Mior fixes.

README.md

@@ -18,6 +18,11 @@ pip3 install -r requirements.txt
 All the code in this repository exploits [LOBSTER](https://lobsterdata.com) data. To have an overview on their structure, please refer
 to the official documentation available at the following [link](https://lobsterdata.com/info/DataStructure.php).
 
+# Preliminary operations
+Before starting any experiment:
+- Open the ```lightning_batch_gd.py``` file and insert the Weights & Biases project's name and API key.
+- Open the ```utils.py``` file and set the standard values of the parameters.
+
 ## Usage
 To start an experiment from scratch, you need to follow these steps:
 - Place the raw data in the `data/nasdaq/raw` folder. The data must be in the LOBSTER format and each folder must be named with the asset's name (e.g. AAPL for Apple stock).

optimizers/lightning_batch_gd.py

@@ -261,7 +261,7 @@ def __init__(
 
     def delete_run(self):
         api = wandb.Api()
-        project_path = "financial_computing_group/Limit_Order_Book"
+        project_path = "<Specify here the name of WB project>" # Specify here the name of WB project.
         runs = api.runs(path=project_path)
         print('Deleting runs...')
         while len(runs) < 1:
@@ -295,7 +295,7 @@ def train(self):
         )
         early_stopping_callback = EarlyStopping("val_loss", patience=self.patience, min_delta=0.003)
 
-        os.environ["WANDB_API_KEY"] = "" # Inser API key
+        os.environ["WANDB_API_KEY"] = ""  # Insert API key
         os.environ["WANDB__SERVICE_WAIT"] = "300"
         try:
             wandb_logger = WandbLogger(