Skip to content

Add support for Anyscale Endpoint, with llama2 70B model for now #8434

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 4 commits into from
Closed

Add support for Anyscale Endpoint, with llama2 70B model for now #8434

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
7773263
Add support for Anyscale Endpoint, with llama2 70B model for now
kylehh Jul 28, 2023
f93a1fd
check with startswith
kylehh Jul 28, 2023
fd65290
fmt
baskaryan Jul 28, 2023
c2cc879
nit
baskaryan Jul 28, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
35 changes: 27 additions & 8 deletions libs/langchain/langchain/llms/anyscale.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,8 @@
from langchain.llms.utils import enforce_stop_tokens
from langchain.utils import get_from_dict_or_env

ANYSCALE_ENDPOINT = "https://console.endpoints.anyscale.com"


class Anyscale(LLM):
"""Anyscale Service models.
Expand Down Expand Up @@ -110,14 +112,31 @@ def _call(
f"{self.anyscale_service_url}{self.anyscale_service_route}"
)
headers = {"Authorization": f"Bearer {self.anyscale_service_token}"}
body = {"prompt": prompt}
resp = requests.post(anyscale_service_endpoint, headers=headers, json=body)

if resp.status_code != 200:
raise ValueError(
f"Error returned by service, status code {resp.status_code}"
)
text = resp.text

if self.anyscale_service_url.startswith(ANYSCALE_ENDPOINT):
body = {
"model": "meta-llama/Llama-2-70b-chat-hf",
Copy link
Collaborator

@baskaryan baskaryan Jul 28, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

if this is a chat model should be implemented as ChatModel. can create a ChatAnyscale(ChatModel) class. happy to help if needed

Copy link
Contributor Author

@kylehh kylehh Jul 28, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the suggest. I will close this PR for now and will work on it later

"messages": [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": prompt},
],
"temperature": 0.7,
}
resp = requests.post(anyscale_service_endpoint, headers=headers, json=body)
if resp.status_code != 200:
raise ValueError(f"Error returned by service")
resp_json = resp.json()
text = resp_json["choices"][0]["message"]["content"]

else:
body = {"prompt": prompt}
resp = requests.post(anyscale_service_endpoint, headers=headers, json=body)

if resp.status_code != 200:
raise ValueError(
f"Error returned by service, status code {resp.status_code}"
)
text = resp.text

if stop is not None:
# This is a bit hacky, but I can't figure out a better way to enforce
Expand Down