The latest articles on DEV Community by Filippo Lollato (@xupremix). https://dev.to/xupremix 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%2F4094863%2F0a0b346b-6733-476a-a868-a5891a5836fb.jpg https://dev.to/xupremix en Filippo Lollato Wed, 26 Aug 2026 01:07:13 +0000 https://dev.to/xupremix/incin-a-rust-machine-learning-framework-for-setting-fire-to-dimensionality-bugs-55h8 https://dev.to/xupremix/incin-a-rust-machine-learning-framework-for-setting-fire-to-dimensionality-bugs-55h8 <p><a href="https://github.com/xupremix/incin" rel="noopener noreferrer">Incin</a> is a deep learning framework in Rust where a tensor’s shape, dtype, device and gradient state all live in its type. Shape, dtype, device mismatches and so on are compiler errors.<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight rust"><code><span class="k">use</span> <span class="nn">incin</span><span class="p">::</span><span class="nn">prelude</span><span class="p">::</span><span class="o">*</span><span class="p">;</span> <span class="err">
</span><span class="k">let</span> <span class="n">x</span> <span class="o">=</span> <span class="n">Cpu</span><span class="nf">.randn</span><span class="p">(</span><span class="nd">shape!</span><span class="p">[</span><span class="mi">4</span><span class="p">,</span> <span class="mi">8</span><span class="p">])</span><span class="o">?</span><span class="p">;</span> <span class="k">let</span> <span class="n">w</span> <span class="o">=</span> <span class="n">Cpu</span><span class="nf">.randn</span><span class="p">(</span><span class="nd">shape!</span><span class="p">[</span><span class="mi">8</span><span class="p">,</span> <span class="mi">2</span><span class="p">])</span><span class="o">?</span><span class="p">;</span> <span class="k">let</span> <span class="n">y</span> <span class="o">=</span> <span class="n">x</span><span class="nf">.matmul</span><span class="p">(</span><span class="o">&amp;</span><span class="n">w</span><span class="p">)</span><span class="o">?</span><span class="p">;</span> <span class="c1">// [4, 8] x [8, 2] -&gt; [4, 2]</span> <span class="err">
</span><span class="k">let</span> <span class="n">bad</span> <span class="o">=</span> <span class="n">Cpu</span><span class="nf">.randn</span><span class="p">(</span><span class="nd">shape!</span><span class="p">[</span><span class="mi">3</span><span class="p">,</span> <span class="mi">8</span><span class="p">])</span><span class="o">?</span><span class="p">;</span> <span class="k">let</span> <span class="n">_</span> <span class="o">=</span> <span class="n">x</span><span class="nf">.matmul</span><span class="p">(</span><span class="o">&amp;</span><span class="n">bad</span><span class="p">)</span><span class="o">?</span><span class="p">;</span> <span class="c1">// inner dims 8 and 3: does not compile</span> </code></pre> </div> <p>The main goal was to find out how much of the tensor contract the type system can genuinely carry, how flexible we can make it, and how pleasant to work with it can be.<br> If you're interested on the features, architecture or anything else:</p> <ul> <li><a href="https://crates.io/crates/incin" rel="noopener noreferrer">Crate</a></li> <li><a href="https://docs.rs/incin/latest/incin" rel="noopener noreferrer">Docs</a></li> <li><a href="https://xupremix.github.io/incin" rel="noopener noreferrer">Book</a></li> </ul> deeplearning machinelearning opensource rust