Skip to content
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

Jump to bottom

Update the format model list #21

Merged
merged 2 commits into from
Aug 30, 2024
Merged

Update the format model list #21

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
51d114e
refactor: Update GPT deployment ID in gpt.py
snakedye Aug 29, 2024
af29ca8
refactor: Update GPT deployment ID in gpt.py
snakedye Aug 30, 2024
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
38 changes: 19 additions & 19 deletions pipeline/gpt.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,10 +3,18 @@
import dspy.adapters
import dspy.utils

MODELS_WITH_RESPONSE_FORMAT = [
"ailab-llm",
"ailab-llm-gpt-4o"
] # List of models that support the response_format option
SUPPORTED_MODELS = {
"gpt-3.5-turbo": {
"max_token": 12000,
"api_version": "2024-02-01",
"response_format": { "type": "json_object" },
},
"gpt-4o": {
"max_token": 4096,
"api_version": "2024-02-15-preview",
"response_format": { "type": "json_object" },
}
}

SPECIFICATION = """
Keys:
Expand Down Expand Up @@ -64,27 +72,19 @@ def __init__(self, api_endpoint, api_key, deployment_id):
if not api_endpoint or not api_key or not deployment_id:
raise ValueError("The API endpoint, key and deployment_id are required to instantiate the GPT class.")

response_format = None
if deployment_id in MODELS_WITH_RESPONSE_FORMAT:
response_format = { "type": "json_object" }

max_token = 12000
api_version = "2024-02-01"
if deployment_id == MODELS_WITH_RESPONSE_FORMAT[0]:
max_token = 3500
elif deployment_id == MODELS_WITH_RESPONSE_FORMAT[1]:
max_token = 4096
api_version="2024-02-15-preview"

config = SUPPORTED_MODELS.get(deployment_id)
if not config:
raise ValueError(f"The deployment_id {deployment_id} is not supported.")

self.dspy_client = dspy.AzureOpenAI(
user="fertiscan",
api_base=api_endpoint,
api_key=api_key,
deployment_id=deployment_id,
# model_type='text',
api_version=api_version,
max_tokens=max_token,
response_format=response_format,
api_version=config.get("api_version"),
max_tokens=config.get("max_token"),
response_format=config.get("response_format"),
)

def create_inspection(self, prompt) -> Prediction:
Expand Down
Loading