Instructions to use joey00072/q_sparse_0.7 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use joey00072/q_sparse_0.7 with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("joey00072/q_sparse_0.7", device_map="auto") - Notebooks
- Google Colab
- Kaggle
| { | |
| "activation": "relu_squared", | |
| "bias": false, | |
| "d_model": 512, | |
| "dropout": 0.2, | |
| "hidden_dim": 2048, | |
| "mlp": "GLU", | |
| "num_heads": 32, | |
| "num_kv_heads": 0, | |
| "num_layers": 32, | |
| "seq_len": 256, | |
| "vocab_size": 50257, | |
| "weight_tying": true | |
| } |