new sample for embeddings

jrcastropy · Feb 22, 2024 · a72512e · a72512e
1 parent e6ac3d5
commit a72512e
Showing 1 changed file with 12 additions and 4 deletions.
diff --git a/test.ipynb b/test.ipynb
@@ -527,18 +527,26 @@
      "text": [
       "512\n",
       "# of embeddings for the first vector: 1\n",
-      "# of embeddings for the second vector: 4\n",
+      "# of embeddings for the second vector: 8\n",
       "Results: {\n",
       "    \"result_list\": [\n",
       "        \"Black cat\",\n",
+      "        \"Black cat\",\n",
+      "        \"white Cat\",\n",
       "        \"white Cat\",\n",
       "        \"Red Cat\",\n",
+      "        \"Red Cat\",\n",
+      "        \"apple\",\n",
       "        \"apple\"\n",
       "    ],\n",
       "    \"scores_llist\": [\n",
       "        0.9999999999999999,\n",
+      "        0.9999999999999999,\n",
+      "        0.7667772992427558,\n",
       "        0.7667772992427558,\n",
       "        0.7071684587870847,\n",
+      "        0.7071684587870847,\n",
+      "        0.2702694570913964,\n",
       "        0.2702694570913964\n",
       "    ]\n",
       "}\n"
@@ -548,7 +556,7 @@
    "source": [
     "from serverless_openai import OpenAIAPI, Similarity\n",
     "from serverless_openai.helpers import cosine_similarity\n",
-    "import os\n",
+    "import os, json\n",
     "\n",
     "# Initialize OpenAIAPI with API Key\n",
     "openai_api_key = os.getenv(\"openai_api_key\")\n",
@@ -563,15 +571,15 @@
     "print(len(res1.result[0]))\n",
     "\n",
     "\n",
-    "prompt_list = [\"Black cat\", \"apple\", \"white Cat\", \"Red Cat\"]\n",
+    "prompt_list = [\"Black cat\", \"apple\", \"white Cat\", \"Red Cat\", \"Black cat\", \"apple\", \"white Cat\", \"Red Cat\"]\n",
     "res2 = ai.embeddings(prompt=prompt_list, model=model, dimensions=512)\n",
     "# matrix = [m[:512] for m in res2.result]\n",
     "if res1.result and res2.result:\n",
     "    print(\"# of embeddings for the first vector:\", len(res1.result))\n",
     "    print(\"# of embeddings for the second vector:\",len(res2.result))\n",
     "\n",
     "    sim = Similarity(vector=res1.result, matrix=res2.result)\n",
-    "    result_list = cosine_similarity(sim, prompt_list)\n",
+    "    result_list = cosine_similarity(sim, prompt_list, topn=10)\n",
     "    print(\"Results:\", json.dumps(result_list, indent=4))"
    ]
   },