Error in analyzing video datasets with Qwen2.5-VL-3B-Instruct

Open

#927 opened on Apr 16, 2025 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language: Python
Stars: 4.4k
Forks: 768
Avg merge: 1d 10h
Merged PRs (30d): 17

Description

I am running the below command python run.py --data Video-MME_8frame --model Qwen2.5-VL-3B-Instruct but I am getting the error:
File "VLMEvalKit_original/VLMEvalKit/vlmeval/vlm/qwen2_vl/model.py", line 192, in generate_inner
images, videos = process_vision_info([messages])
File "/usr/local/lib/python3.10/dist-packages/qwen_vl_utils/vision_process.py", line 332, in process_vision_info
video_inputs.append(fetch_video(vision_info))
File "/usr/local/lib/python3.10/dist-packages/qwen_vl_utils/vision_process.py", line 260, in fetch_video
video = VIDEO_READER_BACKENDSvideo_reader_backend
File "/usr/local/lib/python3.10/dist-packages/qwen_vl_utils/vision_process.py", line 230, in _read_video_decord
nframes = smart_nframes(ele, total_frames=total_frames, video_fps=video_fps)
File "/usr/local/lib/python3.10/dist-packages/qwen_vl_utils/vision_process.py", line 161, in smart_nframes
raise ValueError(f"nframes should in interval [{FRAME_FACTOR}, {total_frames}], but got {nframes}.")
ValueError: nframes should in interval [2, 2227], but got 0.

I am installing transformers==4.49.0 and qwen-vl-utils[decord]==0.0.8. Are there any issues with my packages?

Contributor guide

No contributing guide indexed for this repository

First steps

Read the whole issue, then the project's contributing guide.
Comment on the issue to say you are picking it up — it saves two people doing the same work.
Fork the repository and make your change on a branch.
Open a pull request that references the issue number.

Research direction

Start by reproducing python run.py --data Video-MME_8frame --model Qwen2.5-VL-3B-Instruct with the reported transformers and qwen-vl-utils versions. Read vlmeval/vlm/qwen2_vl/model.py around generate_inner and the referenced qwen_vl_utils/vision_process.py path through smart_nframes. Done means determining whether the failure is caused by package versions or video/frame handling and documenting the confirmed resolution.

Written by the indexing model from the issue text.

Assessment

Tech stack: python
Domain: computer-vision, machine-learning
Issue type: Bug
Difficulty: 3/5
Estimated time: 1-2 days
Activity status: Stale
Clarity: Mostly clear
Newbie friendliness: 35/100