{ "alpha_pattern": { "model.layers.1.mlp.down_proj": 256, "model.layers.1.mlp.gate_proj": 256, "model.layers.1.mlp.up_proj": 256, "model.layers.1.self_attn.k_proj": 256, "model.layers.1.self_attn.o_proj": 256, "model.layers.1.self_attn.q_proj": 256, "model.layers.1.self_attn.v_proj": 256, "model.layers.11.mlp.down_proj": 64, "model.layers.11.mlp.gate_proj": 64, "model.layers.11.mlp.up_proj": 64, "model.layers.11.self_attn.k_proj": 64, "model.layers.11.self_attn.o_proj": 64, "model.layers.11.self_attn.q_proj": 64, "model.layers.11.self_attn.v_proj": 64, "model.layers.14.mlp.down_proj": 64, "model.layers.14.mlp.gate_proj": 64, "model.layers.14.mlp.up_proj": 64, "model.layers.14.self_attn.k_proj": 64, "model.layers.14.self_attn.o_proj": 64, "model.layers.14.self_attn.q_proj": 64, "model.layers.14.self_attn.v_proj": 64, "model.layers.16.mlp.down_proj": 128, "model.layers.16.mlp.gate_proj": 128, "model.layers.16.mlp.up_proj": 128, "model.layers.16.self_attn.k_proj": 128, "model.layers.16.self_attn.o_proj": 128, "model.layers.16.self_attn.q_proj": 128, "model.layers.16.self_attn.v_proj": 128, "model.layers.2.mlp.down_proj": 128, "model.layers.2.mlp.gate_proj": 128, "model.layers.2.mlp.up_proj": 128, "model.layers.2.self_attn.k_proj": 128, "model.layers.2.self_attn.o_proj": 128, "model.layers.2.self_attn.q_proj": 128, "model.layers.2.self_attn.v_proj": 128, "model.layers.21.mlp.down_proj": 128, "model.layers.21.mlp.gate_proj": 128, "model.layers.21.mlp.up_proj": 128, "model.layers.21.self_attn.k_proj": 128, "model.layers.21.self_attn.o_proj": 128, "model.layers.21.self_attn.q_proj": 128, "model.layers.21.self_attn.v_proj": 128, "model.layers.22.mlp.down_proj": 256, "model.layers.22.mlp.gate_proj": 256, "model.layers.22.mlp.up_proj": 256, "model.layers.22.self_attn.k_proj": 256, "model.layers.22.self_attn.o_proj": 256, "model.layers.22.self_attn.q_proj": 256, "model.layers.22.self_attn.v_proj": 256, "model.layers.25.mlp.down_proj": 128, "model.layers.25.mlp.gate_proj": 128, "model.layers.25.mlp.up_proj": 128, "model.layers.25.self_attn.k_proj": 128, "model.layers.25.self_attn.o_proj": 128, "model.layers.25.self_attn.q_proj": 128, "model.layers.25.self_attn.v_proj": 128, "model.layers.27.mlp.down_proj": 256, "model.layers.27.mlp.gate_proj": 256, "model.layers.27.mlp.up_proj": 256, "model.layers.27.self_attn.k_proj": 256, "model.layers.27.self_attn.o_proj": 256, "model.layers.27.self_attn.q_proj": 256, "model.layers.27.self_attn.v_proj": 256, "model.layers.3.mlp.down_proj": 256, "model.layers.3.mlp.gate_proj": 256, "model.layers.3.mlp.up_proj": 256, "model.layers.3.self_attn.k_proj": 256, "model.layers.3.self_attn.o_proj": 256, "model.layers.3.self_attn.q_proj": 256, "model.layers.3.self_attn.v_proj": 256, "model.layers.6.mlp.down_proj": 64, "model.layers.6.mlp.gate_proj": 64, "model.layers.6.mlp.up_proj": 64, "model.layers.6.self_attn.k_proj": 64, "model.layers.6.self_attn.o_proj": 64, "model.layers.6.self_attn.q_proj": 64, "model.layers.6.self_attn.v_proj": 64, "model.layers.7.mlp.down_proj": 64, "model.layers.7.mlp.gate_proj": 64, "model.layers.7.mlp.up_proj": 64, "model.layers.7.self_attn.k_proj": 64, "model.layers.7.self_attn.o_proj": 64, "model.layers.7.self_attn.q_proj": 64, "model.layers.7.self_attn.v_proj": 64 }, "auto_mapping": null, "base_model_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct", "bias": "none", "corda_config": null, "eva_config": null, "exclude_modules": null, "fan_in_fan_out": false, "inference_mode": true, "init_lora_weights": true, "layer_replication": null, "layers_pattern": null, "layers_to_transform": null, "loftq_config": {}, "lora_alpha": 64, "lora_bias": false, "lora_dropout": 0.05, "megatron_config": null, "megatron_core": "megatron.core", "modules_to_save": null, "peft_type": "LORA", "qalora_group_size": 16, "r": 32, "rank_pattern": { "model.layers.1.mlp.down_proj": 128, "model.layers.1.mlp.gate_proj": 128, "model.layers.1.mlp.up_proj": 128, "model.layers.1.self_attn.k_proj": 128, "model.layers.1.self_attn.o_proj": 128, "model.layers.1.self_attn.q_proj": 128, "model.layers.1.self_attn.v_proj": 128, "model.layers.11.mlp.down_proj": 32, "model.layers.11.mlp.gate_proj": 32, "model.layers.11.mlp.up_proj": 32, "model.layers.11.self_attn.k_proj": 32, "model.layers.11.self_attn.o_proj": 32, "model.layers.11.self_attn.q_proj": 32, "model.layers.11.self_attn.v_proj": 32, "model.layers.14.mlp.down_proj": 32, "model.layers.14.mlp.gate_proj": 32, "model.layers.14.mlp.up_proj": 32, "model.layers.14.self_attn.k_proj": 32, "model.layers.14.self_attn.o_proj": 32, "model.layers.14.self_attn.q_proj": 32, "model.layers.14.self_attn.v_proj": 32, "model.layers.16.mlp.down_proj": 64, "model.layers.16.mlp.gate_proj": 64, "model.layers.16.mlp.up_proj": 64, "model.layers.16.self_attn.k_proj": 64, "model.layers.16.self_attn.o_proj": 64, "model.layers.16.self_attn.q_proj": 64, "model.layers.16.self_attn.v_proj": 64, "model.layers.2.mlp.down_proj": 64, "model.layers.2.mlp.gate_proj": 64, "model.layers.2.mlp.up_proj": 64, "model.layers.2.self_attn.k_proj": 64, "model.layers.2.self_attn.o_proj": 64, "model.layers.2.self_attn.q_proj": 64, "model.layers.2.self_attn.v_proj": 64, "model.layers.21.mlp.down_proj": 64, "model.layers.21.mlp.gate_proj": 64, "model.layers.21.mlp.up_proj": 64, "model.layers.21.self_attn.k_proj": 64, "model.layers.21.self_attn.o_proj": 64, "model.layers.21.self_attn.q_proj": 64, "model.layers.21.self_attn.v_proj": 64, "model.layers.22.mlp.down_proj": 128, "model.layers.22.mlp.gate_proj": 128, "model.layers.22.mlp.up_proj": 128, "model.layers.22.self_attn.k_proj": 128, "model.layers.22.self_attn.o_proj": 128, "model.layers.22.self_attn.q_proj": 128, "model.layers.22.self_attn.v_proj": 128, "model.layers.25.mlp.down_proj": 64, "model.layers.25.mlp.gate_proj": 64, "model.layers.25.mlp.up_proj": 64, "model.layers.25.self_attn.k_proj": 64, "model.layers.25.self_attn.o_proj": 64, "model.layers.25.self_attn.q_proj": 64, "model.layers.25.self_attn.v_proj": 64, "model.layers.27.mlp.down_proj": 128, "model.layers.27.mlp.gate_proj": 128, "model.layers.27.mlp.up_proj": 128, "model.layers.27.self_attn.k_proj": 128, "model.layers.27.self_attn.o_proj": 128, "model.layers.27.self_attn.q_proj": 128, "model.layers.27.self_attn.v_proj": 128, "model.layers.3.mlp.down_proj": 128, "model.layers.3.mlp.gate_proj": 128, "model.layers.3.mlp.up_proj": 128, "model.layers.3.self_attn.k_proj": 128, "model.layers.3.self_attn.o_proj": 128, "model.layers.3.self_attn.q_proj": 128, "model.layers.3.self_attn.v_proj": 128, "model.layers.6.mlp.down_proj": 32, "model.layers.6.mlp.gate_proj": 32, "model.layers.6.mlp.up_proj": 32, "model.layers.6.self_attn.k_proj": 32, "model.layers.6.self_attn.o_proj": 32, "model.layers.6.self_attn.q_proj": 32, "model.layers.6.self_attn.v_proj": 32, "model.layers.7.mlp.down_proj": 32, "model.layers.7.mlp.gate_proj": 32, "model.layers.7.mlp.up_proj": 32, "model.layers.7.self_attn.k_proj": 32, "model.layers.7.self_attn.o_proj": 32, "model.layers.7.self_attn.q_proj": 32, "model.layers.7.self_attn.v_proj": 32 }, "revision": null, "target_modules": [ "model.layers.27.self_attn.k_proj", "model.layers.6.mlp.up_proj", "model.layers.22.self_attn.k_proj", "model.layers.6.self_attn.o_proj", "model.layers.16.self_attn.k_proj", "model.layers.1.self_attn.v_proj", "model.layers.3.mlp.down_proj", "model.layers.22.mlp.gate_proj", "model.layers.2.self_attn.k_proj", "model.layers.7.mlp.down_proj", "model.layers.11.mlp.gate_proj", "model.layers.16.self_attn.v_proj", "model.layers.3.mlp.up_proj", "model.layers.22.mlp.up_proj", "model.layers.3.self_attn.v_proj", "model.layers.21.self_attn.q_proj", "model.layers.21.self_attn.o_proj", "model.layers.16.mlp.down_proj", "model.layers.11.self_attn.o_proj", "model.layers.1.self_attn.q_proj", "model.layers.14.mlp.up_proj", "model.layers.25.mlp.up_proj", "model.layers.1.mlp.gate_proj", "model.layers.14.mlp.down_proj", "model.layers.6.self_attn.k_proj", "model.layers.3.self_attn.k_proj", "model.layers.27.self_attn.q_proj", "model.layers.25.self_attn.k_proj", "model.layers.3.self_attn.q_proj", "model.layers.27.mlp.up_proj", "model.layers.6.self_attn.v_proj", "model.layers.27.mlp.down_proj", "model.layers.27.mlp.gate_proj", "model.layers.11.mlp.down_proj", "model.layers.21.mlp.down_proj", "model.layers.1.mlp.down_proj", "model.layers.7.mlp.up_proj", "model.layers.3.self_attn.o_proj", "model.layers.22.mlp.down_proj", "model.layers.1.self_attn.o_proj", "model.layers.11.mlp.up_proj", "model.layers.25.mlp.down_proj", "model.layers.11.self_attn.k_proj", "model.layers.11.self_attn.q_proj", "model.layers.2.mlp.down_proj", "model.layers.14.self_attn.q_proj", "model.layers.22.self_attn.o_proj", "model.layers.27.self_attn.v_proj", "model.layers.21.self_attn.k_proj", "model.layers.25.self_attn.v_proj", "model.layers.22.self_attn.v_proj", "model.layers.25.mlp.gate_proj", "model.layers.6.mlp.gate_proj", "model.layers.27.self_attn.o_proj", "model.layers.7.self_attn.q_proj", "model.layers.1.self_attn.k_proj", "model.layers.2.mlp.gate_proj", "model.layers.14.self_attn.k_proj", "model.layers.2.mlp.up_proj", "model.layers.21.mlp.gate_proj", "model.layers.6.mlp.down_proj", "model.layers.16.self_attn.q_proj", "model.layers.14.self_attn.v_proj", "model.layers.25.self_attn.o_proj", "model.layers.1.mlp.up_proj", "model.layers.6.self_attn.q_proj", "model.layers.14.mlp.gate_proj", "model.layers.11.self_attn.v_proj", "model.layers.7.mlp.gate_proj", "model.layers.2.self_attn.v_proj", "model.layers.7.self_attn.o_proj", "model.layers.16.mlp.up_proj", "model.layers.16.mlp.gate_proj", "model.layers.21.self_attn.v_proj", "model.layers.21.mlp.up_proj", "model.layers.22.self_attn.q_proj", "model.layers.7.self_attn.k_proj", "model.layers.2.self_attn.q_proj", "model.layers.2.self_attn.o_proj", "model.layers.25.self_attn.q_proj", "model.layers.7.self_attn.v_proj", "model.layers.14.self_attn.o_proj", "model.layers.16.self_attn.o_proj", "model.layers.3.mlp.gate_proj" ], "target_parameters": null, "task_type": "CAUSAL_LM", "trainable_token_indices": null, "use_dora": false, "use_qalora": false, "use_rslora": false }