Change Llama Embedding Tutorial to use SFT by default (#14231) (#14303)

chtruong814 · suiyoubi · web-flow · commit b807e7a19d8c · 2025-08-05T08:59:48.000-05:00
Signed-off-by: Ao Tang &lt;aot@nvidia.com&gt;
Co-authored-by: Ao Tang &lt;aot@nvidia.com&gt;
diff --git a/tutorials/llm/embedding/llama_embedding.ipynb b/tutorials/llm/embedding/llama_embedding.ipynb
@@ -216,6 +216,7 @@
     "        name=\"llama32_1b_embedding_finetuning\",\n",
     "        num_nodes=nodes,\n",
     "        num_gpus_per_node=gpus_per_node,\n",
+    "        peft_scheme=None,\n",
     "    )\n",
     "    return recipe\n",
     "\n",
