Update runtime images in LLM fine-tuning with KFTO example

Author: astefanutti

examples/kfto-sft-llm/sft.ipynb

@@ -334,7 +334,7 @@
     "* Amend the resources per worker according to the job requirements\n",
     "* If you use AMD accelerators:\n",
     "  * Change `nvidia.com/gpu` to `amd.com/gpu` in `resources_per_worker`\n",
-    "  * Change `base_image` to `quay.io/modh/training:py311-rocm62-torch241`\n",
+    "  * Change `base_image` to `quay.io/modh/training:py311-rocm62-torch251`\n",
     "* Update the PVC name to the one you've attached to the workbench if needed"
    ]
   },
@@ -356,7 +356,7 @@
     "        \"memory\": \"64Gi\",\n",
     "        \"cpu\": 4,\n",
     "    },\n",
-    "    base_image=\"quay.io/modh/training:py311-cuda121-torch241\",\n",
+    "    base_image=\"quay.io/modh/training:py311-cuda124-torch251\",\n",
     "    env_vars={\n",
     "        # HuggingFace\n",
     "        \"HF_HOME\": \"/mnt/shared/.cache\",\n",