{"library":"torch-ema","type":"library","category":null,"description":"torch-ema is a compact PyTorch library designed for efficiently computing and managing exponential moving averages of model parameters during the training of deep learning models. It helps stabilize training and often leads to improved generalization. The current version is 0.3.0, with the last release in November 2021, indicating a slow release cadence.","language":"python","status":"active","version":"0.3.0","tags":["pytorch","ema","deep learning","training utility","exponential moving average"],"last_verified":"Mon May 25","install":[{"cmd":"pip install torch-ema","imports":["from torch_ema import ExponentialMovingAverage"]}],"homepage":null,"github":"https://github.com/fadel/pytorch_ema","docs":null,"changelog":null,"pypi":"https://pypi.org/project/torch-ema/","npm":null,"openapi_spec":null,"status_page":null,"smithery":null,"compatibility":{"summary":{"python_range":"3.10–3.9","success_rate":100,"avg_install_s":73.8,"avg_import_s":5.11,"wheel_type":"wheel"},"url":"https://checklist.day/v1/registry/torch-ema/compatibility"}}