Skip to content

Commit

Permalink
add yolov11
Browse files Browse the repository at this point in the history
  • Loading branch information
@cxnt
cxnt committed Oct 23, 2024
1 parent 2647494 commit 750eaa7
Show file tree
Hide file tree
Showing 5 changed files with 182 additions and 7 deletions.
4 changes: 2 additions & 2 deletions README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -90,13 +90,13 @@ This application is a versatile tool designed for data transformation tasks (lik
| **Neural Networks** | | | |
| [Apply NN Inference](https://github.com/supervisely-ecosystem/data-nodes/blob/master/src/ui/dtl/actions/neural_networks/apply_nn_inference/README.md#apply-nn-inference) | Apply neural network inference. | + | - |
| [Deploy YOLOv5](https://github.com/supervisely-ecosystem/data-nodes/blob/master/src/ui/dtl/actions/neural_networks/deploy/yolov5.md) | Deploy pretrained or custom YOLOv5 model. | + | - |
| [Deploy YOLO (v8, v9, v10)](https://github.com/supervisely-ecosystem/data-nodes/blob/master/src/ui/dtl/actions/neural_networks/deploy/yolov8.md) | Deploy pretrained or custom YOLO (v8, v9, v10) model. | + | - |
| [Deploy YOLO (v8, v9, v10, v11)](https://github.com/supervisely-ecosystem/data-nodes/blob/master/src/ui/dtl/actions/neural_networks/deploy/yolov8.md) | Deploy pretrained or custom YOLO (v8, v9, v10, v11) model. | + | - |
| [Deploy MMDetection](https://github.com/supervisely-ecosystem/data-nodes/blob/master/src/ui/dtl/actions/neural_networks/deploy/mmdetection.md) | Deploy pretrained or custom MMDetection model. | + | - |
| [Deploy MMSegmentation](https://github.com/supervisely-ecosystem/data-nodes/blob/master/src/ui/dtl/actions/neural_networks/deploy/mmsegmentation.md) | Deploy pretrained or custom MMSegmentation model. | + | - |
| **Other** | | | |
| [Dummy](https://github.com/supervisely-ecosystem/data-nodes/blob/master/src/ui/dtl/actions/other/dummy/README.md#dummy) | Dummy node. Do nothing. Can be used to merge layers. | + | - |
| [Dataset](https://github.com/supervisely-ecosystem/data-nodes/blob/master/src/ui/dtl/actions/other/dataset/README.md#dataset) | All data will be put into a single dataset. | + | - |
| [Split Data](https://github.com/supervisely-ecosystem/data-nodes/blob/mastersrc/ui/dtl/actions/other/split_data/README.md#split-data) | All data will be distributed across datasets based on split method. | + | - |
| [Split Data](https://github.com/supervisely-ecosystem/data-nodes/blob/mastersrc/ui/dtl/actions/other/split_data/README.md#split-data) | All data will be distributed across datasets based on split method. | + | - |
| **Output** | | | |
| [Output Project](https://github.com/supervisely-ecosystem/data-nodes/blob/master/src/ui/dtl/actions/output/output_project/README.md#output_project) | Save data to a new or existing project. | + | - |
| [Create new Project](https://github.com/supervisely-ecosystem/data-nodes/blob/master/src/ui/dtl/actions/output/create_new_project/README.md#create-new-project) | Create a new project and save data there. | + | + |
Expand Down
4 changes: 2 additions & 2 deletions src/ui/dtl/actions/neural_networks/deploy/deploy.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -350,8 +350,8 @@ class DeployYOLOV5Action(DeployBaseAction):

class DeployYOLOV8Action(DeployBaseAction):
name = "deploy_yolo_v8"
title = "Deploy YOLO v8 | v9 | v10"
description = "Deploy YOLO v8 | v9 | v10 models."
title = "Deploy YOLO v8 | v9 | v10 | v11"
description = "Deploy YOLO v8 | v9 | v10 | v11 models."
md_description = DeployBaseAction.read_md_file(dirname(realpath(__file__)) + "/yolov8.md")
need_runtime_selector = True

Expand Down
170 changes: 170 additions & 0 deletions src/ui/dtl/actions/neural_networks/deploy/layout/pretrained_models.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -449,6 +449,176 @@
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.2.0/yolov10x.pt",
},
},
{
"Model": "YOLO11n-det",
"Size (pixels)": "640",
"mAP": "39.5",
"params (M)": "2.6",
"FLOPs (B)": "6.5",
"meta": {
"task_type": "object detection",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt",
},
},
{
"Model": "YOLO11s-det",
"Size (pixels)": "640",
"mAP": "47.0",
"params (M)": "9.4",
"FLOPs (B)": "21.5",
"meta": {
"task_type": "object detection",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11s.pt",
},
},
{
"Model": "YOLO11m-det",
"Size (pixels)": "640",
"mAP": "51.5",
"params (M)": "20.1",
"FLOPs (B)": "68.0",
"meta": {
"task_type": "object detection",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11m.pt",
},
},
{
"Model": "YOLO11l-det",
"Size (pixels)": "640",
"mAP": "53.4",
"params (M)": "25.3",
"FLOPs (B)": "86.9",
"meta": {
"task_type": "object detection",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11l.pt",
},
},
{
"Model": "YOLO11x-det",
"Size (pixels)": "640",
"mAP": "54.7",
"params (M)": "56.9",
"FLOPs (B)": "194.9",
"meta": {
"task_type": "object detection",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11x.pt",
},
},
{
"Model": "YOLO11n-pose",
"Size (pixels)": "640",
"mAP": "50.0",
"params (M)": "2.9",
"FLOPs (B)": "7.6",
"meta": {
"task_type": "pose estimation",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n-pose.pt",
},
},
{
"Model": "YOLO11s-pose",
"Size (pixels)": "640",
"mAP": "58.9",
"params (M)": "9.9",
"FLOPs (B)": "23.2",
"meta": {
"task_type": "pose estimation",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11s-pose.pt",
},
},
{
"Model": "YOLO11m-pose",
"Size (pixels)": "640",
"mAP": "64.9",
"params (M)": "20.9",
"FLOPs (B)": "71.7",
"meta": {
"task_type": "pose estimation",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11m-pose.pt",
},
},
{
"Model": "YOLO11l-pose",
"Size (pixels)": "640",
"mAP": "66.1",
"params (M)": "26.2",
"FLOPs (B)": "90.7",
"meta": {
"task_type": "pose estimation",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11l-pose.pt",
},
},
{
"Model": "YOLO11x-pose",
"Size (pixels)": "640",
"mAP": "69.5",
"params (M)": "58.8",
"FLOPs (B)": "203.3",
"meta": {
"task_type": "pose estimation",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11x-pose.pt",
},
},
{
"Model": "YOLO11n-seg",
"Size (pixels)": "640",
"mAP (box)": "38.9",
"mAP (mask)": "32.0",
"params (M)": "2.9",
"FLOPs (B)": "10.4",
"meta": {
"task_type": "instance segmentation",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n-seg.pt",
},
},
{
"Model": "YOLO11s-seg",
"Size (pixels)": "640",
"mAP (box)": "46.6",
"mAP (mask)": "37.8",
"params (M)": "10.1",
"FLOPs (B)": "35.5",
"meta": {
"task_type": "instance segmentation",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11s-seg.pt",
},
},
{
"Model": "YOLO11m-seg",
"Size (pixels)": "640",
"mAP (box)": "51.5",
"mAP (mask)": "41.5",
"params (M)": "22.4",
"FLOPs (B)": "123.3",
"meta": {
"task_type": "instance segmentation",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11m-seg.pt",
},
},
{
"Model": "YOLO11l-seg",
"Size (pixels)": "640",
"mAP (box)": "53.4",
"mAP (mask)": "42.9",
"params (M)": "27.6",
"FLOPs (B)": "142.2",
"meta": {
"task_type": "instance segmentation",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11l-seg.pt",
},
},
{
"Model": "YOLO11x-seg",
"Size (pixels)": "640",
"mAP (box)": "54.7",
"mAP (mask)": "43.8",
"params (M)": "62.1",
"FLOPs (B)": "319.0",
"meta": {
"task_type": "instance segmentation",
"weights_url": "https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11x-seg.pt",
},
},
]

mmdetection3 = [
Expand Down
11 changes: 8 additions & 3 deletions src/ui/dtl/actions/neural_networks/deploy/yolov8.md
View file
Original file line number Diff line number Diff line change
@@ -1,14 +1,19 @@
# YOLO v8 | v9 | v10
# YOLO v8 | v9 | v10 | v11

`Deploy YOLO v8 | v9 | v10` using [Serve YOLO v8 | v9 | v10](https://ecosystem.supervisely.com/apps/yolov8/serve) app to serve models and can be used to deploy custom and pretrained models that you can use via `Apply NN` layer. Custom models will appear in the custom tab of the table only if you have any trained YOLO v8 | v9 | v10 models in your Team Files. You can train your own model using [Train YOLO v8 | v9 | v10](https://ecosystem.supervisely.com/apps/yolov8/train) app. If you want to use pretrained models, simply select "Pretrained public models" tab in model selector.
`Deploy YOLO v8 | v9 | v10 | v11` using [Serve YOLO v8 | v9 | v10 | v11](https://ecosystem.supervisely.com/apps/yolov8/serve) app to serve models and can be used to deploy custom and pretrained models that you can use via `Apply NN` layer. Custom models will appear in the custom tab of the table only if you have any trained YOLO v8 | v9 | v10 | v11 models in your Team Files. You can train your own model using [Train YOLO v8 | v9 | v10 | v11](https://ecosystem.supervisely.com/apps/yolov8/train) app. If you want to use pretrained models, simply select "Pretrained public models" tab in model selector.

![image](https://github.com/supervisely-ecosystem/data-nodes/assets/48913536/791a50bf-51b5-4bdb-8bda-0f4ed2a3d06b)

- YOLOv8 is a powerful neural network architecture that provides both decent accuracy of predictions and high speed of inference. In comparison to YOLOv5, YOLOv8 uses an anchor-free head (allowing to speed up the non-max suppression (NMS) process), a new backbone, and new loss functions.
- YOLOv9 builds on the advancements of YOLOv8 by further improving the model's performance and efficiency. It incorporates extended feature extraction techniques, advanced loss functions and optimized training processes for better accuracy and faster inference times.
- YOLOv10 introduces consistent dual assignments for NMS-free training and adopts a holistic efficiency-accuracy-driven model design strategy.
- YOLOv11, the latest iteration in the YOLO series, continues the trend of balancing accuracy and efficiency with significant advancements. It introduces a hybrid architecture that combines the strengths of anchor-free and anchor-based methods, allowing the model to adapt dynamically based on the characteristics of the input data.

### Settings:

### How to use:

1. Add `Deploy YOLO v8 | v9 | v10` layer
1. Add `Deploy YOLO v8 | v9 | v10 | v11` layer
2. Open agent settings and select agent and device
3. Open models selector and select one of the available models
4. Press `SERVE`
Expand Down
Binary file added static/images_project_1.jpg
View file
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.

0 comments on commit 750eaa7

Please sign in to comment.