The latest articles on DEV Community by manishn18 (@manishn18). https://dev.to/manishn18 https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4245023%2Fa336198e-1b88-4401-a31f-d1ca857d4e58.png https://dev.to/manishn18 en manishn18 Sat, 14 Nov 2026 19:47:55 +0000 https://dev.to/manishn18/how-i-built-a-deep-learning-library-in-pure-java-and-compared-it-with-pytorch-2j8l https://dev.to/manishn18/how-i-built-a-deep-learning-library-in-pure-java-and-compared-it-with-pytorch-2j8l <ol> <li>Why: learning Java and wanting to understand AI under the hood</li> <li>Tensors and matmul first, then gradient checks for every layer</li> <li>Building Conv2D, LSTM and a mini transformer</li> <li>Making it small: int8 and int4 quantization</li> <li>Testing: 283 tests, including autograd checks</li> <li>The PyTorch comparison: same CNNs, same epochs, accuracy within about 1 point</li> <li>What I learned and what's next Links: <a href="https://github.com/manishn1877/minidl-java/tree/main" rel="noopener noreferrer">https://github.com/manishn1877/minidl-java/tree/main</a> and <a href="https://manishn1877.github.io/minidl-java/" rel="noopener noreferrer">https://manishn1877.github.io/minidl-java/</a> </li> </ol> java deeplearning opensource architecture