### Train FasterNet-T0 on ImageNet-1K (8-GPU)

Source: https://github.com/jierunchen/fasternet/blob/master/README.MD

Example command for training FasterNet-T0 on ImageNet-1K using 8 GPUs. Modify configuration, batch size, and other parameters as required for different variants or setups.

```bash
python train_test.py -g 0,1,2,3,4,5,6,7 --num_nodes 1 -n 4 -b 4096 -e 2000 \
--data_dir ../../data/imagenet --pin_memory --wandb_project_name fasternet \
--model_ckpt_dir ./model_ckpt/$(date +'%Y%m%d_%H%M%S') --cfg cfg/fasternet_t0.yaml
```

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

### Install Project Dependencies

Source: https://github.com/jierunchen/fasternet/blob/master/requirements.txt

Installs all Python packages listed in the requirements.txt file.

```bash
# pip install -r requirements.txt
```

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

### Clone Repository and Install Dependencies

Source: https://github.com/jierunchen/fasternet/blob/master/README.MD

Clones the FasterNet repository and installs all required Python packages listed in 'requirements.txt'.

```bash
git clone https://github.com/JierunChen/FasterNet
cd FasterNet/
pip install -r requirements.txt
```

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

### Install Detection Dependencies

Source: https://github.com/jierunchen/fasternet/blob/master/detection/README.MD

Install mmcv and mmdet for object detection tasks. Ensure compatibility with your CUDA version.

```bash
pip install mmcv-full==1.6.0 -f https://download.openmmlab.com/mmcv/dist/cu113/torch1.11.0/index.html
pip install mmdet==2.25.0
```

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

### Evaluate FasterNet-T0 on ImageNet-1K

Source: https://github.com/jierunchen/fasternet/blob/master/README.MD

Example command to evaluate a pre-trained FasterNet-T0 model on ImageNet-1K using a single GPU. Adjust checkpoint path and data directory as needed.

```bash
python train_test.py -c cfg/fasternet_t0.yaml \
--checkpoint_path model_ckpt/fasternet_t0-epoch=281-val_acc1=71.9180.pth \
--data_dir ../../data/imagenet --test_phase -g 1 -e 125
```

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

### PyTorch Installation with CUDA

Source: https://github.com/jierunchen/fasternet/blob/master/requirements.txt

Installs PyTorch version 1.11.0 and Torchvision version 0.12.0 with CUDA 11.3 support.

```bash
# --extra-index-url https://download.pytorch.org/whl/cu113

torch==1.11.0
torchvision==0.12.0
```

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

### Install MMDetection with CUDA

Source: https://github.com/jierunchen/fasternet/blob/master/requirements.txt

Installs MMDetection version 2.25.0 with specific dependencies for CUDA 11.3 and PyTorch 1.11.0.

```bash
# pip install mmcv-full==1.6.0 -f https://download.openmmlab.com/mmcv/dist/cu113/torch1.11.0/index.html
# pip install mmdet==2.25.0
```

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

### Create and Activate Conda Environment

Source: https://github.com/jierunchen/fasternet/blob/master/README.MD

Sets up a new conda virtual environment named 'fasternet' with Python 3.9.12 and activates it.

```bash
conda create -n fasternet python=3.9.12 -y
conda activate fasternet
```

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

### Create Conda Environment

Source: https://github.com/jierunchen/fasternet/blob/master/requirements.txt

Creates a Conda environment named 'fasternet' with Python 3.9.12.

```bash
# conda create -n fasternet python=3.9.12 -y
```

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

### Activate Conda Environment

Source: https://github.com/jierunchen/fasternet/blob/master/requirements.txt

Activates the Conda environment named 'fasternet'.

```bash
# conda activate fasternet
```

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

### ImageNet-1K Dataset Structure

Source: https://github.com/jierunchen/fasternet/blob/master/README.MD

Illustrates the required directory structure for the ImageNet-1K dataset for image classification tasks.

```text
/path/to/imagenet-1k/
  train/
    class1/
      img1.jpeg
    class2/
      img2.jpeg
  val/
    class1/
      img3.jpeg
    class2/
      img4.jpeg
```

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

### Measure Latency and Throughput

Source: https://github.com/jierunchen/fasternet/blob/master/README.MD

Command to measure model latency on CPU/ARM and throughput on GPU, with options for fused convolution and batch size.

```bash
python train_test.py -c cfg/fasternet_t0.yaml \
--checkpoint_path model_ckpt/fasternet_t0-epoch=281-val_acc1=71.9180.pth \
--data_dir ../../data/imagenet --test_phase -g 1 -e 32  --measure_latency --fuse_conv_bn
```

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

### Measure GPU Throughput

Source: https://github.com/jierunchen/fasternet/blob/master/detection/README.MD

Measure the inference throughput of FasterNet on a GPU. This command uses PyTorch distributed launcher for benchmarking.

```python
CUDA_VISIBLE_DEVICES=0 python -m torch.distributed.launch --nproc_per_node=1 --master_port=29500 \
benchmark.py configs/fasternet/mask_rcnn_fasternet_s_fpn_1x_coco.py \
ckpts/mask_rcnn_fasternet_s_fpn_1x_coco_20221111_063419.pth --launcher pytorch --fuse-conv-bn
```

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

### Train FasterNet on COCO

Source: https://github.com/jierunchen/fasternet/blob/master/detection/README.MD

Train FasterNet-T0 + Mask R-CNN on COCO train2017 using 8 GPUs. This command specifies the configuration file, number of GPUs, work directory, and a pre-trained model checkpoint.

```bash
CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 bash ./dist_train.sh \
configs/fasternet/mask_rcnn_fasternet_s_fpn_1x_coco.py 8 \
--work-dir work_dirs/mask_rcnn_fasternet_s_fpn_1x_coco/$(date +'%Y%m%d_%H%M%S') \
--cfg-options model.pretrained=../model_ckpt/fasternet_t0-epoch=281-val_acc1=71.9180.pth
```

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

### Measure FLOPs

Source: https://github.com/jierunchen/fasternet/blob/master/detection/README.MD

Calculate the Floating Point Operations (FLOPs) for a given FasterNet configuration.

```python
python get_flops.py configs/fasternet/mask_rcnn_fasternet_s_fpn_1x_coco.py
```

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

### Evaluate FasterNet on COCO

Source: https://github.com/jierunchen/fasternet/blob/master/detection/README.MD

Evaluate FasterNet-T0 + Mask R-CNN on COCO val2017 using multiple GPUs. This command performs both bounding box and segmentation evaluation.

```bash
bash ./dist_test.sh configs/fasternet/mask_rcnn_fasternet_s_fpn_1x_coco.py \
 ckpts/mask_rcnn_fasternet_s_fpn_1x_coco_20221111_063419.pth 8 --eval bbox segm
```

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

### FasterNet Citation

Source: https://github.com/jierunchen/fasternet/blob/master/README.MD

BibTeX entry for citing the FasterNet paper. Use this in academic publications.

```bibtex
@article{chen2023run,
  title={Run, Don't Walk: Chasing Higher FLOPS for Faster Neural Networks},
  author={Chen, Jierun and Kao, Shiu-hong and He, Hao and Zhuo, Weipeng and Wen, Song and Lee, Chul-Ho and Chan, S-H Gary},
  journal={arXiv preprint arXiv:2303.03667},
  year={2023}
}
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.