Rename notebook to readme

templates/e2e-dspy-workflow/e2e_dspy.ipynb → templates/e2e-dspy-workflow/README.ipynb

@@ -892,7 +892,6 @@
     "    # \"lora_dynamic_folder\": \"dspy/lora_weights/prodjob_qmulcjw4x8z599m8hkyja8tbmi/meta-llama/Llama-3.2-1B-Instruct\"\n",
     "}\n",
     "\n",
-    "\n",
     "finetuneable_lm = dspy.LM(model=\"meta-llama/Llama-3.2-1B-Instruct\")\n",
     "finetuning_job = await finetuneable_lm.finetune(method, train_path, eval_path, provider=\"anyscale\", train_kwargs=kwargs)\n",
     "model_names = []\n",
@@ -991,6 +990,29 @@
     "all_llamas = {**finetuned_llamas_1b, \"base\": llama_1b}"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# # Sanity check that the finetuned models are working\n",
+    "llama1 = list(finetuned_llamas_1b.values())[3]\n",
+    "with dspy.context(lm=llama1):\n",
+    "    print(llama1.model)\n",
+    "    test_predictor = IntentClassificationModule()\n",
+    "    sample_input = ft_trainset[11]\n",
+    "    print(sample_input)\n",
+    "    print(test_predictor(**sample_input.inputs()).label)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Now let's try optimizing the program with the finetuned model"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -1021,29 +1043,6 @@
     ")"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# # Sanity check that the finetuned models are working\n",
-    "llama1 = list(finetuned_llamas_1b.values())[3]\n",
-    "with dspy.context(lm=llama1):\n",
-    "    print(llama1.model)\n",
-    "    test_predictor = IntentClassificationModule()\n",
-    "    sample_input = ft_trainset[11]\n",
-    "    print(sample_input)\n",
-    "    print(test_predictor(**sample_input.inputs()).label)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "Now let's try optimizing the program with the finetuned model"
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -1063,7 +1062,6 @@
     "    return dspy.Example(text=data[\"example\"][\"text\"], label=data[\"prediction\"][\"label\"]).with_inputs(\"text\")\n",
     "\n",
     "collected_data_examples = [collected_data_to_example(x) for x in collected_data_filtered]\n",
-    "# collected_data_examples[0]\n",
     "\n",
     "devset_synthetic = collected_data_examples[:DEV_SIZE]\n",
     "ft_optimizer_devset = collected_data_examples[DEV_SIZE:DEV_SIZE+OPTIMIZER_NUM_VAL]\n",
@@ -1111,7 +1109,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "if False:\n",
+    "if True:\n",
     "    import json\n",
     "    with open(\"ft_results.json\", \"w\") as f:\n",
     "        json.dump(ft_results, f)\n",