Time Series Forecasting
Transformers
PyTorch
patchtst
Generated from Trainer
time series
forecasting
pretrained models
foundation models
time series foundation models
time-series
Instructions to use ibm-research/patchtst-etth1-pretrain with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ibm-research/patchtst-etth1-pretrain with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, PatchTSTForMaskPretraining tokenizer = AutoTokenizer.from_pretrained("ibm-research/patchtst-etth1-pretrain") model = PatchTSTForMaskPretraining.from_pretrained("ibm-research/patchtst-etth1-pretrain", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download config.json from ibm-research/patchtst-etth1-pretrain: direct link, hf CLI and curl.
- Browser
- Download file 1.26 kB
-
https://huggingface.co/ibm-research/patchtst-etth1-pretrain/resolve/main/config.json
- Command line
-
hf download hf://ibm-research/patchtst-etth1-pretrain/config.json
-
curl -L -o config.json https://huggingface.co/ibm-research/patchtst-etth1-pretrain/resolve/main/config.json
1.26 kB
| { | |
| "_name_or_path": "model/patchtst_etth_pretraining", | |
| "activation_function": "gelu", | |
| "architectures": [ | |
| "PatchTSTForMaskPretraining" | |
| ], | |
| "attention_dropout": 0.0, | |
| "bias": true, | |
| "channel_attention": false, | |
| "channel_consistent_masking": false, | |
| "context_length": 512, | |
| "d_model": 128, | |
| "dropout": 0.3, | |
| "dropout_path": 0.0, | |
| "encoder_attention_heads": 16, | |
| "encoder_ffn_dim": 512, | |
| "encoder_layers": 6, | |
| "ff_dropout": 0.0, | |
| "head_dropout": 0.2, | |
| "init_std": 0.02, | |
| "learn_pe": false, | |
| "mask_input": true, | |
| "mask_patch_ratios": [ | |
| 1, | |
| 1 | |
| ], | |
| "mask_patches": [ | |
| 2, | |
| 3 | |
| ], | |
| "mask_ratio": 0.2, | |
| "mask_type": "random", | |
| "mask_value": 0, | |
| "model_type": "patchtst", | |
| "norm": "BatchNorm", | |
| "num_classes": 0, | |
| "num_input_channels": 7, | |
| "num_output_channels": 1, | |
| "num_patches": 42, | |
| "patch_length": 12, | |
| "pooling": null, | |
| "positional_dropout": 0.0, | |
| "positional_encoding": "sincos", | |
| "pre_norm": false, | |
| "prediction_length": 24, | |
| "prediction_range": null, | |
| "scaling": "mean", | |
| "seed_number": null, | |
| "shared_embedding": true, | |
| "shared_projection": true, | |
| "stride": 12, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.33.0.dev0", | |
| "unmasked_channel_indices": null, | |
| "use_cls_token": false | |
| } | |