Diffusors

Introduction

This project is folked from diffusors

Setup

1. Install the required libraries:

pip install pytorch_lightning wandb

2. Log in to Weights & Biases:

wandb login

Usage

Replace the TensorBoard logger with the Weights & Biases logger as follows:

import wandb
from pytorch_lightning.loggers import WandbLogger

# Initialize wandb
wandb.init(project='my-project', entity='my-entity')

# Create a WandbLogger
wandb_logger = WandbLogger()

Please replace 'my-project' and 'my-entity' with your actual project and entity names.

3. Training

python diffusors/segmentation_diffuser_two.py --train_samples 500 \
                                --val_samples 200 \
                                --test_samples 100 \
                                --datadir . \
                                --epochs 200 \
                                --devices 1 \
                                --strategy ddp \
                                --batch_size 2 \
                                --accelerator gpu \
                                --logsdir . \
                                --precision 16 \
                                --is_use_cycle True

License

MIT