{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "dd03eb44",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "token\n",
      "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\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/david/Documents/python-env-test/venv/lib/python3.10/site-packages/huggingface_hub/hf_api.py:101: FutureWarning: `name` and `organization` input arguments are deprecated and will be removed in v0.10. Pass `repo_id` instead.\n",
      "  warnings.warn(\n",
      "Cloning https://huggingface.co/datasets/HuggingDavid/simple-mnist-flagging into local empty directory.\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "38d85f20bb7d48f8934048f520b5125f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Download file img/tmp7qxdqjtl.png:  46%|####5     | 8.28k/18.1k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ea599ef7307c42c7b1f3db8e453aadc9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Clean file img/tmp7qxdqjtl.png:   6%|5         | 1.00k/18.1k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4badbc924d4a4d04b5469682a6837c9a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Download file img/tmpb9pmlzsj.png: 100%|##########| 15.4k/15.4k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "665fb3a23cc843cba87cdaad930b645a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Clean file img/tmpb9pmlzsj.png:   7%|6         | 1.00k/15.4k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Running on local URL:  http://127.0.0.1:7880\n",
      "\n",
      "To create a public link, set `share=True` in `launch()`.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div><iframe src=\"http://127.0.0.1:7880/\" width=\"900\" height=\"500\" allow=\"autoplay; camera; microphone; clipboard-read; clipboard-write;\" frameborder=\"0\" allowfullscreen></iframe></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(<gradio.routes.App at 0x162231e40>, 'http://127.0.0.1:7880/', None)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2ecf20840bb14b4f96671ee323d83734",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Upload file img/tmpjuysmmri.png: 100%|##########| 17.6k/17.6k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "remote: Scanning LFS files for validity, may be slow...        \n",
      "remote: LFS file scan complete.        \n",
      "To https://huggingface.co/datasets/HuggingDavid/simple-mnist-flagging\n",
      "   4b19b7d..458cf22  main -> main\n",
      "\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "import gradio as gr\n",
    "from torchvision import transforms\n",
    "from PIL import ImageOps\n",
    "import os\n",
    "from dotenv import load_dotenv\n",
    "\n",
    "load_dotenv()\n",
    "\n",
    "hf_writer = gr.HuggingFaceDatasetSaver(os.getenv('HF_TOKEN'), \"simple-mnist-flagging\")\n",
    "\n",
    "def load_model():\n",
    "    model_dict = torch.load('linear_model.pt')\n",
    "    return model_dict\n",
    "\n",
    "model = load_model()\n",
    "convert_tensor = transforms.ToTensor()\n",
    "\n",
    "def predict(img):\n",
    "    img =  ImageOps.grayscale(img).resize((28,28))\n",
    "    image_tensor = convert_tensor(img).view(28*28)\n",
    "    res = image_tensor @ model['weights'] + model['bias']\n",
    "    res = res.sigmoid()\n",
    "    return {\"It's 3\": float(res), \"It's 7\": float(1-res)}\n",
    "\n",
    "title = \"Is it 7 or 3\"\n",
    "description = '<p><center>Write a number, 7 or 3, in the middle.</center></p>'\n",
    "\n",
    "gr.Interface(fn=predict, \n",
    "             inputs=gr.Paint(type=\"pil\", invert_colors=True),\n",
    "             outputs=gr.Label(num_top_classes=2),\n",
    "             title=title,\n",
    "             flagging_options=[\"incorrect\",\"ambiguous\"],\n",
    "             flagging_callback=hf_writer,\n",
    "             description=description,\n",
    "             allow_flagging='manual').launch()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}