train YOLOv8-Nano on COCO with 500 epochs

README_CN.md

@@ -110,7 +110,7 @@ python train.py --cuda -d coco --root path/to/COCO -v yolov1 -bs 16 --max_epoch
 | YOLOv5        | CSPDarkNet-53      |  640  |  √   |  250  |       |                        |                   |  |
 | YOLOX         | CSPDarkNet-L       |  640  |  √   |  300  |       |        46.6            |       66.1        | [ckpt](https://github.com/yjh0410/PyTorch_YOLO_Tutorial/releases/download/yolo_tutorial_ckpt/yolox_coco.pth) |
 | YOLOv7-Nano   | ELANNet-Nano       |  640  |  √   |  300  |       |                        |                   |  |
-| YOLOv7-Tiny   | ELANNet-Tiny       |  640  |  √   |  300  |       |                        |                   |  |
+| YOLOv7-Tiny   | ELANNet-Tiny       |  640  |  √   |  300  |       |        37.7            |       56.6        | [ckpt](https://github.com/yjh0410/PyTorch_YOLO_Tutorial/releases/download/yolo_tutorial_ckpt/yolov7_tiny_coco.pth) |
 | YOLOv7-Large  | ELANNet-Large      |  640  |  √   |  300  |       |                        |                   |  |
 | YOLOv8-Nano   | CSP-ELANNet-Nano   |  640  |  √   |  300  |       |                        |                   |  |
 | YOLOv8-Small  | CSP-ELANNet-Small  |  640  |  √   |  300  |       |                        |                   |  |

train.sh

@@ -7,7 +7,7 @@ python train.py \
         -bs 16 \
         -size 640 \
         --wp_epoch 1 \
-        --max_epoch 300 \
+        --max_epoch 500 \
         --eval_epoch 10 \
         --ema \
         --fp16 \