forklift_yolo11n_v20260514 — YOLO11n det (cvat2 #9 +18 task 增量)

Single class forklift。同 v20260502 stack：YOLO11n / 雙 5090 DDP / 640×640 / batch=64 / optimizer=auto / 100 epochs / patience=30。

v514 增量 (vs v20260502 baseline)：

cvat2 #9 + 18 task / +797 frame（Train +548 / Val +126 / Test +123）
跳過 10 個 EMPTY task (HONCHUAN_CH00_20260504_016-025, 146 frame)

對照表

	v20260501/502 (baseline)	v20260514 (this, +18 task)
test mAP@0.5	0.9882*	0.9880
test mAP@0.5:0.95	0.8503*	0.8542
test P	—	0.9788
test R	—	0.9657

* v20260502 baseline 從 ~/runs_new/forklift_v20260501/run/results.csv 自動讀取（best epoch）。test set 與 v514 略有差異，僅供大致比較。

v514 test eval (full)

{
  "split": "test",
  "weights": "/home/ubuntu/runs_new/forklift_yolo11n_v20260514/run/weights/best.pt",
  "map50": 0.9879759267197881,
  "map75": 0.9406910259996931,
  "map50_95": 0.8542324191107594,
  "mp": 0.9787833227622136,
  "mr": 0.9656709934164212,
  "per_class": {
    "forklift": {
      "ap50": 0.9879759267197881,
      "ap50_95": 0.8542324191107594
    }
  }
}

Final epoch (val) metrics

metric	value
train/box_loss	0.5217
train/cls_loss	0.2502
train/dfl_loss	0.9407
metrics/precision(B)	0.9578
metrics/recall(B)	0.9641
metrics/mAP50(B)	0.9862
metrics/mAP50-95(B)	0.8496
val/box_loss	0.6376
val/cls_loss	0.3423
val/dfl_loss	0.9509

Train / val curves

Eval curves

PR curve

F1 curve

P curve

R curve

Confusion matrix

Test sample inference (conf≥0.25)

t2284_frame_00008610.jpg (preds=0)

t1695_rf_hitsz_train_cut_12153785-CH-1-20220609092022-aa_210_jpg.rf.a3cf94ed9ecbb2c79ffdb2ea4a0d1ce0.jpg (preds=1)

t1701_rf_phantom_train_961899-5841_jpg.rf.91a5d6e85e23e5cc91b9b8c7f6c9edc9.jpg (preds=0)

t1698_kg_walidguirat_train_QWXLQOR02ED4_jpg.rf.a9dc7d9c036524c5d704a09ce8bf1494.jpg (preds=1)

t2295_frame_00006961.jpg (preds=0)

t2295_frame_00007021.jpg (preds=0)

t2284_frame_00008790.jpg (preds=0)

t2295_frame_00007411.jpg (preds=0)

Train args

task: detect
mode: train
model: /home/ubuntu/yolo11n.pt
data: /mnt/ssd/cvat2/datasets/forklift_v514/data.yaml
epochs: 100
time: null
patience: 30
batch: 64
imgsz: 640
save: true
save_period: -1
cache: ram
device: 0,1
workers: 8
project: /home/ubuntu/runs_new/forklift_yolo11n_v20260514
name: run
exist_ok: true
pretrained: true
optimizer: auto
verbose: true
seed: 0
deterministic: true
single_cls: false
rect: false
cos_lr: false
close_mosaic: 10
resume: false
amp: true
fraction: 1.0
profile: false
freeze: null
multi_scale: 0.0
compile: false
overlap_mask: true
mask_ratio: 4
dropout: 0.0
val: true
split: val
save_json: false
conf: null
iou: 0.7
max_det: 300
half: false
dnn: false
plots: true
end2end: null
source: null
vid_stride: 1
stream_buffer: false
visualize: false
augment: false
agnostic_nms: false
classes: null
retina_masks: false
embed: null
show: false
save_frames: false
save_txt: false
save_conf: false
save_crop: false
show_labels: true
show_conf: true
show_boxes: true
line_width: null
format: torchscript
keras: false
optimize: false
int8: false
dynamic: false
simplify: true
opset: null
workspace: null
nms: false
lr0: 0.01
lrf: 0.01
momentum: 0.937
weight_decay: 0.0005
warmup_epochs: 3.0
warmup_momentum: 0.8
warmup_bias_lr: 0.1
box: 7.5
cls: 0.5
dfl: 1.5
pose: 12.0
kobj: 1.0
rle: 1.0
angle: 1.0
nbs: 64
hsv_h: 0.015
hsv_s: 0.7
hsv_v: 0.4
degrees: 0.0
translate: 0.1
scale: 0.5
shear: 0.0
perspective: 0.0
flipud: 0.0
fliplr: 0.5
bgr: 0.0
mosaic: 1.0
mixup: 0.0
cutmix: 0.0
copy_paste: 0.0
copy_paste_mode: flip
auto_augment: randaugment
erasing: 0.4
cfg: null
tracker: botsort.yaml
save_dir: /home/ubuntu/runs_new/forklift_yolo11n_v20260514/run