Nlq fixes and fp16 support #53
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Adds a quickstart notebook that sets up the conda env, pre-processes annotations/clips/frames, runs an eval on a subset of VQ, and also does a small e2e training run. Note: Meta google accounts don't have an access to Colab, use a regular google account to see the notebook instead.
adding camera pose results for the val set
* edge-case handling for clip_uid as None * checking for corrupt clips + recomputing * updated TQDM description
* downscale clips to 700p before saving
* edge-case handling for clip_uid as None
* pre-extract detections for each query * can facilite faster evaluation
* replace test-challenge inference with general inference * use "WorkerWithDevice" strategy to avoid GPU overload in MP * add cached evaluation to utilize pre-computed BBoxes * add "torch.cuda.set_device(device)" in KYSTracker to assign GPU * add "visualize" flag in perform_retrieval* to save RAM
* Use deepspeed library to measure average detector flops
Add VQ2D Quickstart Colab Notebook
NLQ fixes and support FP16 features
* add annot_key * use separate dir per sample for visualization * set data version based on annot file
* read frames from video_reader lazily * sets a similarity threshold to remove noisy peaks * option to set lost_thresh for KYS
* remove end-to-end evaluate_vq2d.py * add evaluate_vq.py to use precomputed results
* account for errors where extracted clip is shorter than expected * add missing reader.close() * add option to extract specific clips
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
7 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.