Part 6/8:
- Performance Metrics: Once the setup is up and running, performance metrics like RAM and GPU utilization are monitored, highlighting the significant resources required to execute such a large model.
Results and Comparisons
After an intricate setup, observing the terminal output displaying the model's attempt to generate a Flappy Bird game serves as a tangible demonstration of its capabilities. With a token output speed averaging around 1.58 tokens per second, the performance is commendable considering the model size and local execution.