| { |
| "alora_invocation_tokens": null, |
| "alpha_pattern": {}, |
| "arrow_config": null, |
| "auto_mapping": null, |
| "base_model_name_or_path": "google/gemma-4-E2B-it", |
| "bias": "none", |
| "corda_config": null, |
| "ensure_weight_tying": true, |
| "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": 16, |
| "lora_bias": false, |
| "lora_dropout": 0.05, |
| "megatron_config": null, |
| "megatron_core": "megatron.core", |
| "modules_to_save": [ |
| "lm_head", |
| "embed_tokens" |
| ], |
| "peft_type": "LORA", |
| "peft_version": "0.18.1", |
| "qalora_group_size": 16, |
| "r": 8, |
| "rank_pattern": {}, |
| "revision": null, |
| "target_modules": [ |
| "per_layer_projection", |
| "17.self_attn.q_proj", |
| "language_model.layers.10.self_attn.k_proj", |
| "language_model.layers.9.mlp.gate_proj", |
| "language_model.layers.14.mlp.gate_proj", |
| "language_model.layers.1.self_attn.q_proj", |
| "26.mlp.down_proj", |
| "language_model.layers.14.self_attn.o_proj", |
| "language_model.layers.5.mlp.up_proj", |
| "input_proj", |
| "17.mlp.down_proj", |
| "25.self_attn.o_proj", |
| "language_model.layers.3.self_attn.v_proj", |
| "22.mlp.gate_proj", |
| "language_model.layers.9.self_attn.o_proj", |
| "33.mlp.down_proj", |
| "language_model.layers.1.mlp.gate_proj", |
| "language_model.layers.0.mlp.up_proj", |
| "language_model.layers.7.self_attn.v_proj", |
| "linear", |
| "28.self_attn.q_proj", |
| "31.mlp.up_proj", |
| "language_model.layers.3.mlp.down_proj", |
| "24.self_attn.q_proj", |
| "23.mlp.gate_proj", |
| "25.self_attn.q_proj", |
| "language_model.layers.13.self_attn.k_proj", |
| "language_model.layers.15.self_attn.o_proj", |
| "33.self_attn.o_proj", |
| "language_model.layers.9.self_attn.q_proj", |
| "31.self_attn.o_proj", |
| "language_model.layers.8.self_attn.k_proj", |
| "language_model.layers.14.mlp.down_proj", |
| "19.self_attn.o_proj", |
| "21.mlp.gate_proj", |
| "27.self_attn.q_proj", |
| "language_model.layers.8.mlp.gate_proj", |
| "language_model.layers.5.self_attn.o_proj", |
| "24.mlp.up_proj", |
| "34.self_attn.q_proj", |
| "language_model.layers.7.mlp.up_proj", |
| "21.self_attn.q_proj", |
| "18.mlp.up_proj", |
| "16.mlp.down_proj", |
| "30.mlp.down_proj", |
| "16.mlp.gate_proj", |
| "language_model.layers.12.self_attn.v_proj", |
| "language_model.layers.0.self_attn.v_proj", |
| "language_model.layers.10.mlp.up_proj", |
| "language_model.layers.14.mlp.up_proj", |
| "language_model.layers.15.mlp.up_proj", |
| "language_model.layers.0.mlp.down_proj", |
| "language_model.layers.10.mlp.down_proj", |
| "20.mlp.up_proj", |
| "26.mlp.gate_proj", |
| "24.mlp.gate_proj", |
| "language_model.layers.0.self_attn.k_proj", |
| "17.mlp.up_proj", |
| "20.self_attn.q_proj", |
| "language_model.layers.10.self_attn.o_proj", |
| "language_model.layers.11.mlp.gate_proj", |
| "16.self_attn.o_proj", |
| "embedding_projection", |
| "34.mlp.gate_proj", |
| "language_model.layers.0.self_attn.o_proj", |
| "language_model.layers.3.self_attn.k_proj", |
| "output_proj", |
| "23.mlp.down_proj", |
| "28.mlp.up_proj", |
| "language_model.layers.4.self_attn.v_proj", |
| "30.self_attn.q_proj", |
| "language_model.layers.13.self_attn.o_proj", |
| "28.mlp.gate_proj", |
| "language_model.layers.12.self_attn.q_proj", |
| "language_model.layers.3.self_attn.q_proj", |
| "language_model.layers.6.self_attn.v_proj", |
| "21.mlp.up_proj", |
| "17.mlp.gate_proj", |
| "29.mlp.down_proj", |
| "20.self_attn.o_proj", |
| "32.self_attn.q_proj", |
| "26.self_attn.q_proj", |
| "16.self_attn.q_proj", |
| "language_model.layers.1.mlp.down_proj", |
| "30.mlp.gate_proj", |
| "28.mlp.down_proj", |
| "per_layer_input_gate", |
| "language_model.layers.2.self_attn.q_proj", |
| "language_model.layers.4.self_attn.k_proj", |
| "34.mlp.down_proj", |
| "33.mlp.gate_proj", |
| "relative_k_proj", |
| "language_model.layers.9.mlp.up_proj", |
| "31.mlp.gate_proj", |
| "language_model.layers.13.mlp.down_proj", |
| "language_model.layers.15.self_attn.q_proj", |
| "language_model.layers.7.self_attn.q_proj", |
| "language_model.layers.4.mlp.down_proj", |
| "28.self_attn.o_proj", |
| "29.self_attn.q_proj", |
| "21.self_attn.o_proj", |
| "language_model.layers.14.self_attn.q_proj", |
| "language_model.layers.3.mlp.gate_proj", |
| "language_model.layers.8.mlp.down_proj", |
| "34.self_attn.o_proj", |
| "language_model.layers.5.self_attn.k_proj", |
| "language_model.layers.12.mlp.gate_proj", |
| "language_model.layers.7.self_attn.o_proj", |
| "27.mlp.down_proj", |
| "18.self_attn.o_proj", |
| "language_model.layers.2.self_attn.k_proj", |
| "language_model.layers.13.mlp.up_proj", |
| "25.mlp.up_proj", |
| "language_model.layers.2.self_attn.o_proj", |
| "language_model.layers.7.mlp.down_proj", |
| "language_model.layers.13.self_attn.v_proj", |
| "language_model.layers.9.mlp.down_proj", |
| "27.mlp.gate_proj", |
| "language_model.layers.1.self_attn.k_proj", |
| "23.self_attn.o_proj", |
| "per_layer_model_projection", |
| "language_model.layers.15.mlp.down_proj", |
| "language_model.layers.12.self_attn.o_proj", |
| "22.self_attn.o_proj", |
| "language_model.layers.8.mlp.up_proj", |
| "language_model.layers.4.mlp.up_proj", |
| "19.mlp.gate_proj", |
| "22.self_attn.q_proj", |
| "language_model.layers.6.self_attn.q_proj", |
| "language_model.layers.1.self_attn.v_proj", |
| "language_model.layers.5.mlp.gate_proj", |
| "19.mlp.down_proj", |
| "27.self_attn.o_proj", |
| "language_model.layers.8.self_attn.q_proj", |
| "26.self_attn.o_proj", |
| "34.mlp.up_proj", |
| "20.mlp.down_proj", |
| "language_model.layers.4.self_attn.q_proj", |
| "language_model.layers.6.self_attn.o_proj", |
| "22.mlp.up_proj", |
| "language_model.layers.14.self_attn.v_proj", |
| "32.self_attn.o_proj", |
| "27.mlp.up_proj", |
| "language_model.layers.2.mlp.down_proj", |
| "language_model.layers.4.mlp.gate_proj", |
| "language_model.layers.2.self_attn.v_proj", |
| "18.self_attn.q_proj", |
| "language_model.layers.10.mlp.gate_proj", |
| "language_model.layers.5.self_attn.v_proj", |
| "18.mlp.down_proj", |
| "20.mlp.gate_proj", |
| "language_model.layers.13.self_attn.q_proj", |
| "language_model.layers.11.self_attn.k_proj", |
| "30.self_attn.o_proj", |
| "19.self_attn.q_proj", |
| "32.mlp.gate_proj", |
| "language_model.layers.0.mlp.gate_proj", |
| "language_model.layers.1.self_attn.o_proj", |
| "language_model.layers.10.self_attn.q_proj", |
| "language_model.layers.8.self_attn.o_proj", |
| "language_model.layers.10.self_attn.v_proj", |
| "language_model.layers.6.mlp.down_proj", |
| "24.self_attn.o_proj", |
| "33.self_attn.q_proj", |
| "26.mlp.up_proj", |
| "language_model.layers.12.mlp.down_proj", |
| "language_model.layers.9.self_attn.k_proj", |
| "30.mlp.up_proj", |
| "language_model.layers.11.mlp.up_proj", |
| "23.self_attn.q_proj", |
| "17.self_attn.o_proj", |
| "language_model.layers.0.self_attn.q_proj", |
| "31.self_attn.q_proj", |
| "25.mlp.gate_proj", |
| "23.mlp.up_proj", |
| "31.mlp.down_proj", |
| "language_model.layers.9.self_attn.v_proj", |
| "language_model.layers.1.mlp.up_proj", |
| "language_model.layers.12.self_attn.k_proj", |
| "22.mlp.down_proj", |
| "24.mlp.down_proj", |
| "19.mlp.up_proj", |
| "language_model.layers.14.self_attn.k_proj", |
| "29.mlp.gate_proj", |
| "29.self_attn.o_proj", |
| "18.mlp.gate_proj", |
| "language_model.layers.3.self_attn.o_proj", |
| "language_model.layers.4.self_attn.o_proj", |
| "language_model.layers.6.self_attn.k_proj", |
| "language_model.layers.15.mlp.gate_proj", |
| "32.mlp.up_proj", |
| "32.mlp.down_proj", |
| "language_model.layers.2.mlp.gate_proj", |
| "29.mlp.up_proj", |
| "language_model.layers.7.self_attn.k_proj", |
| "language_model.layers.12.mlp.up_proj", |
| "language_model.layers.11.self_attn.q_proj", |
| "language_model.layers.11.self_attn.v_proj", |
| "language_model.layers.6.mlp.gate_proj", |
| "language_model.layers.3.mlp.up_proj", |
| "language_model.layers.5.mlp.down_proj", |
| "language_model.layers.13.mlp.gate_proj", |
| "16.mlp.up_proj", |
| "33.mlp.up_proj", |
| "language_model.layers.7.mlp.gate_proj", |
| "language_model.layers.5.self_attn.q_proj", |
| "21.mlp.down_proj", |
| "25.mlp.down_proj", |
| "language_model.layers.6.mlp.up_proj", |
| "language_model.layers.8.self_attn.v_proj", |
| "input_proj_linear", |
| "language_model.layers.11.self_attn.o_proj", |
| "language_model.layers.2.mlp.up_proj", |
| "language_model.layers.11.mlp.down_proj" |
| ], |
| "target_parameters": null, |
| "task_type": "CAUSAL_LM", |
| "trainable_token_indices": null, |
| "use_dora": false, |
| "use_qalora": false, |
| "use_rslora": false |
| } |
