langchain/docs/modules/models/llms/examples/streaming_llm.ipynb
Ankush Gola d3ec00b566
Callbacks Refactor [base] (#3256)
Co-authored-by: Nuno Campos <nuno@boringbits.io>
Co-authored-by: Davis Chase <130488702+dev2049@users.noreply.github.com>
Co-authored-by: Zander Chase <130414180+vowelparrot@users.noreply.github.com>
Co-authored-by: Harrison Chase <hw.chase.17@gmail.com>
2023-04-30 11:14:09 -07:00

274 lines
7.9 KiB
Plaintext

{
"cells": [
{
"cell_type": "markdown",
"id": "6eaf7e66-f49c-42da-8d11-22ea13bef718",
"metadata": {},
"source": [
"# How to stream LLM and Chat Model responses\n",
"\n",
"LangChain provides streaming support for LLMs. Currently, we support streaming for the `OpenAI`, `ChatOpenAI`. and `Anthropic` implementations, but streaming support for other LLM implementations is on the roadmap. To utilize streaming, use a [`CallbackHandler`](https://github.com/hwchase17/langchain/blob/master/langchain/callbacks/base.py) that implements `on_llm_new_token`. In this example, we are using [`StreamingStdOutCallbackHandler`]()."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "4ac0ff54-540a-4f2b-8d9a-b590fec7fe07",
"metadata": {
"tags": []
},
"outputs": [],
"source": [
"from langchain.llms import OpenAI, Anthropic\n",
"from langchain.chat_models import ChatOpenAI\n",
"from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler\n",
"from langchain.schema import HumanMessage"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "77f60a4b-f786-41f2-972e-e5bb8a48dcd5",
"metadata": {
"tags": []
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"\n",
"Verse 1\n",
"I'm sippin' on sparkling water,\n",
"It's so refreshing and light,\n",
"It's the perfect way to quench my thirst\n",
"On a hot summer night.\n",
"\n",
"Chorus\n",
"Sparkling water, sparkling water,\n",
"It's the best way to stay hydrated,\n",
"It's so crisp and so clean,\n",
"It's the perfect way to stay refreshed.\n",
"\n",
"Verse 2\n",
"I'm sippin' on sparkling water,\n",
"It's so bubbly and bright,\n",
"It's the perfect way to cool me down\n",
"On a hot summer night.\n",
"\n",
"Chorus\n",
"Sparkling water, sparkling water,\n",
"It's the best way to stay hydrated,\n",
"It's so crisp and so clean,\n",
"It's the perfect way to stay refreshed.\n",
"\n",
"Verse 3\n",
"I'm sippin' on sparkling water,\n",
"It's so light and so clear,\n",
"It's the perfect way to keep me cool\n",
"On a hot summer night.\n",
"\n",
"Chorus\n",
"Sparkling water, sparkling water,\n",
"It's the best way to stay hydrated,\n",
"It's so crisp and so clean,\n",
"It's the perfect way to stay refreshed."
]
}
],
"source": [
"llm = OpenAI(streaming=True, callbacks=[StreamingStdOutCallbackHandler()], temperature=0)\n",
"resp = llm(\"Write me a song about sparkling water.\")"
]
},
{
"cell_type": "markdown",
"id": "61fb6de7-c6c8-48d0-a48e-1204c027a23c",
"metadata": {
"tags": []
},
"source": [
"We still have access to the end `LLMResult` if using `generate`. However, `token_usage` is not currently supported for streaming."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "a35373f1-9ee6-4753-a343-5aee749b8527",
"metadata": {
"tags": []
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"\n",
"Q: What did the fish say when it hit the wall?\n",
"A: Dam!"
]
},
{
"data": {
"text/plain": [
"LLMResult(generations=[[Generation(text='\\n\\nQ: What did the fish say when it hit the wall?\\nA: Dam!', generation_info={'finish_reason': None, 'logprobs': None})]], llm_output={'token_usage': {}, 'model_name': 'text-davinci-003'})"
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"llm.generate([\"Tell me a joke.\"])"
]
},
{
"cell_type": "markdown",
"id": "a93a4d61-0476-49db-8321-7de92bd74059",
"metadata": {},
"source": [
"Here's an example with the `ChatOpenAI` chat model implementation:"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "22665f16-e05b-473c-a4bd-ad75744ea024",
"metadata": {
"tags": []
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Verse 1:\n",
"Bubbles rising to the top\n",
"A refreshing drink that never stops\n",
"Clear and crisp, it's oh so pure\n",
"Sparkling water, I can't ignore\n",
"\n",
"Chorus:\n",
"Sparkling water, oh how you shine\n",
"A taste so clean, it's simply divine\n",
"You quench my thirst, you make me feel alive\n",
"Sparkling water, you're my favorite vibe\n",
"\n",
"Verse 2:\n",
"No sugar, no calories, just H2O\n",
"A drink that's good for me, don't you know\n",
"With lemon or lime, you're even better\n",
"Sparkling water, you're my forever\n",
"\n",
"Chorus:\n",
"Sparkling water, oh how you shine\n",
"A taste so clean, it's simply divine\n",
"You quench my thirst, you make me feel alive\n",
"Sparkling water, you're my favorite vibe\n",
"\n",
"Bridge:\n",
"You're my go-to drink, day or night\n",
"You make me feel so light\n",
"I'll never give you up, you're my true love\n",
"Sparkling water, you're sent from above\n",
"\n",
"Chorus:\n",
"Sparkling water, oh how you shine\n",
"A taste so clean, it's simply divine\n",
"You quench my thirst, you make me feel alive\n",
"Sparkling water, you're my favorite vibe\n",
"\n",
"Outro:\n",
"Sparkling water, you're the one for me\n",
"I'll never let you go, can't you see\n",
"You're my drink of choice, forevermore\n",
"Sparkling water, I adore."
]
}
],
"source": [
"chat = ChatOpenAI(streaming=True, callbacks=[StreamingStdOutCallbackHandler()], temperature=0)\n",
"resp = chat([HumanMessage(content=\"Write me a song about sparkling water.\")])"
]
},
{
"cell_type": "markdown",
"id": "909ae48b-0f07-4990-bbff-e627f706c93e",
"metadata": {},
"source": [
"Here is an example with the `Anthropic` LLM implementation, which uses their `claude` model."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "eadae4ba-9f21-4ec8-845d-dd43b0edc2dc",
"metadata": {
"tags": []
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"Sparkling water, bubbles so bright,\n",
"\n",
"Fizzing and popping in the light.\n",
"\n",
"No sugar or calories, a healthy delight,\n",
"\n",
"Sparkling water, refreshing and light.\n",
"\n",
"Carbonation that tickles the tongue,\n",
"\n",
"In flavors of lemon and lime unsung.\n",
"\n",
"Sparkling water, a drink quite all right,\n",
"\n",
"Bubbles sparkling in the light."
]
},
{
"data": {
"text/plain": [
"'\\nSparkling water, bubbles so bright,\\n\\nFizzing and popping in the light.\\n\\nNo sugar or calories, a healthy delight,\\n\\nSparkling water, refreshing and light.\\n\\nCarbonation that tickles the tongue,\\n\\nIn flavors of lemon and lime unsung.\\n\\nSparkling water, a drink quite all right,\\n\\nBubbles sparkling in the light.'"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"llm = Anthropic(streaming=True, callbacks=[StreamingStdOutCallbackHandler()], temperature=0)\n",
"llm(\"Write me a song about sparkling water.\")"
]
}
],
"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": "3.10.9"
}
},
"nbformat": 4,
"nbformat_minor": 5
}