Add notebook with instructions to host fondant-datacomp-small index

rom1504 · Mar 15, 2024 · fc6e9f0 · fc6e9f0
1 parent ee0931f
commit fc6e9f0
Showing 1 changed file with 195 additions and 0 deletions.
diff --git a/docs/datacomp12M_back.ipynb b/docs/datacomp12M_back.ipynb
@@ -0,0 +1,195 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "d8728dbd-b542-4a4b-b3db-6cc8c9a18a36",
+   "metadata": {},
+   "source": [
+    "# Run clip-retrieval back with fondant-ai/datacomp-small-clip index"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5708ba35-9ed3-4492-96aa-cd07f31cb8c0",
+   "metadata": {},
+   "source": [
+    "### Create virtual environment"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "0f5908fa-fe50-432a-8475-e926d4e566be",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!python3 -m venv .env\n",
+    "!source .env/bin/activate"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1812814f-6193-4d19-abd8-8a64c4c714d0",
+   "metadata": {},
+   "source": [
+    "## Download index and metadata"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b0893795-2670-4ea4-9035-83eb2709ebfc",
+   "metadata": {},
+   "source": [
+    "### Install requirements"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1cdfd076-f2a4-4809-aaf6-44868662e6ec",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!pip install dask[dataframe] huggingface_hub"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "41bd6c70-fbec-4e83-93bd-d2887ab48e99",
+   "metadata": {},
+   "source": [
+    "### Create the index folder"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "fec09d8d-3aa7-43ea-b356-db93ecd5384c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!mkdir datacomp_small"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "64117004-eb0e-4b07-a0f0-a7fffb9552d8",
+   "metadata": {},
+   "source": [
+    "### Download the index"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b4806e99-805f-4d25-b69d-0799fee7bf76",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!wget -O datacomp_small/image.index https://huggingface.co/datasets/fondant-ai/datacomp-small-clip/resolve/main/faiss?download=true -q --show-progress"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "429ea232-95da-4c48-898f-3170a1bc74e4",
+   "metadata": {},
+   "source": [
+    "### Download the metadata"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "61d44529-5609-4612-852c-d18cf4560075",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import dask.dataframe as dd\n",
+    "from dask.diagnostics import ProgressBar\n",
+    "\n",
+    "ddf = dd.read_parquet(\"hf://datasets/fondant-ai/datacomp-small-clip/id_mapping\")\n",
+    "ddf = ddf.rename(columns={\"image_path\": \"url\"})\n",
+    "ddf = ddf.repartition(npartitions=1)\n",
+    "\n",
+    "with ProgressBar():\n",
+    "    ddf.to_parquet(\"datacomp_small/metadata\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "872de5af-5c66-4249-a5db-aa028f5bca58",
+   "metadata": {},
+   "source": [
+    "## Run clip-retrieval backend"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "302b6927-45b8-4977-926e-8d3f709d6e60",
+   "metadata": {},
+   "source": [
+    "### Install requirements"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b4d1729f-4812-4a7d-9945-16615ecbd51f",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!pip install clip-retrieval"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "580cc9e6-1a2b-4432-98ac-c1e20664bfe1",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "%%writefile indices.json\n",
+    "{\n",
+    "    \"fondant_datacomp_small\": {\n",
+    "        \"indice_folder\": \"datacomp_small\",\n",
+    "        \"columns_to_return\": [\"url\"],\n",
+    "        \"clip_model\": \"open_clip:ViT-B-32/laion2b_s34b_b79k\",\n",
+    "        \"enable_mclip_option\": false,\n",
+    "        \"provide_aesthetic_embeddings\": false\n",
+    "    }\n",
+    "}"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9fbf3533-3864-4abb-86a0-d5816e4deb14",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!clip-retrieval back --port 1234 --indices-paths indices.json --clip_model open_clip:ViT-B-32/laion2b_s34b_b79k"
+   ]
+  }
+ ],
+ "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": "3.10.12"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}