Bailin He


The image shows the versions of each component: ultralytics==8.2.5 torch==2.0.0 torchvision==0.18.0,it seems like torch.meshgrid has the keyword argument 'indexing',why it say unexpected?

I have also encountered this problem. Has anyone solved it？

> Hi, Can I work on this issue? Also, is there anything specific I should keep in mind while adding the installation guide and usage steps? I had the same...

``` model = init_detector(cfg, checkpoint) result = inference_detector(model, img) ``` 这里result每个类别的框数都是几百个，但是置信度并不高

似乎是配置文件中没有设置optimizer_config导致的

> This repository does not support fp16 training. mmdet.apis.train.py的182行 ` # fp16 setting fp16_cfg = cfg.get('fp16', None) if fp16_cfg is not None: optimizer_config = Fp16OptimizerHook( **cfg.optimizer_config, **fp16_cfg, distributed=distributed) elif distributed...

/tensorflow/python/training/py_checkpoint_reader.py file ``` from tensorflow.python.util._pywrap_checkpoint_reader import CheckpointReader try: return CheckpointReader(compat.as_bytes(filepattern)) ``` RuntimeError: Unsuccessful TensorSliceReader constructor: Failed to find any matching files for maskconver_seg_mnv3p5rf_coco_200k/43437096