יwere can i find a colab notebook for this package ?

[underwater84]

hi im looking to try this package , i was wondering do you have a colab notebook i can try ?

[GuyWhoDoesentKnowShit]

from the read me

To reproduce the results in the paper, the codes are simplified as much as possible. You start with the settings inconfig_main.py file, where you set all the settings for:

The Walkforward, K-Cross Validation, and Combinatorial Purged Cross Validation (CPCV) methods. Set how many candles/data points you require for training and validation. Set which tickers you will download from Binance, the minimum buy limits. Set your technical indicators. Computes automatically the exact start and end dates for training and validation, respectively, based on your trade start date and end date. A short description of each folder:

data Contains all your training/validation data in the main folder, and a subfolder which contains trade_data after download using both 0_dl_trainval_data.py and 0_dl_trade_data.py (more later) drl_agents Contains the DRL framework ElegantRL which implements a series of model-free DRL algorithms plots_and_metrics Dump folder for all analysis images and performance metrics produced train Holds all utility functions for DRL training train_results After running either 1_optimize_cpcv.py / 1_optimize_kcv.py / 1_optimize_wf.py will have a folder with your trained DRL agents Then, running and producing similar results to that in the paper are simple, following the numbered Python files as indicated by the number of the filename:

0_dl_trainval_data.py Downloads the train and validation data according to config_main.py 0_dl_trade_data.py Downloads the trade data according to config_main.py 1_optimize_cpcv.py Optimizes hyperparameters with a Combinatorial Purged Cross-validation scheme 1_optimize_kcv.py Optimizes hyperparameters with a K-Fold Cross-validation scheme 1_optimize_wf.py Optimizes hyperparameters with a Walk-forward validation scheme 2_validate.py Shows insights about the training and validation process (select a results folder from train_results) 4_backtestpy Backtests trained DRL agents (enter multiple results folders from train_results in a list) 5_pbo.py Computes PBO for trained DRL agents (enter multiple results folders from train_results in a list) Simply run the scripts in the above order. Please note the trained agents are auto-saved to the folder train_results. That is where you can find your trained DRL agents!

[YangletLiu]

from the read me

To reproduce the results in the paper, the codes are simplified as much as possible. You start with the settings inconfig_main.py file, where you set all the settings for:

The Walkforward, K-Cross Validation, and Combinatorial Purged Cross Validation (CPCV) methods. Set how many candles/data points you require for training and validation. Set which tickers you will download from Binance, the minimum buy limits. Set your technical indicators. Computes automatically the exact start and end dates for training and validation, respectively, based on your trade start date and end date. A short description of each folder:

data Contains all your training/validation data in the main folder, and a subfolder which contains trade_data after download using both 0_dl_trainval_data.py and 0_dl_trade_data.py (more later) drl_agents Contains the DRL framework ElegantRL which implements a series of model-free DRL algorithms plots_and_metrics Dump folder for all analysis images and performance metrics produced train Holds all utility functions for DRL training train_results After running either 1_optimize_cpcv.py / 1_optimize_kcv.py / 1_optimize_wf.py will have a folder with your trained DRL agents Then, running and producing similar results to that in the paper are simple, following the numbered Python files as indicated by the number of the filename:

0_dl_trainval_data.py Downloads the train and validation data according to config_main.py 0_dl_trade_data.py Downloads the trade data according to config_main.py 1_optimize_cpcv.py Optimizes hyperparameters with a Combinatorial Purged Cross-validation scheme 1_optimize_kcv.py Optimizes hyperparameters with a K-Fold Cross-validation scheme 1_optimize_wf.py Optimizes hyperparameters with a Walk-forward validation scheme 2_validate.py Shows insights about the training and validation process (select a results folder from train_results) 4_backtestpy Backtests trained DRL agents (enter multiple results folders from train_results in a list) 5_pbo.py Computes PBO for trained DRL agents (enter multiple results folders from train_results in a list) Simply run the scripts in the above order. Please note the trained agents are auto-saved to the folder train_results. That is where you can find your trained DRL agents!

Thanks for clarification!