{ "bos_token_id" : 0, "chunk_size" : 256, "conv_kernel" : 4, "eos_token_id" : 0, "expand" : 2, "head_dim" : 64, "hidden_act" : "silu", "hidden_size" : 768, "initializer_range" : 0.10000000000000001, "layer_norm_epsilon" : 1.0000000000000001e-05, "model_type" : "mamba2", "n_groups" : 1, "num_heads" : 24, "num_hidden_layers" : 24, "pad_token_id" : 0, "quantization" : { "bits" : 4, "group_size" : 64, "mode" : "affine" }, "quantization_config" : { "bits" : 4, "group_size" : 64, "mode" : "affine" }, "rescale_prenorm_residual" : 0, "residual_in_fp32" : 1, "rms_norm" : 1, "state_size" : 128, "tie_word_embeddings" : 1, "time_step_floor" : 0.0001, "time_step_limit" : [ 0, 1e+308 ], "time_step_max" : 0.10000000000000001, "time_step_min" : 0.001, "time_step_rank" : 256, "transformers_version" : "4.45.0", "use_bias" : 0, "use_cache" : 1, "use_conv_bias" : 1, "vocab_size" : 50288 }