{ "architectures": [ "NemotronDenseForCausalLM" ], "auto_map": { "AutoConfig": "modeling_nemotron_dense.NemotronDenseConfig", "AutoModelForCausalLM": "modeling_nemotron_dense.NemotronDenseForCausalLM" }, "dtype": "float32", "head_dim": 128, "hidden_act": "relu2", "hidden_size": 2048, "intermediate_size": 9216, "max_position_embeddings": 131072, "model_type": "nemotron_dense", "norm_eps": 1e-05, "num_attention_heads": 16, "num_hidden_layers": 28, "num_key_value_heads": 8, "rope_parameters": { "partial_rotary_factor": 1.0, "rope_theta": 100000000, "rope_type": "default" }, "tie_word_embeddings": false, "transformers_version": "5.4.0", "vocab_size": 131072 }