| { |
| "model_type": "neuroblast", |
| "architectures": [ |
| "NeuroBLASTForCausalLM" |
| ], |
| "auto_map": { |
| "AutoConfig": "configuration_neuroblast.NeuroBLASTConfig", |
| "AutoModelForCausalLM": "modeling_neuroblast.NeuroBLASTForCausalLM" |
| }, |
| "association_cortex_hidden_size": 1024, |
| "association_gradient_scale": 0.8, |
| "cross_attention_gradient_scale": 0.8, |
| "dropout": 0.0, |
| "eos_token_id": [ |
| 2, |
| 32009 |
| ], |
| "gradient_scaling_enabled": true, |
| "hidden_size": 1536, |
| "intermediate_size": 3072, |
| "kv_dim": 12288, |
| "max_position_embeddings": 131072, |
| "motor_cortex_hidden_size": 1024, |
| "norm_topk_prob": true, |
| "num_association_cortex_layers": 3, |
| "num_experts": null, |
| "num_experts_per_tok": 4, |
| "num_heads": 32, |
| "num_motor_cortex_layers": 3, |
| "num_sensory_cortex_layers": 3, |
| "pad_token_id": 32000, |
| "precomputed_total_layers": 18, |
| "rope_scaling": { |
| "factor": 32.0, |
| "high_freq_factor": 4.0, |
| "low_freq_factor": 1.0, |
| "original_max_position_embeddings": 8192, |
| "rope_type": "llama3" |
| }, |
| "sensory_cortex_hidden_size": 1024, |
| "sensory_gradient_scale": 0.95, |
| "tie_word_embeddings": false, |
| "torch_dtype": "bfloat16", |
| "transformers_version": "4.53.2", |
| "use_zero_memory": true, |
| "vocab_size": 32024, |
| "zero_memory_layers": [ |
| 15 |
| ] |
| } |
|
|
