Skip to content

[Fix] of "evaluation of llava_vid on mvbench" #541

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

Merged
Luodian merged 7 commits into from
Feb 19, 2025
Merged

[Fix] of "evaluation of llava_vid on mvbench" #541

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
e267f62
Update README.md
zhshj0110 Feb 10, 2025
22ec659
Update README.md
zhshj0110 Feb 11, 2025
e798dd0
Merge branch 'EvolvingLMMs-Lab:main' into main
zhshj0110 Feb 12, 2025
c3c0e3d
Merge branch 'EvolvingLMMs-Lab:main' into main
zhshj0110 Feb 14, 2025
5a06467
[Fix] of "mvbench missing videos"
zhshj0110 Feb 14, 2025
23e5b39
[Fix] of "evaluation of llava_vid on mvbench"
zhshj0110 Feb 14, 2025
1dd3b1f
Fixing lmms_eval/models/llava_vid.py
zhshj0110 Feb 15, 2025
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
6 changes: 5 additions & 1 deletion lmms_eval/models/llava_vid.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,3 +1,4 @@
import glob
import math
import os
from datetime import timedelta
Expand Down Expand Up @@ -416,6 +417,8 @@ def generate_until(self, requests) -> List[str]:
visuals = doc_to_visual(self.task_dict[task][split][doc_id])
# visuals = [visuals]
# visuals = self.flatten(visuals)
if os.path.isdir(visuals[0]):
visuals = glob.glob(visuals[0] + "/*")
videos = []
try:
# for visual in visuals:
Expand All @@ -440,7 +443,8 @@ def generate_until(self, requests) -> List[str]:
frame_idx = sampled_indices.tolist()
frame_time = [i / fps for i in frame_idx]
frame_time = ",".join([f"{i:.2f}s" for i in frame_time])
video = [visuals[i] for i in frame_idx]
# video = [visuals[i] for i in frame_idx]
video = np.stack([np.array(Image.open(visuals[i])) for i in frame_idx], axis=0)

video = self._image_processor.preprocess(video, return_tensors="pt")["pixel_values"].cuda()
if self.torch_dtype == "bfloat16":
Expand Down
14 changes: 7 additions & 7 deletions lmms_eval/tasks/mvbench/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,22 +16,22 @@
from lmms_eval.tasks._task_utils.file_utils import generate_submission_file

DATA_LIST = {
"object_interaction": "star/Charades_segment",
"action_sequence": "star/Charades_segment",
"action_prediction": "star/Charades_segment",
"action_localization": "sta/sta_video_segment",
"object_interaction": "star/Charades_v1_480",
"action_sequence": "star/Charades_v1_480",
"action_prediction": "star/Charades_v1_480",
"action_localization": "sta_video",
"moving_count": "clevrer/video_validation",
"fine_grained_pose": "nturgbd_convert",
"fine_grained_pose": "nturgbd",
"character_order": "perception/videos",
"object_shuffle": "perception/videos",
"egocentric_navigation": "vlnqa",
"moving_direction": "clevrer/video_validation",
"episodic_reasoning": "tvqa/video_fps3_hq_segment",
"episodic_reasoning": "tvqa/frames_fps3_hq",
"fine_grained_action": "Moments_in_Time_Raw/videos",
"scene_transition": "scene_qa/video",
"state_change": "perception/videos",
"moving_attribute": "clevrer/video_validation",
"action_antonym": "ssv2_video_mp4",
"action_antonym": "ssv2_video",
"unexpected_action": "FunQA_test/test",
"counterfactual_inference": "clevrer/video_validation",
"object_existence": "clevrer/video_validation",
Expand Down