without ever explicitly querying out-of-distribution actions.
"""
from .constants import (
DEFAULT_DATA_PATH,
MARKET_FEATURE_COLUMNS,
PORTFOLIO_FEATURE_COLUMNS,
ACTIONS,
N_ACTIONS,
ACTION_INDEX_BY_NAME,
STARTING_CASH,
DRAWDOWN_LIMIT,
)
from .env import BTCTradingEnv
from .dataset import build_offline_rl_dataset, OfflineRLDataset
from .iql_trainer import IQLTrainer, IQLConfig
|