Training models

Getting Flowing

Our repository makes it incredibly easy to take existing CSV file(s) and train a model to forecast on them. We have some basic tutorials below with more extensive documentation underneath.

Basic Tutorial Video

Tutorial Notebook

Solar Forecasting Notebook

COVID-19 Forecasting Notebook

Setting up experiments

We take inspiration from AllenNLP and use JSON configuration files to effectively track experiments. To get started define a configuration file in the following format:

{
"model_name": string,
"model_type": string 
"model_params": {
  ""
 },
"dataset_params":
{. "type": string  
   "training_path": list 
   "validation_path": list
   "test_path": list
   "batch_size":integer,
   "forecast_history":integer,
   "forecast_length":integer 
},
"training_params":
{
   "criterion":string
   "optimizer": string
   "optim_params":{
        "lr": number
        "momentum": number 
        }
   "epochs": integer

}
"GCS":{
   "run_save_path": string,
   "project_id": string,
   "credential_path":string
   
}

"wandb": {
          "name":string,
          "tags":List[string]
         }

}

Model parameters should include all necessary parameters to initialize your model for instance if your model init function is def __init__(self, hidden_layer_dim, encoder_layers) then your model_params would be

"model_params":
   {
    "hidden_layer_dim: integer
    "encoder_layers": integer
   }

Dataset parameters include things like batch size, file paths, and anything the data loader will need. Data params also allows you define a train_end and a valid_start parameters if all your data is in the same CSV file. We also support automated methods to interpolate data.

"dataset_params":
    {  "class": "default",
       "training_path": file_path,
       "validation_path": file_path,
       "test_path": file_path,
       "batch_size":wandb_config["batch_size"],
       "forecast_history":wandb_config["forecast_history"],
       "forecast_length":wandb_config["out_seq_length"],
       "train_end": int(train_number),
       "valid_start":int(train_number+1),
       "valid_end": int(validation_number),
       "target_col": ["milk_price"],
       "relevant_cols": ["milk_price", "gallons_produced"],
       "scaler": "StandardScaler", 
       "interpolate": False
    }

Forward Params In 99% of cases this will simply be:

"forward_params":{
}

Inference Parameters

Optional parameters

• early_stopping

"early_stopping":{
"patience": 2
}

• sweep (either true or false). This must be set to true if you plan to use a Wandb sweep.
• wandb (either true or false). Set to false if you are planning on using a sweep. Only set to true if you are using Wandb without a sweep.

Defining Weights and Biases Sweep Configuration