LLM Learning Hub

workspace/llm-course/home

GRU

A GRU (Gated Recurrent Unit) is a simplified LSTM with fewer gates (update and reset), making it faster while retaining much of the long-range modeling capability. It merges the cell and hidden states.

What is GRU?

A GRU (Gated Recurrent Unit) is a simplified LSTM with fewer gates (update and reset), making it faster while retaining much of the long-range modeling capability. It merges the cell and hidden states.

A GRU (Gated Recurrent Unit) is a simplified LSTM with fewer gates (update and reset), making it faster while retaining much of the long-range modeling capability. It merges the cell and hidden states.

Where is it used?

GRUs are used in lightweight sequence models for mobile NLP, time-series forecasting, and as efficient baselines before scaling up to transformers.

How to build it

Use `nn.GRU(input_size, hidden_size, num_layers=2)`, compare parameter count and speed with an equivalent LSTM, and measure accuracy on a sequence classification task.