{
  "architectures": [
    "SmolLM3ForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": 128012,
  "hidden_act": "silu",
  "hidden_size": 2048,
  "initializer_range": 0.02,
  "intermediate_size": 11008,
  "layer_types": [
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention"
  ],
  "max_position_embeddings": 65536,
  "max_window_layers": 28,
  "mlp_bias": false,
  "model_type": "smollm3",
  "no_rope_layer_interval": 4,
  "no_rope_layers": [
    1,
    1,
    1,
    0,
    1,
    1,
    1,
    0,
    1,
    1,
    1,
    0,
    1,
    1,
    1,
    0,
    1,
    1,
    1,
    0,
    1,
    1,
    1,
    0,
    1,
    1,
    1,
    0,
    1,
    1,
    1,
    0,
    1,
    1,
    1,
    0
  ],
  "num_attention_heads": 16,
  "num_hidden_layers": 36,
  "num_key_value_heads": 4,
  "pad_token_id": 128004,
  "pretraining_tp": 2,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 5000000.0,
  "sliding_window": null,
  "tie_word_embeddings": true,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.54.0.dev0",
  "use_cache": false,
  "use_sliding_window": false,
  "vocab_size": 128256,
  "quantization": {
    "group_size": 64,
    "bits": 4,
    "mode": "affine",
    "model.layers.0.mlp.down_proj": {
      "bits": 6,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.0.self_attn.k_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.0.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.0.self_attn.q_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.0.self_attn.v_proj": {
      "bits": 6,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.1.mlp.down_proj": {
      "bits": 6,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.1.self_attn.k_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.1.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.1.self_attn.q_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.1.self_attn.v_proj": {
      "bits": 6,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.10.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.10.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.11.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.11.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.12.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.12.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.13.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.13.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.14.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.14.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.15.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.15.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.16.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.16.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.17.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.17.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.18.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.18.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.19.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.19.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.2.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.2.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.20.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.20.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.21.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.21.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.22.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.22.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.23.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.23.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.24.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.24.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.25.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.25.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.26.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.26.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.27.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.28.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.3.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.4.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.29.self_attn.k_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.29.self_attn.v_proj": {
      "bits": 6,
      "group_size": 64,
      "mode": "affine"
    }
  },
  "quantization_config": {
    "group_size": 64,
    "bits": 4,
    "mode": "affine",
    "model.layers.0.mlp.down_proj": {
      "bits": 6,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.0.self_attn.k_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.0.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.0.self_attn.q_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.0.self_attn.v_proj": {
      "bits": 6,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.1.mlp.down_proj": {
      "bits": 6,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.1.self_attn.k_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.1.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.1.self_attn.q_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.1.self_attn.v_proj": {
      "bits": 6,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.10.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.10.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.11.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.11.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.12.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.12.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.13.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.13.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.14.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.14.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.15.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.15.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.16.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.16.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.17.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.17.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.18.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.18.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.19.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.19.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.2.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.2.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.20.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.20.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.21.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.21.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.22.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.22.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.23.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.23.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.24.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.24.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.25.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.25.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.26.mlp.down_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.26.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.27.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.28.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.3.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.4.self_attn.o_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.29.self_attn.k_proj": {
      "bits": 5,
      "group_size": 64,
      "mode": "affine"
    },
    "model.layers.29.self_attn.v_proj": {
      "bits": 6,
      "group_size": 64,
      "mode": "affine"
    }
  }
}