Instructions to use CodeIsAbstract/HybridModelScratch_lightn_Conti with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use CodeIsAbstract/HybridModelScratch_lightn_Conti with Transformers:
# Load model directly from transformers import HybridFourierLM model = HybridFourierLM.from_pretrained("CodeIsAbstract/HybridModelScratch_lightn_Conti", device_map="auto") - Notebooks
- Google Colab
- Kaggle
| { | |
| "architectures": [ | |
| "HybridFourierLM" | |
| ], | |
| "bos_token_id": 0, | |
| "dropout": 0.05, | |
| "dtype": "float32", | |
| "eos_token_id": 0, | |
| "latent_dim": 768, | |
| "layer_types": [ | |
| "linear", | |
| "linear", | |
| "linear", | |
| "softmax", | |
| "linear", | |
| "linear", | |
| "linear", | |
| "softmax", | |
| "linear", | |
| "linear", | |
| "linear", | |
| "softmax" | |
| ], | |
| "model_type": "hybrid_fourier_lm", | |
| "num_layers": 12, | |
| "num_modes": 64, | |
| "pad_token_id": 1, | |
| "tie_word_embeddings": true, | |
| "time_scale": 128.0, | |
| "transformers_version": "5.13.0", | |
| "use_cache": false, | |
| "vocab_size": 50277 | |
| } | |