{ "_name_or_path": "meta-llama/Llama-2-7b-hf", "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "auto_map": { "AutoModelForCausalLM": "qpal_modelling_llama.QPalLlamaForCausalLM" }, "bos_token_id": 1, "eos_token_id": 2, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 11008, "max_position_embeddings": 4096, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 32, "pretraining_tp": 1, "qpal_quant_config": { "modules": { "model.layers.0.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 4, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.0.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 3, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.0.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 3, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.0.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 3, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.0.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 3, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.0.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 3, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.0.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 3, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.1.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.1.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.1.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.1.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 3, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.1.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 3, 4 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.1.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 3, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.1.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.10.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.10.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.10.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.10.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.10.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.10.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.10.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 8, 9 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 10 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.11.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.11.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.11.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.11.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.11.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.11.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.11.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 9, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 10 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.12.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": [ 7, 8 ], "L": 16, "V": 2, "bias": false, "in_features": 11008, "in_part": [ 5504, 5504 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.12.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.12.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.12.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.12.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.12.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.12.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 8, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.13.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 11008, "in_part": [ 5504, 5504 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.13.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.13.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.13.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.13.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 7, 8 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.13.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.13.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 9, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 10 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.14.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 11008, "in_part": [ 5504, 5504 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.14.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.14.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.14.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 4, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.14.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.14.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.14.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 8, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.15.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 11008, "in_part": [ 5504, 5504 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.15.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.15.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.15.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.15.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 7, 8 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.15.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.15.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 8, 9 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 10 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.16.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.16.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.16.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.16.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.16.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.16.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.16.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 9, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 10 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.17.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.17.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.17.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.17.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 3, 4 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.17.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.17.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.17.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.18.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.18.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.18.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.18.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.18.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.18.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.18.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 10, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 11 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.19.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.19.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 7, 8 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.19.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.19.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.19.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.19.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 4, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.19.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 7, 8 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.2.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 11008, "in_part": [ 5504, 5504 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.2.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.2.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.2.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 4, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.2.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.2.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.2.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 9, 10 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 11 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.20.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.20.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.20.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.20.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 3, 4 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.20.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.20.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.20.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.21.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": [ 7, 8 ], "L": 16, "V": 2, "bias": false, "in_features": 11008, "in_part": [ 5504, 5504 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.21.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.21.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.21.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.21.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.21.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 4, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.21.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 9, 10 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 11 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.22.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.22.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.22.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.22.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.22.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.22.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.22.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 7, 8 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.23.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.23.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.23.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.23.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 4, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.23.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.23.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.23.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 8, 9 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 10 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.24.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.24.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.24.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.24.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.24.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.24.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.24.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 8, 9 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 10 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.25.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.25.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.25.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.25.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.25.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.25.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.25.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 8, 9 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 10 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.26.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 11008, "in_part": [ 5504, 5504 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.26.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.26.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.26.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.26.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.26.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.26.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.27.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.27.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.27.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.27.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.27.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.27.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.27.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 8, 9 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 10 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.28.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 11008, "in_part": [ 5504, 5504 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.28.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.28.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.28.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.28.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.28.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.28.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 7, 8 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.29.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 11008, "in_part": [ 5504, 5504 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.29.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.29.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.29.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 3, 4 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.29.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.29.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 3, 4 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.29.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 7, 8 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.3.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.3.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.3.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.3.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.3.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.3.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.3.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 9, 10 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 11 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.30.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.30.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.30.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.30.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.30.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.30.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 4, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.30.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 8, 9 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 10 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.31.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.31.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.31.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.31.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 4, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.31.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.31.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 4, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.31.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.4.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.4.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.4.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.4.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 4, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.4.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.4.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.4.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 9, 10 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 11 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.5.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 11008, "in_part": [ 5504, 5504 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.5.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.5.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.5.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 3, 4 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.5.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.5.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.5.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 10, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 11 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.6.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.6.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.6.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.6.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.6.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.6.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.6.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 9, 10 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 11 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.7.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.7.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.7.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.7.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.7.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.7.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.7.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 9, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 10 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.8.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": [ 6, 7 ], "L": 16, "V": 2, "bias": false, "in_features": 11008, "in_part": [ 5504, 5504 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.8.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.8.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.8.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.8.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 7, 8 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.8.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 4, 5 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.8.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.9.mlp.down_proj": { "bias": false, "dtype": "float32", "hadU": 11008, "hadV": 4096, "in_features": 11008, "linear": { "KV": 7, "L": 16, "V": 2, "bias": false, "in_features": 11008, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.9.mlp.gate_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": 6, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.9.mlp.up_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 11008, "in_features": 4096, "linear": { "KV": [ 5, 6 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 11008, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 11008, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.9.self_attn.k_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": [ 3, 4 ], "L": 16, "V": 2, "bias": false, "in_features": 4096, "in_part": [ 2048, 2048 ], "linear_cls": "CombtLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.9.self_attn.o_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 5, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.9.self_attn.q_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 4, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 9 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 }, "model.layers.9.self_attn.v_proj": { "bias": false, "dtype": "float32", "hadU": 4096, "hadV": 4096, "in_features": 4096, "linear": { "KV": 10, "L": 16, "V": 2, "bias": false, "in_features": 4096, "linear_cls": "QTIPLinearTCQ", "linear_dtype": "float32", "out_features": 4096, "td_x": 16, "td_y": 16, "tlut_bits": 11 }, "module_type": "IncoherentLinear", "out_features": 4096, "rot_info": "skip_r", "scale": 32.0 } } }, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 10000.0, "tie_word_embeddings": false, "torch_dtype": "float16", "transformers_version": "4.45.2", "use_cache": true, "vocab_size": 32000 }