| { | |
| "architecture": "AMOR", | |
| "format": "custom-pytorch-amor-v1", | |
| "model_kwargs": { | |
| "d_model": 1024, | |
| "n_layer": 24, | |
| "d_ff": 1984, | |
| "head_dim": 64, | |
| "d_state": 128, | |
| "d_conv": 4, | |
| "expand": 2, | |
| "n_groups": 1, | |
| "backbone": "gdn", | |
| "n_amor_blocks": 3, | |
| "residual_mode": "classic", | |
| "vocab_size": 128256, | |
| "n_heads": 16, | |
| "attention_mode": "full_with_mask", | |
| "max_seq_len": 4096, | |
| "init_alpha": 0.0, | |
| "init_threshold": 0.3, | |
| "gate_offset": 0.03, | |
| "offset_mode": "adaptive", | |
| "offset_k": 0.2 | |
| }, | |
| "tokenizer": "teknium/Llama-3.1-AlternateTokenizer", | |
| "tokenizer_revision": "5590cc63198b0f111f4289daf006da0a12044173", | |
| "training_context_length": 3072, | |
| "tied_weights": { | |
| "lm_head.weight": "embed.weight" | |
| }, | |
| "note": "Use load_model.py; not compatible with Transformers AutoModel or GGUF.", | |
| "tokenizer_revision_provenance": "Pinned at release packaging; original training-time revision not recorded." | |
| } | |