Built site for gh-pages

.nojekyll

@@ -1 +1 @@
-3ace7c03
+4641f014

docs/config-reference.html

@@ -1506,8 +1506,8 @@ gtag('config', 'G-9KYCVJBNMQ', { 'anonymize_ip': true});
 <span id="cb1-724"><a href="#cb1-724" aria-hidden="true" tabindex="-1"></a><span class="fu">bfloat16</span><span class="kw">:</span><span class="at"> bool | None</span></span>
 <span id="cb1-725"><a href="#cb1-725" aria-hidden="true" tabindex="-1"></a><span class="co"># No AMP (automatic mixed precision)</span></span>
 <span id="cb1-726"><a href="#cb1-726" aria-hidden="true" tabindex="-1"></a><span class="fu">float16</span><span class="kw">:</span><span class="at"> bool | None</span></span>
-<span id="cb1-727"><a href="#cb1-727" aria-hidden="true" tabindex="-1"></a><span class="co"># Use CUDA tf32 - require &gt;=ampere</span></span>
-<span id="cb1-728"><a href="#cb1-728" aria-hidden="true" tabindex="-1"></a><span class="fu">tf32</span><span class="kw">:</span><span class="at"> bool | None</span></span>
+<span id="cb1-727"><a href="#cb1-727" aria-hidden="true" tabindex="-1"></a><span class="co"># bool to use CUDA tf32 or 'auto' for automatic detection - require &gt;=ampere</span></span>
+<span id="cb1-728"><a href="#cb1-728" aria-hidden="true" tabindex="-1"></a><span class="fu">tf32</span><span class="kw">:</span><span class="at"> Literal['auto'] | bool | None = auto</span></span>
 <span id="cb1-729"><a href="#cb1-729" aria-hidden="true" tabindex="-1"></a><span class="fu">float32</span><span class="kw">:</span><span class="at"> bool | None</span></span>
 <span id="cb1-730"><a href="#cb1-730" aria-hidden="true" tabindex="-1"></a></span>
 <span id="cb1-731"><a href="#cb1-731" aria-hidden="true" tabindex="-1"></a><span class="co"># Whether to use gradient checkpointing. Available options are: true, false, 'offload',</span></span>

index.html

@@ -809,7 +809,7 @@ gtag('config', 'G-9KYCVJBNMQ', { 'anonymize_ip': true});
 <ul>
 <li>2026/03:
 <ul>
-<li>New model support has been added in Axolotl for [<a href="https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/qwen3.5">Qwen3.5, Qwen3.5 MoE</a>, <a href="https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/glm47-flash">GLM-4.7-Flash</a>, <a href="https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/glm46v">GLM-4.6V</a>, and <a href="https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/glm45">GLM-4.5-Air</a>.</li>
+<li>New model support has been added in Axolotl for <a href="https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/mistral4">Mistral Small 4</a>, <a href="https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/qwen3.5">Qwen3.5, Qwen3.5 MoE</a>, <a href="https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/glm47-flash">GLM-4.7-Flash</a>, <a href="https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/glm46v">GLM-4.6V</a>, and <a href="https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/glm45">GLM-4.5-Air</a>.</li>
 <li><a href="https://docs.axolotl.ai/docs/expert_quantization.html">MoE expert quantization</a> support (via <code>quantize_moe_experts: true</code>) greatly reduces VRAM when training MoE models (FSDP2 compat).</li>
 </ul></li>
 <li>2026/02: