converted rwkv.mdx into .ipynb

docs/docs/integrations/llms/rwkv.ipynb

@@ -0,0 +1,160 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "0c47d7c4",
+   "metadata": {},
+   "source": [
+    "# RWKV\n",
+    "\n",
+    ">[RWKV](https://www.rwkv.com/) (pronounced RwaKuv) language model is an RNN \n",
+    "> with GPT-level LLM performance, \n",
+    "> and it can also be directly trained like a GPT transformer (parallelizable).\n",
+    "\n",
+    "## Overview\n",
+    "\n",
+    "RWKV is combining the best of RNN and transformer - great performance, fast inference, \n",
+    "fast training, saves VRAM, \"infinite\" ctxlen, and free text embedding. \n",
+    "Moreover, it's 100% attention-free, and a LFAI project.\n",
+    "\n",
+    "### Rwkv models recommended VRAM\n",
+    "\n",
+    "| Model | 8bit | bf16/fp16 | fp32 |\n",
+    "|-------|------|-----------|------|\n",
+    "| 14B   | 16GB | 28GB      | >50GB |\n",
+    "| 7B    | 8GB  | 14GB      | 28GB |\n",
+    "| 3B    | 2.8GB| 6GB       | 12GB |\n",
+    "| 1b5   | 1.3GB| 3GB       | 6GB |\n",
+    "\n",
+    "See the [rwkv pip](https://pypi.org/project/rwkv/) page for more information about strategies, \n",
+    "including streaming and CUDA support.\n",
+    "\n",
+    "## Setup\n",
+    "\n",
+    "- Install the Python `rwkv` and `tokenizer` packages"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "57081891",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!pip install rwkv tokenizer"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "0be7a218",
+   "metadata": {},
+   "source": [
+    "- Download a [RWKV model](https://huggingface.co/BlinkDL/rwkv-4-raven/tree/main) and place it in your desired directory\n",
+    "- Download a [tokens file](https://raw.githubusercontent.com/BlinkDL/ChatRWKV/main/20B_tokenizer.json)\n",
+    "\n",
+    "## Instantiation"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "45d32186",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain_community.llms import RWKV"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4f629ee2",
+   "metadata": {
+    "lines_to_next_cell": 0
+   },
+   "source": [
+    "To use the RWKV wrapper, you need to provide the path to the pre-trained model file and the tokenizer's configuration."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3ad21947",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model = RWKV(\n",
+    "    model=\"./models/RWKV-4-Raven-3B-v7-Eng-20230404-ctx4096.pth\",\n",
+    "    strategy=\"cpu fp32\",\n",
+    "    tokens_path=\"./rwkv/20B_tokenizer.json\"\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3538ec3e",
+   "metadata": {},
+   "source": [
+    "## Invocation"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "27003f1c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def generate_prompt(instruction, input=None):\n",
+    "    if input:\n",
+    "        return f\"\"\"Below is an instruction that describes a task, paired with an input that provides\n",
+    "further context. Write a response that appropriately completes the request.\n",
+    "\n",
+    "# Instruction:\n",
+    "{instruction}\n",
+    "\n",
+    "# Input:\n",
+    "{input}\n",
+    "\n",
+    "# Response:\n",
+    "\"\"\"\n",
+    "    else:\n",
+    "        return f\"\"\"Below is an instruction that describes a task. Write a response that \n",
+    "appropriately completes the request.\n",
+    "\n",
+    "# Instruction:\n",
+    "{instruction}\n",
+    "\n",
+    "# Response:\n",
+    "\"\"\"\n",
+    "\n",
+    "response = model.invoke(generate_prompt(\"Once upon a time, \"))"
+   ]
+  }
+ ],
+ "metadata": {
+  "jupytext": {
+   "cell_metadata_filter": "-all",
+   "main_language": "python",
+   "notebook_metadata_filter": "-all"
+  },
+  "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
+}