+ reduction = 0.5, inchannels = 3, nclasses = 1000)</code></pre><p>Create a DenseNet model with specified configuration. Currently supported values are (121, 161, 169, 201) (<a href="https://arxiv.org/abs/1608.06993">reference</a>).</p><p><strong>Arguments</strong></p><ul><li><code>config</code>: the configuration of the model</li><li><code>pretrain</code>: whether to load the model with pre-trained weights for ImageNet.</li><li><code>growth_rate</code>: the output feature map growth probability of dense blocks (i.e. <code>k</code> in the ref)</li><li><code>reduction</code>: the factor by which the number of feature maps is scaled across each transition</li><li><code>inchannels</code>: the number of input channels</li><li><code>nclasses</code>: the number of output classes</li></ul><div class="admonition is-warning" id="Warning-2d899de5f1fbe38f"><header class="admonition-header">Warning<a class="admonition-anchor" href="#Warning-2d899de5f1fbe38f" title="Permalink"></a></header><div class="admonition-body"><p><code>DenseNet</code> does not currently support pretrained weights.</p></div></div><p>See also <a href="#Metalhead.densenet"><code>Metalhead.densenet</code></a>.</p></div><a class="docs-sourcelink" target="_blank" href="https://github.com/FluxML/Metalhead.jl/blob/1d8f8a7108153a878d3a87b76a2916e7901025ce/src/convnets/densenet.jl#L119-L139">source</a></section></details></article><h2 id="The-core-function"><a class="docs-heading-anchor" href="#The-core-function">The core function</a><a id="The-core-function-1"></a><a class="docs-heading-anchor-permalink" href="#The-core-function" title="Permalink"></a></h2><article><details class="docstring" open="true"><summary id="Metalhead.densenet"><a class="docstring-binding" href="#Metalhead.densenet"><code>Metalhead.densenet</code></a> — <span class="docstring-category">Function</span></summary><section><div><pre><code class="language-julia hljs">densenet(nblocks::AbstractVector{Int}; growth_rate = 32,
0 commit comments