| { | |
| "architectures": [ | |
| "FlashSTU" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "config.FlashSTUConfig", | |
| "AutoModel": "model.FlashSTU" | |
| }, | |
| "bias": false, | |
| "bsz": 1, | |
| "dropout": 0.0, | |
| "hidden_act": "swish", | |
| "hidden_size": 1536, | |
| "intermediate_size": 18432, | |
| "model_type": "FlashSTU", | |
| "n_embd": 1536, | |
| "n_heads": 8, | |
| "n_layers": 26, | |
| "num_eigh": 24, | |
| "seq_len": 8192, | |
| "softcap": 50.0, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.44.0", | |
| "use_approx": true, | |
| "use_attn": true, | |
| "use_flash_fft": true, | |
| "use_hankel_L": false, | |
| "vocab_size": 200064, | |
| "window_size": 1024 | |
| } | |