35 lines
838 B
Python
35 lines
838 B
Python
import torch
|
|
import pathlib
|
|
|
|
# training parameters
|
|
BATCH_SIZE = 8
|
|
NUM_EPOCHS = 10
|
|
NUM_WORKERS = 4
|
|
DEVICE = torch.device('cuda') if torch.cuda.is_available() else torch.device('cpu')
|
|
|
|
# input parameters
|
|
IMG_SIZE = (7, 7) # inches
|
|
IMG_DPI = 256
|
|
RESIZE_TO = 416 # ToDo: alter this parameter to (7, 7) * 256 [IMG_SIZE * IMG_DPI]
|
|
|
|
# dataset parameters
|
|
CLASSES = ['__backgroud__', '1']
|
|
NUM_CLASSES = len(CLASSES)
|
|
|
|
# data snippet paramters
|
|
MIN_FREQ = 200
|
|
MAX_FREQ = 1500
|
|
DELTA_FREQ = 200
|
|
FREQ_OVERLAP = 25
|
|
|
|
DELTA_TIME = 60*10
|
|
TIME_OVERLAP = 60*1
|
|
|
|
# output parameters
|
|
DATA_DIR = 'data/dataset'
|
|
OUTDIR = 'model_outputs'
|
|
INFERENCE_OUTDIR = 'inference_outputs'
|
|
for required_folders in [DATA_DIR, OUTDIR, INFERENCE_OUTDIR]:
|
|
if not pathlib.Path(required_folders).exists():
|
|
pathlib.Path(required_folders).mkdir(parents=True, exist_ok=True)
|