Why is the top 5 results not fixed and changing every time when using the code demo_fideo_text_detrieval.ipynb for text search and video retrieval

[alanMachineLeraning]

Why is the top 5 results not fixed and changing every time when using the code demo_fideo_text_detrieval.ipynb for text search and video retrieval

[UnableToUseGit]

text: A playful dog slides down a snowy hill, wagging its tail with delight. ~ prob: 0.3967
text: A person stands on the snowy floor, pushing a sled loaded with blankets, preparing for a fun-filled ride. ~ prob: 0.3358
text: A person bundled up in a blanket walks through the snowy landscape, enjoying the serene winter scenery. ~ prob: 0.0860
text: A man in a gray sweater plays fetch with his dog in the snowy yard, throwing a toy and watching it run. ~ prob: 0.0641
text: A person in a blue jacket walks their pet on a leash, enjoying a peaceful winter walk among the trees. ~ prob: 0.0605

Are my results the same as yours?

[UnableToUseGit]

Why is the top 5 results not fixed and changing every time when using the code demo_fideo_text_detrieval.ipynb for text search and video retrieval

I also encountered this problem,

[alanMachineLeraning]

no

…

---Original--- From: ***@***.***> Date: Sat, Mar 1, 2025 09:31 AM To: ***@***.***>; Cc: ***@***.******@***.***>; Subject: Re: [OpenGVLab/InternVideo] Why is the top 5 results not fixed andchanging every time when using the code demo_fideo_text_detrieval.ipynb fortext search and video retrieval (Issue #257) text: A playful dog slides down a snowy hill, wagging its tail with delight. ~ prob: 0.3967 text: A person stands on the snowy floor, pushing a sled loaded with blankets, preparing for a fun-filled ride. ~ prob: 0.3358 text: A person bundled up in a blanket walks through the snowy landscape, enjoying the serene winter scenery. ~ prob: 0.0860 text: A man in a gray sweater plays fetch with his dog in the snowy yard, throwing a toy and watching it run. ~ prob: 0.0641 text: A person in a blue jacket walks their pet on a leash, enjoying a peaceful winter walk among the trees. ~ prob: 0.0605 Are my results the same as yours? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***> UnableToUseGit left a comment (OpenGVLab/InternVideo#257) text: A playful dog slides down a snowy hill, wagging its tail with delight. ~ prob: 0.3967 text: A person stands on the snowy floor, pushing a sled loaded with blankets, preparing for a fun-filled ride. ~ prob: 0.3358 text: A person bundled up in a blanket walks through the snowy landscape, enjoying the serene winter scenery. ~ prob: 0.0860 text: A man in a gray sweater plays fetch with his dog in the snowy yard, throwing a toy and watching it run. ~ prob: 0.0641 text: A person in a blue jacket walks their pet on a leash, enjoying a peaceful winter walk among the trees. ~ prob: 0.0605 Are my results the same as yours? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>

[UnableToUseGit]

I solved this bug. The reason was that the model weights were not loaded correctly. Make sure that the "pretrained_path" in internvideo2_stage2_config.py is correctly assigned. This point is not mentioned in the repo's DEMO_USAGE_GUIDE.

[alanMachineLeraning]

thank you very much！