Time Series Forecasting
Transformers
Safetensors
deepwind
wind-power
probabilistic-forecasting
mixture-of-experts
zero-shot
foundation-model
Instructions to use Hexian-2001/DeepWind1.0-33M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Hexian-2001/DeepWind1.0-33M with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("Hexian-2001/DeepWind1.0-33M", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Upload config.json with huggingface_hub
Browse files- config.json +12 -1
config.json
CHANGED
|
@@ -2,12 +2,22 @@
|
|
| 2 |
"architectures": [
|
| 3 |
"DeepWindModel"
|
| 4 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
"context_length": 8192,
|
| 6 |
"d_ff": 1024,
|
| 7 |
"d_model": 384,
|
| 8 |
"dense_act_fn": "silu",
|
| 9 |
"dropout": 0.05,
|
| 10 |
"dtype": "float32",
|
|
|
|
| 11 |
"input_patch_size": 16,
|
| 12 |
"input_patch_stride": 16,
|
| 13 |
"model_type": "deepwind",
|
|
@@ -17,6 +27,7 @@
|
|
| 17 |
"num_known_variates": 10,
|
| 18 |
"num_layers": 6,
|
| 19 |
"num_patch_stats": 9,
|
|
|
|
| 20 |
"quantiles": [
|
| 21 |
0.01,
|
| 22 |
0.05,
|
|
@@ -40,7 +51,7 @@
|
|
| 40 |
0.95,
|
| 41 |
0.99
|
| 42 |
],
|
| 43 |
-
"transformers_version": "4.57.
|
| 44 |
"use_arcsinh": true,
|
| 45 |
"use_coord_embed": true,
|
| 46 |
"use_load_balance_loss": true,
|
|
|
|
| 2 |
"architectures": [
|
| 3 |
"DeepWindModel"
|
| 4 |
],
|
| 5 |
+
"aux_loss_weight": 0.02,
|
| 6 |
+
"channel_loss_weights": [
|
| 7 |
+
1,
|
| 8 |
+
0,
|
| 9 |
+
0,
|
| 10 |
+
0,
|
| 11 |
+
0,
|
| 12 |
+
0
|
| 13 |
+
],
|
| 14 |
"context_length": 8192,
|
| 15 |
"d_ff": 1024,
|
| 16 |
"d_model": 384,
|
| 17 |
"dense_act_fn": "silu",
|
| 18 |
"dropout": 0.05,
|
| 19 |
"dtype": "float32",
|
| 20 |
+
"gmm_components": 3,
|
| 21 |
"input_patch_size": 16,
|
| 22 |
"input_patch_stride": 16,
|
| 23 |
"model_type": "deepwind",
|
|
|
|
| 27 |
"num_known_variates": 10,
|
| 28 |
"num_layers": 6,
|
| 29 |
"num_patch_stats": 9,
|
| 30 |
+
"pred_head_type": "quantile",
|
| 31 |
"quantiles": [
|
| 32 |
0.01,
|
| 33 |
0.05,
|
|
|
|
| 51 |
0.95,
|
| 52 |
0.99
|
| 53 |
],
|
| 54 |
+
"transformers_version": "4.57.6",
|
| 55 |
"use_arcsinh": true,
|
| 56 |
"use_coord_embed": true,
|
| 57 |
"use_load_balance_loss": true,
|