Nixtla/neuralforecast v1.1.0

on GitHub

What's Changed

• [FIX] Update license in #285
• [FEAT] Exogenous variables in #286
• scalers class in #288
• General documentation improvements in #287
• Fixed README links and added SoTA runs to examples in #291
• Improved documentation in #293
• Improved documentation in #297
• Improved RNN-based/BaseRecurrent/Windows in #298
• Save load in #292
• Fix normalizers in #301
• Improved example notebooks, changed numeration in #302
• Static variables in #305
• Turned Pytorch losses into torch.nn.module classes in #311
• Changed enable_checkpointing to False default in #309
• Correct main link pointers in #314
• Rnns normalizers in #316
• rnns with decoders, autos, usage examples, fix val in #320
• Improved documentation3 in #322
• getting started with LSTM and NHITS in #323
• recovered MQ-NHITS in #327

Full Changelog: v1.0.0...v1.1.0