{ "_name_or_path": "ibm/patchtsmixer-etth1-pretrain", "architectures": [ "PatchTSMixerForForecasting" ], "channel_consistent_masking": true, "distribution_output": "student_t", "dropout": 0.5, "expansion_factor": 3, "forecast_channel_indices": null, "forecast_len": 96, "gated_attn": true, "head_agg": "max_pool", "head_dropout": 0.7, "init_std": 0.02, "input_size": 7, "learn_pe": false, "loss": "mse", "mask_patch_ratios": [ 1, 1 ], "mask_patches": [ 2, 3 ], "mask_ratio": 0.5, "mask_type": "random", "mask_value": 0, "masked_loss": true, "mode": "common_channel", "model_type": "patchtsmixer", "n_classes": 3, "n_targets": 3, "norm_mlp": "LayerNorm", "num_features": 48, "num_layers": 2, "num_parallel_samples": 100, "num_patches": 32, "output_range": null, "patch_last": true, "patch_len": 16, "pe": "zeros", "post_init": false, "revin": true, "scaling": "std", "seed_number": null, "self_attn": false, "self_attn_heads": 1, "seq_len": 512, "stride": 16, "torch_dtype": "float32", "transformers_version": "4.35.0.dev0", "use_pe": false }