biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Misc][OpenAI] deprecate max_tokens in favor of new max_completion_tokens field for chat completion endpoint (#9837)

Browse Source
This commit is contained in:
Guillaume Calmettes
2024-10-31 02:15:56 +01:00
committed by GitHub
parent 64384bbcdf
commit abbfb6134d
14 changed files with 140 additions and 118 deletions
Download Patch File Download Diff File Expand all files Collapse all files

38
tests/entrypoints/openai/test_vision.py
View File

@@ -78,11 +78,12 @@ async def test_single_chat_session_image(client: openai.AsyncOpenAI,
}]
# test single completion
chat_completion = await client.chat.completions.create(model=model_name,
messages=messages,
max_tokens=10,
logprobs=True,
top_logprobs=5)
chat_completion = await client.chat.completions.create(
model=model_name,
messages=messages,
max_completion_tokens=10,
logprobs=True,
top_logprobs=5)
assert len(chat_completion.choices) == 1
choice = chat_completion.choices[0]
@@ -101,7 +102,7 @@ async def test_single_chat_session_image(client: openai.AsyncOpenAI,
chat_completion = await client.chat.completions.create(
model=model_name,
messages=messages,
max_tokens=10,
max_completion_tokens=10,
)
message = chat_completion.choices[0].message
assert message.content is not None and len(message.content) >= 0
@@ -134,7 +135,7 @@ async def test_single_chat_session_image_beamsearch(client: openai.AsyncOpenAI,
model=model_name,
messages=messages,
n=2,
max_tokens=10,
max_completion_tokens=10,
logprobs=True,
top_logprobs=5,
extra_body=dict(use_beam_search=True))
@@ -169,11 +170,12 @@ async def test_single_chat_session_image_base64encoded(
}]
# test single completion
chat_completion = await client.chat.completions.create(model=model_name,
messages=messages,
max_tokens=10,
logprobs=True,
top_logprobs=5)
chat_completion = await client.chat.completions.create(
model=model_name,
messages=messages,
max_completion_tokens=10,
logprobs=True,
top_logprobs=5)
assert len(chat_completion.choices) == 1
choice = chat_completion.choices[0]
@@ -192,7 +194,7 @@ async def test_single_chat_session_image_base64encoded(
chat_completion = await client.chat.completions.create(
model=model_name,
messages=messages,
max_tokens=10,
max_completion_tokens=10,
)
message = chat_completion.choices[0].message
assert message.content is not None and len(message.content) >= 0
@@ -226,7 +228,7 @@ async def test_single_chat_session_image_base64encoded_beamsearch(
model=model_name,
messages=messages,
n=2,
max_tokens=10,
max_completion_tokens=10,
extra_body=dict(use_beam_search=True))
assert len(chat_completion.choices) == 2
assert chat_completion.choices[
@@ -259,7 +261,7 @@ async def test_chat_streaming_image(client: openai.AsyncOpenAI,
chat_completion = await client.chat.completions.create(
model=model_name,
messages=messages,
max_tokens=10,
max_completion_tokens=10,
temperature=0.0,
)
output = chat_completion.choices[0].message.content
@@ -269,7 +271,7 @@ async def test_chat_streaming_image(client: openai.AsyncOpenAI,
stream = await client.chat.completions.create(
model=model_name,
messages=messages,
max_tokens=10,
max_completion_tokens=10,
temperature=0.0,
stream=True,
)
@@ -320,7 +322,7 @@ async def test_multi_image_input(client: openai.AsyncOpenAI, model_name: str,
await client.chat.completions.create(
model=model_name,
messages=messages,
max_tokens=10,
max_completion_tokens=10,
temperature=0.0,
)
@@ -337,7 +339,7 @@ async def test_multi_image_input(client: openai.AsyncOpenAI, model_name: str,
chat_completion = await client.chat.completions.create(
model=model_name,
messages=messages,
max_tokens=10,
max_completion_tokens=10,
temperature=0.0,
)
message = chat_completion.choices[0].message