phoenix/litellm-mirror
0
0
Fork 1
mirror of https://github.com/BerriAI/litellm.git synced 2025-04-25 18:54:30 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Reverted error while refactoring

Browse source
This commit is contained in:
Duc Pham 2023-11-10 01:35:41 +07:00
parent 07e8cf1d9a
commit eeac3954d5
2 changed files with 5 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

4
litellm/tests/test_utils.py
View file

@ -38,7 +38,7 @@ def test_multiple_messages_trimming():
{"role": "user", "content": "This is a long message that will exceed the token limit."}, {"role": "user", "content": "This is a long message that will exceed the token limit."},
{"role": "user", "content": "This is another long message that will also exceed the limit."} {"role": "user", "content": "This is another long message that will also exceed the limit."}
] ]
trimmed_messages = trim_messages(messages_copy=messages, model="gpt-3.5-turbo", max_tokens=20) trimmed_messages = trim_messages(messages=messages, model="gpt-3.5-turbo", max_tokens=20)
# print(get_token_count(messages=trimmed_messages, model="gpt-3.5-turbo")) # print(get_token_count(messages=trimmed_messages, model="gpt-3.5-turbo"))
assert(get_token_count(messages=trimmed_messages, model="gpt-3.5-turbo")) <= 20 assert(get_token_count(messages=trimmed_messages, model="gpt-3.5-turbo")) <= 20
# test_multiple_messages_trimming() # test_multiple_messages_trimming()
@ -48,7 +48,7 @@ def test_multiple_messages_no_trimming():
{"role": "user", "content": "This is a long message that will exceed the token limit."}, {"role": "user", "content": "This is a long message that will exceed the token limit."},
{"role": "user", "content": "This is another long message that will also exceed the limit."} {"role": "user", "content": "This is another long message that will also exceed the limit."}
] ]
trimmed_messages = trim_messages(messages_copy=messages, model="gpt-3.5-turbo", max_tokens=100) trimmed_messages = trim_messages(messages=messages, model="gpt-3.5-turbo", max_tokens=100)
print("Trimmed messages") print("Trimmed messages")
print(trimmed_messages) print(trimmed_messages)
assert(messages==trimmed_messages) assert(messages==trimmed_messages)

6
litellm/utils.py
View file

@ -4477,7 +4477,7 @@ def completion_with_config(config: Union[dict, str], **kwargs):
except: except:
continue continue
if prompt_larger_than_model: if prompt_larger_than_model:
messages = trim_messages(messages_copy=messages, model=max_model) messages = trim_messages(messages=messages, model=max_model)
kwargs["messages"] = messages kwargs["messages"] = messages
kwargs["model"] = model kwargs["model"] = model
@ -4666,7 +4666,7 @@ def shorten_message_to_fit_limit(
# this code is borrowed from https://github.com/KillianLucas/tokentrim/blob/main/tokentrim/tokentrim.py # this code is borrowed from https://github.com/KillianLucas/tokentrim/blob/main/tokentrim/tokentrim.py
# Credits for this code go to Killian Lucas # Credits for this code go to Killian Lucas
def trim_messages( def trim_messages(
messages_copy, messages,
model: Optional[str] = None, model: Optional[str] = None,
trim_ratio: float = 0.75, trim_ratio: float = 0.75,
return_response_tokens: bool = False, return_response_tokens: bool = False,
@ -4687,7 +4687,7 @@ def trim_messages(
""" """
# Initialize max_tokens # Initialize max_tokens
# if users pass in max tokens, trim to this amount # if users pass in max tokens, trim to this amount
messages_copy = copy.deepcopy(messages_copy) messages_copy = copy.deepcopy(messages)
try: try:
print_verbose(f"trimming messages") print_verbose(f"trimming messages")
if max_tokens == None: if max_tokens == None: