ros/LLaMA-Factory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

support multiimage inference

Browse Source 
Former-commit-id: 8083e4607549e805eb308c4e93c8aa256202f438
This commit is contained in:
hiyouga
2024-11-01 07:25:20 +00:00
parent a63e624eca
commit 8cea5cd967
7 changed files with 103 additions and 63 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
src/llamafactory/chat/chat_model.py
View File

@@ -64,15 +64,15 @@ class ChatModel:
messages: Sequence[Dict[str, str]],
system: Optional[str] = None,
tools: Optional[str] = None,
image: Optional["ImageInput"] = None,
video: Optional["VideoInput"] = None,
images: Optional[Sequence["ImageInput"]] = None,
videos: Optional[Sequence["VideoInput"]] = None,
**input_kwargs,
) -> List["Response"]:
r"""
Gets a list of responses of the chat model.
"""
task = asyncio.run_coroutine_threadsafe(
self.achat(messages, system, tools, image, video, **input_kwargs), self._loop
self.achat(messages, system, tools, images, videos, **input_kwargs), self._loop
)
return task.result()
@@ -81,28 +81,28 @@ class ChatModel:
messages: Sequence[Dict[str, str]],
system: Optional[str] = None,
tools: Optional[str] = None,
image: Optional["ImageInput"] = None,
video: Optional["VideoInput"] = None,
images: Optional[Sequence["ImageInput"]] = None,
videos: Optional[Sequence["VideoInput"]] = None,
**input_kwargs,
) -> List["Response"]:
r"""
Asynchronously gets a list of responses of the chat model.
"""
return await self.engine.chat(messages, system, tools, image, video, **input_kwargs)
return await self.engine.chat(messages, system, tools, images, videos, **input_kwargs)
def stream_chat(
self,
messages: Sequence[Dict[str, str]],
system: Optional[str] = None,
tools: Optional[str] = None,
image: Optional["ImageInput"] = None,
video: Optional["VideoInput"] = None,
images: Optional[Sequence["ImageInput"]] = None,
videos: Optional[Sequence["VideoInput"]] = None,
**input_kwargs,
) -> Generator[str, None, None]:
r"""
Gets the response token-by-token of the chat model.
"""
generator = self.astream_chat(messages, system, tools, image, video, **input_kwargs)
generator = self.astream_chat(messages, system, tools, images, videos, **input_kwargs)
while True:
try:
task = asyncio.run_coroutine_threadsafe(generator.__anext__(), self._loop)
@@ -115,14 +115,14 @@ class ChatModel:
messages: Sequence[Dict[str, str]],
system: Optional[str] = None,
tools: Optional[str] = None,
image: Optional["ImageInput"] = None,
video: Optional["VideoInput"] = None,
images: Optional[Sequence["ImageInput"]] = None,
videos: Optional[Sequence["VideoInput"]] = None,
**input_kwargs,
) -> AsyncGenerator[str, None]:
r"""
Asynchronously gets the response token-by-token of the chat model.
"""
async for new_token in self.engine.stream_chat(messages, system, tools, image, video, **input_kwargs):
async for new_token in self.engine.stream_chat(messages, system, tools, images, videos, **input_kwargs):
yield new_token
def get_scores(