facebookresearch / facebookresearch/detectron2
LVIS1.0 model weights not found
- Dominant language
- Python
- Stars
- 34.7k
- Forks
- 7.9k
- PR merge metrics
- No merged PRs in 30d
Description
## Instructions To Reproduce the Issue:
1. Full runnable code or full changes you made:
```python
cfg = get_cfg()
modelPath="LVISv1-InstanceSegmentation/mask_rcnn_R_50_FPN_1x.yaml"
# add project-specific config (e.g., TensorMask) here if you're not running a model in detectron2's core library
cfg.merge_from_file(model_zoo.get_config_file(modelPath))
cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = 0.5 # set threshold for this model
# Find a model from detectron2's model zoo. You can use the https://dl.fbaipublicfiles... url as well
cfg.MODEL.WEIGHTS = model_zoo.get_checkpoint_url(modelPath)
predictor = DefaultPredictor(cfg)
outputs = predictor(im)
```
3. __Full logs__ you observed:
```
RuntimeError Traceback (most recent call last)
in ()
5 cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = 0.5 # set threshold for this model
6 # Find a model from detectron2's model zoo. You can use the https://dl.fbaipublicfiles... url as well
----> 7 cfg.MODEL.WEIGHTS = model_zoo.get_checkpoint_url(modelPath)
8 predictor = DefaultPredictor(cfg)
9 outputs = predictor(im)
/usr/local/lib/python3.6/dist-packages/detectron2/model_zoo/model_zoo.py in get_checkpoint_url(config_path)
98 suffix = _ModelZooUrls.CONFIG_PATH_TO_URL_SUFFIX[config_path]
99 return _ModelZooUrls.S3_PREFIX + name + "/" + suffix
--> 100 raise RuntimeError("{} not available in Model Zoo!".format(name))
101
102
RuntimeError: LVISv1-InstanceSegmentation/mask_rcnn_R_50_FPN_1x not available in Model Zoo!
```
## Environment:
Detectron2 Tutorial Colab
Do y'all plan on adding the lvis1.0 model to the model zoo? I found the config (https://github.com/facebookresearch/detectron2/tree/master/configs/LVISv1-InstanceSegmentation) but it looks like the model weights won't load.
Contributor guide
Research direction
Start in detectron2/model_zoo/model_zoo.py, especially the get_checkpoint_url path shown in the traceback, and compare its model-name mapping with configs/LVISv1-InstanceSegmentation. Re-run the supplied Colab snippet after identifying the expected LVISv1 checkpoint. Done means the listed config resolves to a valid weight URL and DefaultPredictor loads it without the reported RuntimeError.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- computer-vision
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 32/100