Comet
Comet machine learning platform integrates with your existing infrastructure and tools so you can manage, visualize, and optimize modelsβfrom training runs to production monitoring
In this guide we will demonstrate how to track your Langchain Experiments, Evaluation Metrics, and LLM Sessions with Comet.
Example Project: Comet with LangChain
Install Comet and Dependenciesβ
%pip install --upgrade --quiet comet_ml langchain langchain-openai google-search-results spacy textstat pandas
!{sys.executable} -m spacy download en_core_web_sm
Initialize Comet and Set your Credentialsβ
You can grab your Comet API Key here or click the link after initializing Comet
import comet_ml
comet_ml.init(project_name="comet-example-langchain")
Set OpenAI and SerpAPI credentialsβ
You will need an OpenAI API Key and a SerpAPI API Key to run the following examples
import os
os.environ["OPENAI_API_KEY"] = "..."
# os.environ["OPENAI_ORGANIZATION"] = "..."
os.environ["SERPAPI_API_KEY"] = "..."
Scenario 1: Using just an LLMβ
from langchain.callbacks import CometCallbackHandler, StdOutCallbackHandler
from langchain_openai import OpenAI
comet_callback = CometCallbackHandler(
project_name="comet-example-langchain",
complexity_metrics=True,
stream_logs=True,
tags=["llm"],
visualizations=["dep"],
)
callbacks = [StdOutCallbackHandler(), comet_callback]
llm = OpenAI(temperature=0.9, callbacks=callbacks, verbose=True)
llm_result = llm.generate(["Tell me a joke", "Tell me a poem", "Tell me a fact"] * 3)
print("LLM result", llm_result)
comet_callback.flush_tracker(llm, finish=True)
API Reference:
Scenario 2: Using an LLM in a Chainβ
from langchain.callbacks import CometCallbackHandler, StdOutCallbackHandler
from langchain.chains import LLMChain
from langchain_core.prompts import PromptTemplate
from langchain_openai import OpenAI
comet_callback = CometCallbackHandler(
complexity_metrics=True,
project_name="comet-example-langchain",
stream_logs=True,
tags=["synopsis-chain"],
)
callbacks = [StdOutCallbackHandler(), comet_callback]
llm = OpenAI(temperature=0.9, callbacks=callbacks)
template = """You are a playwright. Given the title of play, it is your job to write a synopsis for that title.
Title: {title}
Playwright: This is a synopsis for the above play:"""
prompt_template = PromptTemplate(input_variables=["title"], template=template)
synopsis_chain = LLMChain(llm=llm, prompt=prompt_template, callbacks=callbacks)
test_prompts = [{"title": "Documentary about Bigfoot in Paris"}]
print(synopsis_chain.apply(test_prompts))
comet_callback.flush_tracker(synopsis_chain, finish=True)
Scenario 3: Using An Agent with Toolsβ
from langchain.agents import initialize_agent, load_tools
from langchain.callbacks import CometCallbackHandler, StdOutCallbackHandler
from langchain_openai import OpenAI
comet_callback = CometCallbackHandler(
project_name="comet-example-langchain",
complexity_metrics=True,
stream_logs=True,
tags=["agent"],
)
callbacks = [StdOutCallbackHandler(), comet_callback]
llm = OpenAI(temperature=0.9, callbacks=callbacks)
tools = load_tools(["serpapi", "llm-math"], llm=llm, callbacks=callbacks)
agent = initialize_agent(
tools,
llm,
agent="zero-shot-react-description",
callbacks=callbacks,
verbose=True,
)
agent.run(
"Who is Leo DiCaprio's girlfriend? What is her current age raised to the 0.43 power?"
)
comet_callback.flush_tracker(agent, finish=True)
Scenario 4: Using Custom Evaluation Metricsβ
The CometCallbackManager
also allows you to define and use Custom Evaluation Metrics to assess generated outputs from your model. Let's take a look at how this works.
In the snippet below, we will use the ROUGE metric to evaluate the quality of a generated summary of an input prompt.
%pip install --upgrade --quiet rouge-score
from langchain.callbacks import CometCallbackHandler, StdOutCallbackHandler
from langchain.chains import LLMChain
from langchain_core.prompts import PromptTemplate
from langchain_openai import OpenAI
from rouge_score import rouge_scorer
class Rouge:
def __init__(self, reference):
self.reference = reference
self.scorer = rouge_scorer.RougeScorer(["rougeLsum"], use_stemmer=True)
def compute_metric(self, generation, prompt_idx, gen_idx):
prediction = generation.text
results = self.scorer.score(target=self.reference, prediction=prediction)
return {
"rougeLsum_score": results["rougeLsum"].fmeasure,
"reference": self.reference,
}
reference = """
The tower is 324 metres (1,063 ft) tall, about the same height as an 81-storey building.
It was the first structure to reach a height of 300 metres.
It is now taller than the Chrysler Building in New York City by 5.2 metres (17 ft)
Excluding transmitters, the Eiffel Tower is the second tallest free-standing structure in France .
"""
rouge_score = Rouge(reference=reference)
template = """Given the following article, it is your job to write a summary.
Article:
{article}
Summary: This is the summary for the above article:"""
prompt_template = PromptTemplate(input_variables=["article"], template=template)
comet_callback = CometCallbackHandler(
project_name="comet-example-langchain",
complexity_metrics=False,
stream_logs=True,
tags=["custom_metrics"],
custom_metrics=rouge_score.compute_metric,
)
callbacks = [StdOutCallbackHandler(), comet_callback]
llm = OpenAI(temperature=0.9)
synopsis_chain = LLMChain(llm=llm, prompt=prompt_template)
test_prompts = [
{
"article": """
The tower is 324 metres (1,063 ft) tall, about the same height as
an 81-storey building, and the tallest structure in Paris. Its base is square,
measuring 125 metres (410 ft) on each side.
During its construction, the Eiffel Tower surpassed the
Washington Monument to become the tallest man-made structure in the world,
a title it held for 41 years until the Chrysler Building
in New York City was finished in 1930.
It was the first structure to reach a height of 300 metres.
Due to the addition of a broadcasting aerial at the top of the tower in 1957,
it is now taller than the Chrysler Building by 5.2 metres (17 ft).
Excluding transmitters, the Eiffel Tower is the second tallest
free-standing structure in France after the Millau Viaduct.
"""
}
]
print(synopsis_chain.apply(test_prompts, callbacks=callbacks))
comet_callback.flush_tracker(synopsis_chain, finish=True)
Callback Tracerβ
There is another integration with Comet:
See an example.
from langchain.callbacks.tracers.comet import CometTracer