Instructions to use openclimatefix/perceiver-io with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use openclimatefix/perceiver-io with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("openclimatefix/perceiver-io", dtype="auto") - Notebooks
- Google Colab
- Kaggle
File size: 440 Bytes
f86f1fd | 1 2 | {"architecture": "Perceiver", "input_channels": 16, "sat_channels": 12, "forecast_steps": 6, "lr": 0.002, "input_size": 32, "max_frequency": 16.0, "depth": 6, "num_latents": 256, "cross_heads": 1, "latent_heads": 8, "cross_dim_heads": 8, "latent_dim": 256, "weight_tie_layers": false, "self_per_cross_attention": 2, "dim": 32, "logits_dim": null, "queries_dim": 32, "latent_dim_heads": 64, "visualize": true, "preprocessor_type": "metnet"}
|