# DeepinDetection

**Repository Path**: mikigo/deepindetection

## Basic Information

- **Project Name**: DeepinDetection
- **Description**: Based on MMDetection, use FasterRCNN, ready to train on Deepin/UOS dataset.
- **Primary Language**: Unknown
- **License**: Apache-2.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 0
- **Created**: 2024-11-12
- **Last Updated**: 2024-11-15

## Categories & Tags

**Categories**: Uncategorized

**Tags**: detection, deepin, fasterrcnn

## README

# DeepinDetection

Based on MMDetection, use FasterRCNN, ready to train on Deepin/UOS dataset.

## config file 
[configs/deepin/faster-rcnn_r101_fpn_2x_coco.py](configs/deepin/faster-rcnn_r101_fpn_2x_coco.py) :

```python
# 

_base_ = './faster-rcnn_r50_fpn_2x_coco.py'

model = dict(
    roi_head=dict(
        bbox_head=dict(
            num_classes=2
        )
    )
)

# dataset
dataset_type = 'CocoDataset'
data_root = 'data/coco/'

classes = (
    "dde_file_manager_icon",
    "dde_launcher_icon",
)
backend_args = None

train_pipeline = [
    dict(type='LoadImageFromFile', backend_args=backend_args),
    dict(type='LoadAnnotations', with_bbox=True),
    dict(type='Resize', scale=(1333, 800), keep_ratio=True),
    dict(type='RandomFlip', prob=0.5),
    dict(type='PackDetInputs')
]
test_pipeline = [
    dict(type='LoadImageFromFile', backend_args=backend_args),
    dict(type='Resize', scale=(1333, 800), keep_ratio=True),
    # If you don't have a gt annotation, delete the pipeline
    dict(type='LoadAnnotations', with_bbox=True),
    dict(
        type='PackDetInputs',
        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
                   'scale_factor'))
]

train_dataloader = dict(
    batch_size=2,
    num_workers=2,
    persistent_workers=True,
    sampler=dict(type='DefaultSampler', shuffle=True),
    batch_sampler=dict(type='AspectRatioBatchSampler'),
    dataset=dict(
        type=dataset_type,
        data_root=data_root,
        metainfo=dict(classes=classes),
        ann_file='annotations/train.json',
        data_prefix=dict(img='train2017/'),
        filter_cfg=dict(filter_empty_gt=True, min_size=32),
        pipeline=train_pipeline,
        backend_args=backend_args))
val_dataloader = dict(
    batch_size=1,
    num_workers=2,
    persistent_workers=True,
    drop_last=False,
    sampler=dict(type='DefaultSampler', shuffle=False),
    dataset=dict(
        type=dataset_type,
        data_root=data_root,
        metainfo=dict(classes=classes),
        ann_file='annotations/val.json',
        data_prefix=dict(img='val2017/'),
        test_mode=True,
        pipeline=test_pipeline,
        backend_args=backend_args))
test_dataloader = val_dataloader

val_evaluator = dict(
    type='CocoMetric',
    ann_file=data_root + 'annotations/val.json',
    metric='bbox',
    format_only=False,
    backend_args=backend_args)
test_evaluator = val_evaluator

# optimizer
optim_wrapper = dict(
    type='OptimWrapper',
    optimizer=dict(type='SGD', lr=0.02 / 10, momentum=0.9, weight_decay=0.0001))

# Default setting for scaling LR automatically
#   - `enable` means enable scaling LR automatically
#       or not by default.
#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
auto_scale_lr = dict(enable=False, base_batch_size=16 / 8)
```
## Train

```bash
python tools/train.py configs/deepin/faster-rcnn_r101_fpn_2x_coco.py
```

## Inference

```bash
python infer.py
```