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

Qwen batch #5

Closed
wants to merge 3 commits into from
Closed

Qwen batch #5

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
897e197
change prompt and max parralel requests for qwen
HRashidi Nov 11, 2024
4afbb75
bump aana version
HRashidi Nov 15, 2024
a24232e
change async task
HRashidi Nov 27, 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
3 changes: 2 additions & 1 deletion aana_chat_with_video/configs/deployments.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -43,10 +43,11 @@
"instance": VLLMDeployment.options(
num_replicas=1,
ray_actor_options={"num_gpus": 0.25},
max_ongoing_requests=500,
user_config=VLLMConfig(
model="Qwen/Qwen2-VL-2B-Instruct",
dtype=Dtype.AUTO,
gpu_memory_reserved=12000,
gpu_memory_reserved=15000,
max_model_len=32768,
enforce_eager=True,
default_sampling_params=SamplingParams(
Expand Down
2 changes: 1 addition & 1 deletion aana_chat_with_video/endpoints/index_video.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -162,7 +162,7 @@ async def run( # noqa: C901

timestamps.extend(frames_dict["timestamps"])
frame_ids.extend(frames_dict["frame_ids"])
chat_prompt = "Describe the content of the following image in a single sentence:"
chat_prompt = "Caption the content of the following image in a single and concise sentence"
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@HRashidi Can you try to call the captioning model not in batches but by scheduling multiple asyncio tasks like here https://github.com/mobiusml/aana_episodic_retrieval/blob/1a431ed39f70c01af72ccbbb89aeb66271ce38af/aana_episodic_retrieval/endpoints/episodic_summary.py#L120? It should be much faster.

dialogs = [
ImageChatDialog.from_prompt(prompt=chat_prompt, images=[frame]) for frame in frames_dict["frames"]
]
Expand Down
Loading
Loading