Fix typos in README.md

Author: rod409

automotive/camera-3d-detection/README.md

@@ -22,13 +22,13 @@ docker run -it -v ./inference:/inference -v <path to nuscenes dataset>:/inferenc
 ## Run the model in performance mode
 ```
 cd /inference/automotive/camera-3d-detection
-python main.py --dataset nuscenes --dataset-path ./output/data --checkpoint ./data/bevformer_tiny.onnx --config ./projects/config/bevformer/bevformer_tiny.py --scene-file ./data/scene_lengths.pkl
+python main.py --dataset nuscenes --dataset-path ./output/data --checkpoint ./data/bevformer_tiny.onnx --config ./projects/configs/bevformer/bevformer_tiny.py --scene-file ./data/scene_lengths.pkl
 ```
 The dataset being mounted twice is intentional due to some hard coded paths in the original BEVFormer code. This is the solution for now until we push a fix.
 
 ## Run in accuracy model and accuracy checker
 ```
 cd /inference/automotive/camera-3d-detection
-python main.py --dataset nuscenes --dataset-path ./output/data --checkpoint ./output/data/bevformer_tiny.onnx --config projects/config/bevformer/bevformer_tiny.py --scene-file ./output/data/scene_lengths.pkl --accuracy
-python accuracy_nuscenes.py --mlperf-accuracy-file ./output/mlperf_log_accuracy.json --config projects/config/bevformer/bevformer_tiny.py --nuscenes-dir ./output/data
+python main.py --dataset nuscenes --dataset-path ./output/data --checkpoint ./output/data/bevformer_tiny.onnx --config projects/configs/bevformer/bevformer_tiny.py --scene-file ./output/data/scene_lengths.pkl --accuracy
+python accuracy_nuscenes.py --mlperf-accuracy-file ./output/mlperf_log_accuracy.json --config projects/configs/bevformer/bevformer_tiny.py --nuscenes-dir ./output/data
 ```