{
  "A_init_range": [
    1,
    16
  ],
  "D_has_hdim": false,
  "architectures": [
    "Mamba2ForCausalLM"
  ],
  "bos_token_id": 1,
  "chunk_size": 256,
  "conv_init": null,
  "conv_kernel": 4,
  "dt_init_floor": 0.0001,
  "dt_limit": [
    0.0,
    {
      "__float__": "Infinity"
    }
  ],
  "dt_max": 0.1,
  "dt_min": 0.001,
  "dtype": "bfloat16",
  "eos_token_id": 2,
  "expand": 2,
  "fuse_cross_entropy": true,
  "fuse_linear_cross_entropy": false,
  "fuse_norm": true,
  "head_dim": 64,
  "head_init_design": null,
  "hidden_act": "silu",
  "hidden_size": 512,
  "initializer_range": 0.02,
  "max_position_embeddings": 4096,
  "model_type": "mamba2",
  "n_groups": 1,
  "norm_before_gate": false,
  "norm_eps": 1e-05,
  "num_heads": 16,
  "num_hidden_layers": 42,
  "pad_token_id": 0,
  "rescale_prenorm_residual": true,
  "residual_in_fp32": true,
  "rms_norm": true,
  "rmsnorm": true,
  "state_size": 128,
  "tie_word_embeddings": false,
  "time_step_floor": 0.0001,
  "time_step_max": 0.1,
  "time_step_min": 0.001,
  "time_step_rank": 128,
  "transformers_version": "5.9.0",
  "use_bias": false,
  "use_cache": true,
  "use_conv_bias": true,
  "use_l2warp": false,
  "vocab_size": 32000
}