Release training configurations for fusion models (#257)

* [Major] Add fusion model configs.

* [Minor] Update training instructions.

README.md

@@ -183,6 +183,18 @@ For LiDAR-only BEV segmentation model, please run:
 torchpack dist-run -np 8 python tools/train.py configs/nuscenes/seg/lidar-centerpoint-bev128.yaml
 ```
 
+For BEVFusion detection model, please run:
+```bash
+torchpack dist-run -np 8 python tools/train.py configs/nuscenes/det/transfusion/secfpn/camera+lidar/swint_v0p075/convfuser.yaml --model.encoders.camera.backbone.init_cfg.checkpoint pretrained/swint-nuimages-pretrained.pth --load_from pretrained/lidar-only-det.pth 
+```
+
+For BEVFusion segmentation model, please run:
+```bash
+torchpack dist-run -np 8 python tools/train.py configs/nuscenes/seg/fusion-bev256d2-lss.yaml --model.encoders.camera.backbone.init_cfg.checkpoint pretrained/swint-nuimages-pretrained.pth
+```
+
+Note: please run `tools/test.py` separately after training to get the final evaluation metrics.
+
 ## FAQs
 
 Q: Can we directly use the info files prepared by mmdetection3d?

configs/nuscenes/det/transfusion/secfpn/camera+lidar/default.yaml

@@ -53,3 +53,10 @@ model:
           - [0, 0, [1, 1, 0]]
           - [0, 0]
         block_type: basicblock
+
+lr_config: null
+
+optimizer:
+  lr: 2.0e-4
+
+max_epochs: 6

configs/nuscenes/det/transfusion/secfpn/camera+lidar/swint_v0p075/default.yaml

@@ -43,3 +43,9 @@ model:
       test_cfg:
         grid_size: [1440, 1440, 41]
 
+lr_config:
+  policy: CosineAnnealing
+  warmup: linear
+  warmup_iters: 500
+  warmup_ratio: 0.33333333
+  min_lr_ratio: 1.0e-3

configs/nuscenes/seg/fusion-bev256d2-lss.yaml

@@ -17,6 +17,9 @@ model:
         out_indices: [1, 2, 3]
         with_cp: false
         convert_weights: true
+        init_cfg:
+          type: Pretrained
+          checkpoint: https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_tiny_patch4_window7_224.pth
       neck:
         type: GeneralizedLSSFPN
         in_channels: [192, 384, 768]
@@ -104,7 +107,24 @@ model:
     map:
       in_channels: 512
 
-
 optimizer:
   type: AdamW
   lr: 1.0e-4
+  weight_decay: 0.01
+  paramwise_cfg:
+    custom_keys:
+      absolute_pos_embed:
+        decay_mult: 0
+      relative_position_bias_table:
+        decay_mult: 0
+
+optimizer_config:
+  grad_clip:
+    max_norm: 35
+    norm_type: 2
+
+lr_config:
+  policy: cyclic
+
+momentum_config:
+  policy: cyclic