github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2025-07-16 09:48:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

fixing total cost finetuned model giving zero (#5144)

Browse Source 
# OpanAI finetuned model giving zero tokens cost

Very simple fix to the previously committed solution to allowing
finetuned Openai models.

Improves #5127 

---------

Co-authored-by: Dev 2049 <dev.dev2049@gmail.com>
This commit is contained in:
Tommaso De Lorenzo 2023-05-24 19:04:08 +02:00 committed by GitHub
parent 94cf391ef1
commit 52714cedd4
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 36 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
langchain/callbacks/openai_info.py
View File

@ -26,26 +26,34 @@ MODEL_COST_PER_1K_TOKENS = {
"code-davinci-002": 0.02,
"ada-finetuned": 0.0016,
"babbage-finetuned": 0.0024,
"curie-finetuned": 0.0120,
"davinci-finetuned": 0.1200,
"curie-finetuned": 0.012,
"davinci-finetuned": 0.12,
}
def standardize_model_name(
model_name: str,
is_completion: bool = False,
) -> str:
model_name = model_name.lower()
if "ft-" in model_name:
return model_name.split(":")[0] + "-finetuned"
elif is_completion and model_name.startswith("gpt-4"):
return model_name + "-completion"
else:
return model_name
def get_openai_token_cost_for_model(
model_name: str, num_tokens: int, is_completion: bool = False
) -> float:
# handling finetuned models
if "ft-" in model_name:
model_name = f"{model_name.split(':')[0]}-finetuned"
suffix = "-completion" if is_completion and model_name.startswith("gpt-4") else ""
model = model_name.lower() + suffix
if model not in MODEL_COST_PER_1K_TOKENS:
model_name = standardize_model_name(model_name, is_completion=is_completion)
if model_name not in MODEL_COST_PER_1K_TOKENS:
raise ValueError(
f"Unknown model: {model_name}. Please provide a valid OpenAI model name."
"Known models are: " + ", ".join(MODEL_COST_PER_1K_TOKENS.keys())
)
return MODEL_COST_PER_1K_TOKENS[model] * num_tokens / 1000
return MODEL_COST_PER_1K_TOKENS[model_name] * num_tokens / 1000
class OpenAICallbackHandler(BaseCallbackHandler):
@ -91,8 +99,8 @@ class OpenAICallbackHandler(BaseCallbackHandler):
token_usage = response.llm_output["token_usage"]
completion_tokens = token_usage.get("completion_tokens", 0)
prompt_tokens = token_usage.get("prompt_tokens", 0)
model_name = response.llm_output.get("model_name")
if model_name and model_name in MODEL_COST_PER_1K_TOKENS:
model_name = standardize_model_name(response.llm_output.get("model_name", ""))
if model_name in MODEL_COST_PER_1K_TOKENS:
completion_cost = get_openai_token_cost_for_model(
model_name, completion_tokens, is_completion=True
)

16
tests/unit_tests/callbacks/test_openai_info.py
View File

@ -44,3 +44,19 @@ def test_on_llm_end_custom_model(handler: OpenAICallbackHandler) -> None:
)
handler.on_llm_end(response)
assert handler.total_cost == 0
def test_on_llm_end_finetuned_model(handler: OpenAICallbackHandler) -> None:
response = LLMResult(
generations=[],
llm_output={
"token_usage": {
"prompt_tokens": 2,
"completion_tokens": 1,
"total_tokens": 3,
},
"model_name": "ada:ft-your-org:custom-model-name-2022-02-15-04-21-04",
},
)
handler.on_llm_end(response)
assert handler.total_cost > 0