update demo example

demo.ipynb

@@ -193,16 +193,24 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 12,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "==> prediciton:  There is one dog in the picture.\n"
+     ]
+    }
+   ],
    "source": [
     "import requests\n",
     "from PIL import Image\n",
     "img_url = 'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg'\n",
-    "raw_image = Image.open(requests.get(img_url, stream=True).raw).convert('RGB')\n",
+    "raw_image = Image.open(requests.get(img_url, stream=True).raw)\n",
     "query = \"how many dogs are in the picture?\"\n",
-    "inputs = image_processor([img], return_tensors=\"pt\", image_aspect_ratio='anyres')\n",
+    "inputs = image_processor([raw_image], return_tensors=\"pt\", image_aspect_ratio='anyres')\n",
     "prompt = apply_prompt_template(query)\n",
     "language_inputs = tokenizer([prompt], return_tensors=\"pt\")\n",
     "inputs.update(language_inputs)\n",
@@ -216,6 +224,13 @@
     "print(\"==> prediciton: \", prediction)\n",
     "# output: ==> prediciton: There is one dog in the picture."
    ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
   }
  ],
  "metadata": {