https://qtcovi.github.io/tag/deep-learning/Deep LearningHugo Blox Builder (https://hugoblox.com)en-usMon, 01 Jan 2024 00:00:00 +0000https://qtcovi.github.io/media/icon_hu11734318148517933569.pnghttps://qtcovi.github.io/tag/deep-learning/https://qtcovi.github.io/software/schnet4aim/Mon, 01 Jan 2024 00:00:00 +0000https://qtcovi.github.io/software/schnet4aim/<h2 id="about">About</h2> <p><strong>SchNet4AIM</strong> is a code designed to train <a href="https://doi.org/10.1063/1.5019779" target="_blank" rel="noopener">SchNet</a> deep-learning models on atomic (1-body) and pairwise (2-body) properties formulated within the Quantum Theory of Atoms in Molecules (QTAIM). It is built as a targeted modification of <a href="https://github.com/atomistic-machine-learning/schnetpack" target="_blank" rel="noopener">SchNetPack</a>, retaining only the components relevant for 1p/2p property training.</p> <p>Key features:</p> <ul> <li>Train on <strong>atomic</strong> (charges, energies, volumes) or <strong>pairwise</strong> (delocalization indices, IQA interaction energies) QTAIM properties.</li> <li>Supports <strong>JSON</strong> and <strong>ASE-SQLite</strong> database formats.</li> <li>Runs on <strong>CPU and GPU</strong> (GPU recommended for speed).</li> </ul> <h2 id="installation">Installation</h2> <div class="highlight"><pre tabindex="0" class="chroma"><code class="language-bash" data-lang="bash"><span class="line"><span class="cl">git clone https://github.com/QTCOVI/SchNet4AIM.git </span></span><span class="line"><span class="cl"><span class="nb">cd</span> SchNet4AIM </span></span><span class="line"><span class="cl">pip install -r requirements.txt </span></span></code></pre></div><h2 id="links">Links</h2> <ul> <li><a href="https://github.com/QTCOVI/SchNet4AIM" target="_blank" rel="noopener">GitHub repository</a></li> </ul>