yolo11n模型转换bin后部署X5失败

Kyhero · 2025 年1 月 28 日 17:12

参考帖子YOLOv11，地瓜RDK X5开发板，TROS端到端140FPS！ - RDK Model Zoo - 地瓜机器人论坛

1. 板卡型号 RDK X5

2. 系统软件版本：-
[RDK OS Version]:

3.1.1

[RDK Kernel Version]:

Linux ubuntu 6.1.83 #1 SMP PREEMPT Fri Jan 10 15:29:06 CST 2025 aarch64 aarch64 aarch64 GNU/Linux

[RDK Miniboot Version]:

U-Boot 2022.10-g505104e8-dirty (Sep 26 2024 - 15:01:37 +0800)

3.hrt_model_exec model_info --model_file yolov11n_nv12.bin

root@08495013b420:/data/horizon_x5/codes/model_outputyolov11# hrt_model_exec model_info --model_file yolov11n_nv12.bin

hrt_model_exec model_info --model_file yolov11n_nv12.bin

I0000 00:00:00.000000 401 vlog_is_on.cc:197] RAW: Set VLOG level for “*” to 3

core[0] open!

core[1] open!

[HBRT] set log level as 0. version = 3.15.54.0

[DNN] Runtime version = 1.23.10_(3.15.54 HBRT)

[A][DNN][packed_model.cpp:247][Model](2025-01-29,00:56:43.195.6) [HorizonRT] The model builder version = 1.23.8

Load model to DDR cost 118.099ms.

This model file has 1 model:

[yolov11n_nv12]

---------------------------------------------------------------------

[model name]: yolov11n_nv12

input[0]:

input source: HB_DNN_INPUT_FROM_PYRAMID

valid shape: (1,3,640,640,)

aligned shape: (1,3,640,640,)

aligned byte size: 614400

tensor type: HB_DNN_IMG_TYPE_NV12

tensor layout: HB_DNN_LAYOUT_NCHW

quanti type: NONE

stride: (0,0,0,0,)

output[0]:

valid shape: (1,80,80,64,)

aligned shape: (1,80,80,64,)

aligned byte size: 1638400

tensor type: HB_DNN_TENSOR_TYPE_F32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: NONE

stride: (1638400,20480,256,4,)

output[1]:

valid shape: (1,40,40,64,)

aligned shape: (1,40,40,64,)

aligned byte size: 409600

tensor type: HB_DNN_TENSOR_TYPE_F32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: NONE

stride: (409600,10240,256,4,)

output[2]:

valid shape: (1,20,20,64,)

aligned shape: (1,20,20,64,)

aligned byte size: 102400

tensor type: HB_DNN_TENSOR_TYPE_F32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: NONE

stride: (102400,5120,256,4,)

output[3]:

valid shape: (1,80,80,80,)

aligned shape: (1,80,80,80,)

aligned byte size: 2048000

tensor type: HB_DNN_TENSOR_TYPE_F32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: NONE

stride: (2048000,25600,320,4,)

output[4]:

valid shape: (1,40,40,80,)

aligned shape: (1,40,40,80,)

aligned byte size: 512000

tensor type: HB_DNN_TENSOR_TYPE_F32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: NONE

stride: (512000,12800,320,4,)

output[5]:

valid shape: (1,20,20,80,)

aligned shape: (1,20,20,80,)

aligned byte size: 128000

tensor type: HB_DNN_TENSOR_TYPE_F32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: NONE

stride: (128000,6400,320,4,)-

5.yolov11nworkconfig.json 内容

{

“model_file”: “/home/sunrise/share/hobot_dnn-develop/dnn_node_example/config/yolov11n_nv12.bin”,

“dnn_Parser”: “yolov8”,

“model_output_count”: 6,

“reg_max”: 16,

“class_num”: 80,

“cls_names_list”: “config/coco.list”,

“strides”: [8, 16, 32],

“score_threshold”: 0.25,

“nms_threshold”: 0.7,

“nms_top_k”: 300

}-

6、hb_perf yolo11n_detect_bayese_640x640_nv12.bin

7、反向量移除后-
hb_model_modifier yolov11n_nv12.bin -r /model.23/cv2.0/cv2.0.2/Conv_output_0_HzDequantize -r /model.23/cv2.1/cv2.1.2/Conv_output_0_HzDequantize -r /model.23/cv2.2/cv2.2.2/Conv_output_0_HzDequantize-

hrt_model_exec model_info --model_file yolov11n_nv12_modified.bin

I0000 00:00:00.000000 1664 vlog_is_on.cc:197] RAW: Set VLOG level for “*” to 3

core[0] open!

core[1] open!

[HBRT] set log level as 0. version = 3.15.54.0

[DNN] Runtime version = 1.23.10_(3.15.54 HBRT)

[A][DNN][packed_model.cpp:247][Model](2025-02-05,19:51:40.267.242) [HorizonRT] The model builder version = 1.23.8

Load model to DDR cost 104.054ms.

This model file has 1 model:

[yolov11n_nv12]

---------------------------------------------------------------------

[model name]: yolov11n_nv12

input[0]:

input source: HB_DNN_INPUT_FROM_PYRAMID

valid shape: (1,3,640,640,)

aligned shape: (1,3,640,640,)

aligned byte size: 614400

tensor type: HB_DNN_IMG_TYPE_NV12

tensor layout: HB_DNN_LAYOUT_NCHW

quanti type: NONE

stride: (0,0,0,0,)

output[0]:

valid shape: (1,80,80,80,)

aligned shape: (1,80,80,80,)

aligned byte size: 2048000

tensor type: HB_DNN_TENSOR_TYPE_F32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: NONE

stride: (2048000,25600,320,4,)

output[1]:

valid shape: (1,80,80,64,)

aligned shape: (1,80,80,64,)

aligned byte size: 1638400

tensor type: HB_DNN_TENSOR_TYPE_S32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: SCALE

stride: (1638400,20480,256,4,)

scale data: 0.000423609,0.000424303,0.000391664,0.000369442,0.000296873,0.000307984,0.000205902,0.000243054,0.000218922,0.000169183,0.000138454,0.000190798,0.000184895,0.000160589,0.000139148,0.00017361,0.000321352,0.000353644,0.000314408,0.000330033,0.000294964,0.00026163,0.000226388,0.000197568,0.000148003,0.00017361,0.000182985,0.000172221,0.000154426,0.000135069,0.000115624,0.000128645,0.000381248,0.000395137,0.000380901,0.000256422,0.000232464,0.000304512,0.000236283,0.000270658,0.000221006,0.000177777,0.000191145,0.000171787,0.000149305,0.000131857,0.000118923,0.000158853,0.000414928,0.000393053,0.000386109,0.000285241,0.000345484,0.000240971,0.000252776,0.000225346,0.000267533,0.000224999,0.000152951,0.000140364,0.000122048,0.000104947,9.27078e-05,0.00015703,

quantizeAxis: 3

output[2]:

valid shape: (1,40,40,80,)

aligned shape: (1,40,40,80,)

aligned byte size: 512000

tensor type: HB_DNN_TENSOR_TYPE_F32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: NONE

stride: (512000,12800,320,4,)

output[3]:

valid shape: (1,40,40,64,)

aligned shape: (1,40,40,64,)

aligned byte size: 409600

tensor type: HB_DNN_TENSOR_TYPE_S32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: SCALE

stride: (409600,10240,256,4,)

scale data: 0.000216406,0.000217248,0.000204618,0.000152832,0.000119487,0.000116792,0.000106687,0.000133717,9.67513e-05,0.000110561,9.32147e-05,8.07103e-05,7.08583e-05,7.05215e-05,6.69007e-05,8.83308e-05,0.000198218,0.000192324,0.000164705,0.00016622,0.000177335,0.000139948,0.000131444,0.000112077,9.36357e-05,7.98682e-05,8.42469e-05,7.89841e-05,7.1532e-05,6.35746e-05,5.51962e-05,5.79329e-05,0.000222637,0.00022129,0.000184914,0.000160326,0.000156031,0.000132286,0.000123107,9.80144e-05,0.000115697,0.000101635,8.00788e-05,6.58482e-05,8.09629e-05,8.68151e-05,8.53416e-05,0.000108877,0.000183735,0.000186935,0.000162515,0.000156789,0.000138854,0.000172283,0.000131612,0.000101467,0.000128412,0.000103403,8.2647e-05,6.69007e-05,6.45009e-05,6.7827e-05,6.55955e-05,8.36154e-05,

quantizeAxis: 3

output[4]:

valid shape: (1,20,20,80,)

aligned shape: (1,20,20,80,)

aligned byte size: 128000

tensor type: HB_DNN_TENSOR_TYPE_F32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: NONE

stride: (128000,6400,320,4,)

output[5]:

valid shape: (1,20,20,64,)

aligned shape: (1,20,20,64,)

aligned byte size: 102400

tensor type: HB_DNN_TENSOR_TYPE_S32

tensor layout: HB_DNN_LAYOUT_NHWC

quanti type: SCALE

stride: (102400,5120,256,4,)

scale data: 0.000817041,0.000808105,0.000702783,0.00062459,0.000628101,0.000690655,0.000697677,0.000726401,0.000583099,0.000456075,0.000538737,0.000400223,0.000499161,0.000427032,0.000447777,0.000520864,0.000788956,0.000883426,0.000710443,0.00071172,0.000606079,0.000641824,0.000482565,0.000465331,0.000479693,0.000543524,0.00038235,0.000376924,0.000348838,0.000294741,0.000191654,9.05607e-05,0.000868106,0.000870021,0.000636399,0.000675336,0.000615653,0.000583099,0.000626186,0.000420968,0.000447139,0.000595866,0.000437883,0.00050714,0.00037086,0.000319316,0.000344689,0.000372456,0.000771083,0.000699592,0.000637356,0.000611185,0.000627781,0.00064342,0.000687464,0.00056395,0.000610866,0.000575759,0.000602249,0.000432138,0.000432457,0.000372456,0.000406606,0.000438521,

quantizeAxis: 3

---------------------------------------------------------------------

root@08495013b420:/data/horizon_x5/codes/model_outputyolov11#

使用 yolov11n_nv12_modified.bin 还是报同样的错-

8、ros2 launch dnn_node_example dnn_node_example_feedback.launch.py dnn_example_config_file:=config/yolov11workconfig.json dnn_example_image:=config/target.jpg-

然后执行root@ubuntu:/home/sunrise/share/hobot_dnn-develop/dnn_node_example# ros2 launch dnn_node_example dnn_node_example_feedback.launch.py dnn_example_config_file:=config/yolov11workconfig.json dnn_example_image:=config/target.jpg

[INFO] [launch]: All log files can be found below /root/.ros/log/2025-02-05-19-49-27-527307-ubuntu-5525

[INFO] [launch]: Default logging verbosity is set to INFO

dnn_node_example_path is /opt/tros/humble/lib/dnn_node_example

cp_cmd is cp -r /opt/tros/humble/lib/dnn_node_example/config .

[INFO] [example-1]: process started with pid [5528]

[example-1] [WARN] [1738756168.143595718] [dnn_example_node]: Parameter:

[example-1] feed_type(0:local, 1:sub): 0

[example-1] image: config/target.jpg

[example-1] image_type: 0

[example-1] dump_render_img: 1

[example-1] is_shared_mem_sub: 0

[example-1] config_file: config/yolov11workconfig.json

[example-1] msg_pub_topic_name: hobot_dnn_detection

[example-1] info_msg_pub_topic_name: hobot_dnn_detection_info

[example-1] ros_img_topic_name: /image

[example-1] sharedmem_img_topic_name: /hbmem_img

[example-1] [WARN] [1738756168.144440794] [dnn_example_node]: Load [80] class types from file [config/coco.list]

[example-1] [WARN] [1738756168.144577876] [dnn_example_node]: Parameter:

[example-1] model_file_name: /home/sunrise/share/yolov11/yolov11n_nv12_modified.bin

[example-1] model_name:

[example-1] [INFO] [1738756168.144651251] [dnn]: Node init.

[example-1] [INFO] [1738756168.144678000] [dnn_example_node]: Set node para.

[example-1] [WARN] [1738756168.144720500] [dnn_example_node]: model_file_name_: /home/sunrise/share/yolov11/yolov11n_nv12_modified.bin, task_num: 4

[example-1] [INFO] [1738756168.144768875] [dnn]: Model init.

[example-1] [BPU_PLAT]BPU Platform Version(1.3.6)!

[example-1] [HBRT] set log level as 0. version = 3.15.55.0

[example-1] [DNN] Runtime version = 1.24.5_(3.15.55 HBRT)

[example-1] [A][DNN][packed_model.cpp:247][Model](2025-02-05,19:49:28.228.259) [HorizonRT] The model builder version = 1.23.8

[example-1] [INFO] [1738756168.312662474] [dnn]: The model input 0 width is 640 and height is 640

[example-1] [INFO] [1738756168.312831097] [dnn]:

[example-1] Model Info:

[example-1] name: yolov11n_nv12.

[example-1] [input]

[example-1] - (0) Layout: NCHW, Shape: [1, 3, 640, 640], Type: HB_DNN_IMG_TYPE_NV12.

[example-1] [output]

[example-1] - (0) Layout: NHWC, Shape: [1, 80, 80, 64], Type: HB_DNN_TENSOR_TYPE_S32.

[example-1] - (1) Layout: NHWC, Shape: [1, 40, 40, 64], Type: HB_DNN_TENSOR_TYPE_S32.

[example-1] - (2) Layout: NHWC, Shape: [1, 20, 20, 64], Type: HB_DNN_TENSOR_TYPE_S32.

[example-1] - (3) Layout: NHWC, Shape: [1, 80, 80, 80], Type: HB_DNN_TENSOR_TYPE_F32.

[example-1] - (4) Layout: NHWC, Shape: [1, 40, 40, 80], Type: HB_DNN_TENSOR_TYPE_F32.

[example-1] - (5) Layout: NHWC, Shape: [1, 20, 20, 80], Type: HB_DNN_TENSOR_TYPE_F32.

[example-1]

[example-1] [INFO] [1738756168.312877222] [dnn]: Task init.

[example-1] [INFO] [1738756168.314481374] [dnn]: Set task_num [4]

[example-1] [WARN] [1738756168.314562748] [dnn_example_node]: Get model name: yolov11n_nv12 from load model.

[example-1] [INFO] [1738756168.314682289] [dnn_example_node]: The model input width is 640 and height is 640

[example-1] [WARN] [1738756168.314733122] [dnn_example_node]: Create ai msg publisher with topic_name: hobot_dnn_detection

[example-1] [INFO] [1738756168.328336252] [dnn_example_node]: Dnn node feed with local image: config/target.jpg

[example-1] [INFO] [1738756168.387182856] [dnn_example_node]: Output from frame_id: feedback, stamp: 0.0

[ERROR] [example-1]: process has died [pid 5528, exit code -11, cmd ‘/opt/tros/humble/lib/dnn_node_example/example --ros-args --log-level info --ros-args --params-file /tmp/launch_params_lntzvdw_ --params-file /tmp/launch_params_1tbgy6b8 --params-file /tmp/launch_params_km52tfwe --params-file /tmp/launch_params_t2ozwn1v --params-file /tmp/launch_params_ej85ixnj’].

root@ubuntu:/home/sunrise/share/hobot_dnn-develop/dnn_node_example#

9、Python部署是可以的-
[RDK_YOLO] [20:15:30.401] [INFO] Namespace(model_path=‘yolov11n_nv12_modified.bin’, test_img=‘target.jpg’, img_save_path=‘jupyter_result.jpg’, classes_num=80, reg=16, iou_thres=0.45, conf_thres=0.25)

[BPU_PLAT]BPU Platform Version(1.3.6)!

[HBRT] set log level as 0. version = 3.15.55.0

[DNN] Runtime version = 1.24.5_(3.15.55 HBRT)

[A][DNN][packed_model.cpp:247][Model](2025-02-05,20:15:30.484.553) [HorizonRT] The model builder version = 1.23.8

[W][DNN]bpu_model_info.cpp:491][Version](2025-02-05,20:15:30.573.448) Model: yolov11n_nv12. Inconsistency between the hbrt library version 3.15.55.0 and the model build version 3.15.54.0 detected, in order to ensure correct model results, it is recommended to use compilation tools and the BPU SDK from the same OpenExplorer package.

[RDK_YOLO] [20:15:30.577] [DEBUG] Load D-Robotics Quantize model time = 175.82 ms

[RDK_YOLO] [20:15:30.578] [INFO] → input tensors

[RDK_YOLO] [20:15:30.578] [INFO] intput[0], name=images, type=uint8, shape=(1, 3, 640, 640)

[RDK_YOLO] [20:15:30.579] [INFO] → output tensors

[RDK_YOLO] [20:15:30.579] [INFO] output[0], name=output0, type=int32, shape=(1, 80, 80, 64)

[RDK_YOLO] [20:15:30.579] [INFO] output[1], name=469, type=int32, shape=(1, 40, 40, 64)

[RDK_YOLO] [20:15:30.579] [INFO] output[2], name=477, type=int32, shape=(1, 20, 20, 64)

[RDK_YOLO] [20:15:30.579] [INFO] output[3], name=491, type=float32, shape=(1, 80, 80, 80)

[RDK_YOLO] [20:15:30.580] [INFO] output[4], name=505, type=float32, shape=(1, 40, 40, 80)

[RDK_YOLO] [20:15:30.580] [INFO] output[5], name=519, type=float32, shape=(1, 20, 20, 80)

[RDK_YOLO] [20:15:30.580] [INFO] self.s_bboxes_scale.shape=(1, 64), self.m_bboxes_scale.shape=(1, 64), self.l_bboxes_scale.shape=(1, 64)

[RDK_YOLO] [20:15:30.580] [INFO] self.weights_static.shape = (1, 1, 16)

[RDK_YOLO] [20:15:30.582] [INFO] self.s_anchor.shape = (6400, 2), self.m_anchor.shape = (1600, 2), self.l_anchor.shape = (400, 2)

[RDK_YOLO] [20:15:30.583] [INFO] iou threshol = 0.45, conf threshol = 0.25

[RDK_YOLO] [20:15:30.583] [INFO] sigmoid_inverse threshol = -1.10

[RDK_YOLO] [20:15:30.618] [DEBUG] bgr8 to nv12 time = 10.66 ms

[RDK_YOLO] [20:15:30.631] [DEBUG] forward time = 12.02 ms

[RDK_YOLO] [20:15:30.638] [DEBUG] c to numpy time = 6.22 ms

[RDK_YOLO] [20:15:30.644] [DEBUG] Post Process time = 6.31 ms

[RDK_YOLO] [20:15:30.645] [INFO] Draw Results:

[RDK_YOLO] [20:15:30.645] [INFO] (112, 336, 560, 496) → couch: 0.64

[RDK_YOLO] [20:15:30.646] [INFO] (496, 208, 624, 464) → potted plant: 0.51

[RDK_YOLO] [20:15:30.647] [INFO] (68, 188, 100, 236) → potted plant: 0.25

[RDK_YOLO] [20:15:30.664] [INFO] saved in path: “./jupyter_result.jpg”

root@ubuntu:/home/sunrise/share/yolov11#-
就是tros部署不行-

从官网下载的bin文件

wget https://archive.d-robotics.cc/downloads/rdk\_model\_zoo/rdk\_x5/yolo11n\_detect\_bayese\_640x640\_nv12\_modified.bin

执行 YOLOv11_Detect_YUV420SP.py 是可以，但是使用tros还是报错

[ERROR] [websocket-4]: process has died [pid 5914, exit code -2, cmd ‘/opt/tros/humble/lib/websocket/websocket --ros-args --log-level warn --ros-args --params-file /tmp/launch_params_9t65onpw --params-file /tmp/launch_params_zj4vy6ch --params-file /tmp/launch_params_q8_2xngk --params-file /tmp/launch_params_3fmln45g --params-file /tmp/launch_params_5tnsuqrz --params-file /tmp/launch_params_eq1_a038’].

这里报的是-2，我生成的报的-11

yolov11n_nv12_modified.bin

CauchyKesai · 2025 年2 月 5 日 11:35

这里需要按照RDK Model Zoo仓库的README对bbox信息的框的反量化节点进行移除：-

仓库：https://github.com/D-Robotics/rdk\_model\_zoo/tree/main-

https://github.com/D-Robotics/rdk\_model\_zoo/blob/main/demos/detect/YOLO11/YOLO11-Detect\_YUV420SP/README\_cn.md#移除bbox信息3个输出头的反量化节点

SkyXZ · 2025 年2 月 1 日 13:05

mipi_cam 节点日志显示是mipi摄像头的/hbmem_img 话题图像数据，建议检查一下MIPI摄像头是否正确连接了或者正确配置了，3.1.1 MIPI摄像头使用 | RDK DOC

Kyhero · 2025 年2 月 5 日 01:51

摄像头是正常的，按照官方文档里跑的yolo 模型是正常的

Kyhero · 2025 年2 月 5 日 12:01

超哥，反向量移除了也还是不行，移除后的信息更新在文档里，

Kyhero · 2025 年2 月 5 日 12:19

用你写的python部署，是可以的，就是tros不行

CauchyKesai · 2025 年2 月 5 日 12:31

这里需要按照RDK Model Zoo仓库的README，在导出为onnx时，对输出头的顺序进行调转，保证与README所描述的bin模型的输出头顺序是一样的。-
顺序为：-
stride为8的分类头-
stride为8的bbox头（移除反量化节点）

stride为16的分类头-
stride为16的bbox头（移除反量化节点）-
stride为32的分类头-
stride为32的bbox头（移除反量化节点）

仓库：https://github.com/D-Robotics/rdk\_model\_zoo/tree/ma

Kyhero · 2025 年2 月 7 日 02:32

输出头修改这三种都尝试过了-

移除反量化节点对比 hb_model_modifier.log-
2025-02-07 10:01:57,879 file: hb_model_modifier.py func: hb_model_modifier line No: 409 input: “/model.23/cv2.0/cv2.0.2/Conv_output_0_quantized”

input: “/model.23/cv2.0/cv2.0.2/Conv_x_scale”

output: “475”

op_type: “Dequantize”

attribute {

type: INT

i: 3

}-
执行 hb_model_modifier yolov11_ball_nv12.bin -r /model.23/cv2.0/cv2.0.2/Conv_output_0_HzDequantize -r /model.23/cv2.1/cv2.1.2/Conv_output_0_HzDequantize -r /model.23/cv2.2/cv2.2.2/Conv_output_0_HzDequantize-

现在就是不知道卡在那个点了，这张图里只有对比只有标题不一样了

Kyhero · 2025 年2 月 7 日 04:07

超哥，我从官网下载的bin文件

wget https://archive.d-robotics.cc/downloads/rdk\_model\_zoo/rdk\_x5/yolo11n\_detect\_bayese\_640x640\_nv12\_modified.bin-
执行 YOLOv11_Detect_YUV420SP.py 是可以，但是使用tros还是报错-
[ERROR] [websocket-4]: process has died [pid 5914, exit code -2, cmd ‘/opt/tros/humble/lib/websocket/websocket --ros-args --log-level warn --ros-args --params-file /tmp/launch_params_9t65onpw --params-file /tmp/launch_params_zj4vy6ch --params-file /tmp/launch_params_q8_2xngk --params-file /tmp/launch_params_3fmln45g --params-file /tmp/launch_params_5tnsuqrz --params-file /tmp/launch_params_eq1_a038’].

这里报的是-2，我生成的报的-11-

CauchyKesai · 2025 年2 月 7 日 12:31

Python如果跑挂了，就手动修改一下输出头的顺序-
stride为8的分类头-
stride为8的bbox头（移除反量化节点）

stride为16的分类头-
stride为16的bbox头（移除反量化节点）-
stride为32的分类头-
stride为32的bbox头（移除反量化节点）

CauchyKesai · 2025 年2 月 7 日 12:32

TROS的一定要和板子上的bin模型顺序一致，同时要升级到最新的3.1.1系统。

Kyhero · 2025 年2 月 7 日 14:02

系统是最新的 [RDK OS Version]: 3.1.1-
TROS的一定要和板子上的bin模型顺序这个顺序指的是哪个

CauchyKesai · 2025 年2 月 8 日 02:22

系统是最新的了，这个没问题。-

TROS自带的使用的bin模型路径在对应的workconfig.json文件里面有指明，是存储在板子上的

Kyhero · 2025 年2 月 10 日 02:02

workconfig.json 是这个那里没配对吗

Kyhero · 2025 年2 月 10 日 02:34

“model_file”: “/home/sunrise/share/yolov11/yolo11n_detect_bayese_640x640_nv12_modified.bin”, 这里写都是绝对路径

robot_9527 · 2025 年3 月 26 日 06:54

兄弟你这个问题解决了吗，我也遇到了

hjkl · 2025 年3 月 29 日 08:38

你好，请问3.1.0升级到3.1.1需要重装镜像吗？使用sudo apt full-upgrade命令更新后还是3.1.0