| { | |
| "model_type": "custom_gpt", | |
| "architectures": ["GPT"], | |
| "auto_map": { | |
| "AutoConfig": "modeling_my_model.GPTConfig", | |
| "AutoModelForCausalLM": "modeling_my_model.GPT" | |
| }, | |
| "block_size": 1024, | |
| "vocab_size": 50304, | |
| "n_layer": 24, | |
| "n_head": 16, | |
| "n_embd": 1024, | |
| "dropout": 0.0, | |
| "bias": false, | |
| "hc_num_streams": 1, | |
| "hc_num_fracs": 1, | |
| "hc_disable": true, | |
| "mhc": false, | |
| "sinkhorn_iters": 10, | |
| "sinkhorn_tau": 0.05, | |
| "mhc_h_res_proj": "sinkhorn", | |
| "ns_steps": 5, | |
| "ns_eps": 1e-7, | |
| "ns_coeffs": [3.0, -3.2, 1.2] | |
| } | |