config.json

{
  "architectures": [
    "ChemQ3MTPForCausalLM"
  ],
  "attention_dropout": 0.1,
  "bos_token_id": 0,
  "dtype": "float32",
  "entropy_controller_config": {
    "adaptation_rate": 0.01,
    "max_entropy": 3.0,
    "min_entropy": 0.5,
    "target_entropy": 1.5
  },
  "eos_token_id": 1,
  "head_dim": 64,
  "hidden_act": "silu",
  "hidden_size": 320,
  "horizon_weights": [
    1.0,
    0.9,
    0.81
  ],
  "initializer_range": 0.02,
  "intermediate_size": 1280,
  "layer_types": [
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention"
  ],
  "max_position_embeddings": 128,
  "max_window_layers": 28,
  "model_type": "chemq3_mtp",
  "num_attention_heads": 4,
  "num_future_tokens": 3,
  "num_hidden_layers": 6,
  "num_key_value_heads": 2,
  "pad_token_id": 2,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 10000.0,
  "sliding_window": null,
  "tie_word_embeddings": true,
  "transformers_version": "4.56.1",
  "use_cache": false,
  "use_mtp_training": true,
  "use_sliding_window": false,
  "vocab_size": 782
}