Tiny GPT in Go. Optimised for Understanding. Trained on Jules Verne Books

A developer released a minimal GPT implementation written entirely in Go, trained on Jules Verne novels. The model generates short text fragments like "Mysterious Island" and takes about 40 minutes to train on an M3 MacBook Air. The project prioritizes educational clarity over performance, removing batch dimensions and external dependencies to serve as a companion to Karpathy's "Neural Networks: Zero to Hero" course.

Simple GPT implementation in pure Go. Trained on favourite Jules Verne books. What kind of response you can expect from the model: Mysterious Island. Well. My days must follow Or this: Captain Nemo, in two hundred thousand feet weary in the existence of the world. bash $ go run . It takes about 40 minutes to train on MacBook Air M3. The trained weights will be saved to model-1.234M file. If you rerun the model, it will pick up the saved weights and continue training. The loss should decrease each time, indicating that the model is learning something useful. You can train on your own dataset by pointing the data.dataset variable to your text corpus. To run in chat-only mode once the training is done: bash $ go run . -chat You can use this repository as a companion to the Neural Networks: Zero to Hero https://karpathy.ai/zero-to-hero.html course. Use git checkout <tag to see how the model has evolved over time: naive , bigram , multihead , block , residual , full . In main test.go https://github.com/zakirullin/gpt-go/blob/main/main test.go you will find explanations starting from basic neuron example: // Our neuron has 2 inputs and 1 output number of columns in weight matrix . // Its goal is to predict next number in the sequence. input := V{1, 2} // {x1, x2} weight := M{ {2}, // how much x1 contributes to the output {3}, // how much x2 contributes to the output } All the way to self-attention mechanism: // To calculate the sum of all previous tokens, we can multiply by this triangular matrix: tril := M{ {1, 0, 0, 0}, // first token attends only at itself "cat" , it can't look into the future {1, 1, 0, 0}, // second token attends at itself and the previous token "cat" + ", " {1, 1, 1, 0}, // third token attends at itself and the two previous tokens "cat" + ", " + "dog" {1, 1, 1, 1}, // fourth token attends at itself and all the previous tokens "cat" + ", " + "dog" + " and" }.Var // So, at this point each embedding is enriched with the information from all the previous tokens. // That's the crux of self-attention. enrichedEmbeds := MatMul tril, inputEmbeds No batches. I've given up the complexity of the batch dimension for the sake of better understanding. It's far easier to build intuition with 2D matrices, rather than with 3D tensors. Besides, batches aren't inherent to the transformer architecture. For better gradient smoothing gradient accumulation was tried. The effect was negligible, so it was removed as well. Removed gonum . The gonum.matmul gave us ~30% performance boost, but it brought additional dependency. We're not striving for maximum efficiency here, rather for radical simplicity. Current matmul implementation is quite effective, and it's only 40 lines of plain readable code. You don't need to read them to understand the code : Attention Is All You Need https://arxiv.org/abs/1706.03762 Deep Residual Learning https://arxiv.org/abs/1512.03385 DeepMind WaveNet https://arxiv.org/abs/1609.03499 Batch Normalization https://arxiv.org/abs/1502.03167 Deep NN + huge data = breakthrough performance https://papers.nips.cc/paper files/paper/2012/hash/c399862d3b9d6b76c8436e924a68c45b-Abstract.html OpenAI GPT-3 paper https://arxiv.org/abs/2005.14165 Analyzing the Structure of Attention https://arxiv.org/abs/1906.04284 Many thanks to Andrej Karpathy https://github.com/karpathy for his brilliant Neural Networks: Zero to Hero https://karpathy.ai/zero-to-hero.html course.