a2rl.GPTBuilder.evaluate

GPTBuilder.evaluate(context_len=20, sample=True, horizon=200)

This is to evaluate the raw GPT model.

Parameters:

• context_len (int) – These is a sequence of GPT tokens

• sample (bool) – Enable sampling.

• horizon (int) – The number of GPT token to predict based on input GPT token sequence.

Return type:

Axes

Returns:

Matplotlib Axes.