{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Baichuan Text Embeddings\n",
    "\n",
    "As of today (Jan 25th, 2024) BaichuanTextEmbeddings ranks #1 in C-MTEB (Chinese Multi-Task Embedding Benchmark) leaderboard.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Leaderboard (Under Overall -> Chinese section): https://huggingface.co/spaces/mteb/leaderboard"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Official Website: https://platform.baichuan-ai.com/docs/text-Embedding\n",
    "\n",
    "An API key is required to use this embedding model. You can get one by registering at https://platform.baichuan-ai.com/docs/text-Embedding."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "BaichuanTextEmbeddings support 512 token window and preduces vectors with 1024 dimensions. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Please NOTE that BaichuanTextEmbeddings only supports Chinese text embedding. Multi-language support is coming soon."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain_community.embeddings import BaichuanTextEmbeddings\n",
    "\n",
    "embeddings = BaichuanTextEmbeddings(baichuan_api_key=\"sk-*\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Alternatively, you can set API key this way:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "\n",
    "os.environ[\"BAICHUAN_API_KEY\"] = \"YOUR_API_KEY\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "text_1 = \"今天天气不错\"\n",
    "text_2 = \"今天阳光很好\"\n",
    "\n",
    "query_result = embeddings.embed_query(text_1)\n",
    "query_result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "doc_result = embeddings.embed_documents([text_1, text_2])\n",
    "doc_result"
   ]
  }
 ],
 "metadata": {
  "language_info": {
   "name": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}