NeuronDS/CL_forecasting_foundation_model

1{2  "architectures": [3    "PatchTSMixerForPrediction"4  ],5  "channel_consistent_masking": true,6  "context_length": 336,7  "d_model": 16,8  "distribution_output": "student_t",9  "dropout": 0.2,10  "expansion_factor": 2,11  "gated_attn": true,12  "head_aggregation": "max_pool",13  "head_dropout": 0.2,14  "init_std": 0.02,15  "loss": "mse",16  "mask_type": "random",17  "mask_value": 0,18  "masked_loss": true,19  "mode": "common_channel",20  "model_type": "patchtsmixer",21  "norm_eps": 1e-05,22  "norm_mlp": "LayerNorm",23  "num_forecast_mask_patches": [24    225  ],26  "num_input_channels": 5,27  "num_layers": 8,28  "num_parallel_samples": 100,29  "num_patches": 42,30  "num_targets": 3,31  "output_range": null,32  "patch_last": true,33  "patch_length": 8,34  "patch_stride": 8,35  "positional_encoding_type": "sincos",36  "post_init": false,37  "prediction_channel_indices": null,38  "prediction_length": 24,39  "random_mask_ratio": 0.5,40  "scaling": "std",41  "self_attn": false,42  "self_attn_heads": 1,43  "torch_dtype": "float32",44  "transformers_version": "4.37.2",45  "unmasked_channel_indices": null,46  "use_positional_encoding": false47}48