Upload 4 files

README.md

@@ -0,0 +1,7 @@
+---
+base_model: meta-llama/Llama-3.2-11B-Vision-Instruct
+library_name: peft
+---
+### Framework versions
+
+- PEFT 0.13.2

adapter_config.json

@@ -0,0 +1,39 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "meta-llama/Llama-3.2-11B-Vision-Instruct",
+  "bias": "none",
+  "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": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 64,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "k_proj",
+    "v_proj",
+    "o_proj",
+    "gate_proj",
+    "up_proj",
+    "down_proj",
+    "fc1",
+    "fc2"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39ffd520f9657c71854c3078dcbadded4e10862daf6421511f73da4ee2251cc7
+size 413177536

nas_lora_config.json

@@ -0,0 +1 @@
+{"base_model.model.vision_model.transformer.layers.0.self_attn.q_proj": 8, "base_model.model.vision_model.transformer.layers.0.self_attn.k_proj": 4, "base_model.model.vision_model.transformer.layers.0.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.0.self_attn.o_proj": 16, "base_model.model.vision_model.transformer.layers.0.mlp.fc1": 64, "base_model.model.vision_model.transformer.layers.0.mlp.fc2": 32, "base_model.model.vision_model.transformer.layers.1.self_attn.q_proj": 16, "base_model.model.vision_model.transformer.layers.1.self_attn.k_proj": 32, "base_model.model.vision_model.transformer.layers.1.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.1.self_attn.o_proj": 4, "base_model.model.vision_model.transformer.layers.1.mlp.fc1": 32, "base_model.model.vision_model.transformer.layers.1.mlp.fc2": 4, "base_model.model.vision_model.transformer.layers.2.self_attn.q_proj": 4, "base_model.model.vision_model.transformer.layers.2.self_attn.k_proj": 4, "base_model.model.vision_model.transformer.layers.2.self_attn.v_proj": 8, "base_model.model.vision_model.transformer.layers.2.self_attn.o_proj": 8, "base_model.model.vision_model.transformer.layers.2.mlp.fc1": 16, "base_model.model.vision_model.transformer.layers.2.mlp.fc2": 4, "base_model.model.vision_model.transformer.layers.3.self_attn.q_proj": 4, "base_model.model.vision_model.transformer.layers.3.self_attn.k_proj": 64, "base_model.model.vision_model.transformer.layers.3.self_attn.v_proj": 4, "base_model.model.vision_model.transformer.layers.3.self_attn.o_proj": 32, "base_model.model.vision_model.transformer.layers.3.mlp.fc1": 4, "base_model.model.vision_model.transformer.layers.3.mlp.fc2": 16, "base_model.model.vision_model.transformer.layers.4.self_attn.q_proj": 16, "base_model.model.vision_model.transformer.layers.4.self_attn.k_proj": 32, "base_model.model.vision_model.transformer.layers.4.self_attn.v_proj": 8, "base_model.model.vision_model.transformer.layers.4.self_attn.o_proj": 32, "base_model.model.vision_model.transformer.layers.4.mlp.fc1": 8, "base_model.model.vision_model.transformer.layers.4.mlp.fc2": 32, "base_model.model.vision_model.transformer.layers.5.self_attn.q_proj": 64, "base_model.model.vision_model.transformer.layers.5.self_attn.k_proj": 4, "base_model.model.vision_model.transformer.layers.5.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.5.self_attn.o_proj": 8, "base_model.model.vision_model.transformer.layers.5.mlp.fc1": 64, "base_model.model.vision_model.transformer.layers.5.mlp.fc2": 8, "base_model.model.vision_model.transformer.layers.6.self_attn.q_proj": 4, "base_model.model.vision_model.transformer.layers.6.self_attn.k_proj": 16, "base_model.model.vision_model.transformer.layers.6.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.6.self_attn.o_proj": 32, "base_model.model.vision_model.transformer.layers.6.mlp.fc1": 64, "base_model.model.vision_model.transformer.layers.6.mlp.fc2": 8, "base_model.model.vision_model.transformer.layers.7.self_attn.q_proj": 4, "base_model.model.vision_model.transformer.layers.7.self_attn.k_proj": 32, "base_model.model.vision_model.transformer.layers.7.self_attn.v_proj": 16, "base_model.model.vision_model.transformer.layers.7.self_attn.o_proj": 16, "base_model.model.vision_model.transformer.layers.7.mlp.fc1": 64, "base_model.model.vision_model.transformer.layers.7.mlp.fc2": 4, "base_model.model.vision_model.transformer.layers.8.self_attn.q_proj": 64, "base_model.model.vision_model.transformer.layers.8.self_attn.k_proj": 32, "base_model.model.vision_model.transformer.layers.8.self_attn.v_proj": 16, "base_model.model.vision_model.transformer.layers.8.self_attn.o_proj": 16, "base_model.model.vision_model.transformer.layers.8.mlp.fc1": 32, "base_model.model.vision_model.transformer.layers.8.mlp.fc2": 8, "base_model.model.vision_model.transformer.layers.9.self_attn.q_proj": 16, "base_model.model.vision_model.transformer.layers.9.self_attn.k_proj": 16, "base_model.model.vision_model.transformer.layers.9.self_attn.v_proj": 16, "base_model.model.vision_model.transformer.layers.9.self_attn.o_proj": 8, "base_model.model.vision_model.transformer.layers.9.mlp.fc1": 32, "base_model.model.vision_model.transformer.layers.9.mlp.fc2": 4, "base_model.model.vision_model.transformer.layers.10.self_attn.q_proj": 8, "base_model.model.vision_model.transformer.layers.10.self_attn.k_proj": 8, "base_model.model.vision_model.transformer.layers.10.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.10.self_attn.o_proj": 4, "base_model.model.vision_model.transformer.layers.10.mlp.fc1": 32, "base_model.model.vision_model.transformer.layers.10.mlp.fc2": 32, "base_model.model.vision_model.transformer.layers.11.self_attn.q_proj": 16, "base_model.model.vision_model.transformer.layers.11.self_attn.k_proj": 32, "base_model.model.vision_model.transformer.layers.11.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.11.self_attn.o_proj": 4, "base_model.model.vision_model.transformer.layers.11.mlp.fc1": 8, "base_model.model.vision_model.transformer.layers.11.mlp.fc2": 64, "base_model.model.vision_model.transformer.layers.12.self_attn.q_proj": 16, "base_model.model.vision_model.transformer.layers.12.self_attn.k_proj": 32, "base_model.model.vision_model.transformer.layers.12.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.12.self_attn.o_proj": 4, "base_model.model.vision_model.transformer.layers.12.mlp.fc1": 4, "base_model.model.vision_model.transformer.layers.12.mlp.fc2": 8, "base_model.model.vision_model.transformer.layers.13.self_attn.q_proj": 64, "base_model.model.vision_model.transformer.layers.13.self_attn.k_proj": 4, "base_model.model.vision_model.transformer.layers.13.self_attn.v_proj": 4, "base_model.model.vision_model.transformer.layers.13.self_attn.o_proj": 64, "base_model.model.vision_model.transformer.layers.13.mlp.fc1": 32, "base_model.model.vision_model.transformer.layers.13.mlp.fc2": 64, "base_model.model.vision_model.transformer.layers.14.self_attn.q_proj": 16, "base_model.model.vision_model.transformer.layers.14.self_attn.k_proj": 8, "base_model.model.vision_model.transformer.layers.14.self_attn.v_proj": 4, "base_model.model.vision_model.transformer.layers.14.self_attn.o_proj": 4, "base_model.model.vision_model.transformer.layers.14.mlp.fc1": 4, "base_model.model.vision_model.transformer.layers.14.mlp.fc2": 16, "base_model.model.vision_model.transformer.layers.15.self_attn.q_proj": 4, "base_model.model.vision_model.transformer.layers.15.self_attn.k_proj": 16, "base_model.model.vision_model.transformer.layers.15.self_attn.v_proj": 64, "base_model.model.vision_model.transformer.layers.15.self_attn.o_proj": 32, "base_model.model.vision_model.transformer.layers.15.mlp.fc1": 4, "base_model.model.vision_model.transformer.layers.15.mlp.fc2": 4, "base_model.model.vision_model.transformer.layers.16.self_attn.q_proj": 8, "base_model.model.vision_model.transformer.layers.16.self_attn.k_proj": 4, "base_model.model.vision_model.transformer.layers.16.self_attn.v_proj": 4, "base_model.model.vision_model.transformer.layers.16.self_attn.o_proj": 32, "base_model.model.vision_model.transformer.layers.16.mlp.fc1": 8, "base_model.model.vision_model.transformer.layers.16.mlp.fc2": 64, "base_model.model.vision_model.transformer.layers.17.self_attn.q_proj": 16, "base_model.model.vision_model.transformer.layers.17.self_attn.k_proj": 64, "base_model.model.vision_model.transformer.layers.17.self_attn.v_proj": 16, "base_model.model.vision_model.transformer.layers.17.self_attn.o_proj": 16, "base_model.model.vision_model.transformer.layers.17.mlp.fc1": 32, "base_model.model.vision_model.transformer.layers.17.mlp.fc2": 4, "base_model.model.vision_model.transformer.layers.18.self_attn.q_proj": 8, "base_model.model.vision_model.transformer.layers.18.self_attn.k_proj": 32, "base_model.model.vision_model.transformer.layers.18.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.18.self_attn.o_proj": 16, "base_model.model.vision_model.transformer.layers.18.mlp.fc1": 4, "base_model.model.vision_model.transformer.layers.18.mlp.fc2": 4, "base_model.model.vision_model.transformer.layers.19.self_attn.q_proj": 64, "base_model.model.vision_model.transformer.layers.19.self_attn.k_proj": 8, "base_model.model.vision_model.transformer.layers.19.self_attn.v_proj": 64, "base_model.model.vision_model.transformer.layers.19.self_attn.o_proj": 32, "base_model.model.vision_model.transformer.layers.19.mlp.fc1": 4, "base_model.model.vision_model.transformer.layers.19.mlp.fc2": 8, "base_model.model.vision_model.transformer.layers.20.self_attn.q_proj": 4, "base_model.model.vision_model.transformer.layers.20.self_attn.k_proj": 16, "base_model.model.vision_model.transformer.layers.20.self_attn.v_proj": 4, "base_model.model.vision_model.transformer.layers.20.self_attn.o_proj": 4, "base_model.model.vision_model.transformer.layers.20.mlp.fc1": 8, "base_model.model.vision_model.transformer.layers.20.mlp.fc2": 32, "base_model.model.vision_model.transformer.layers.21.self_attn.q_proj": 8, "base_model.model.vision_model.transformer.layers.21.self_attn.k_proj": 32, "base_model.model.vision_model.transformer.layers.21.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.21.self_attn.o_proj": 32, "base_model.model.vision_model.transformer.layers.21.mlp.fc1": 64, "base_model.model.vision_model.transformer.layers.21.mlp.fc2": 32, "base_model.model.vision_model.transformer.layers.22.self_attn.q_proj": 32, "base_model.model.vision_model.transformer.layers.22.self_attn.k_proj": 32, "base_model.model.vision_model.transformer.layers.22.self_attn.v_proj": 64, "base_model.model.vision_model.transformer.layers.22.self_attn.o_proj": 64, "base_model.model.vision_model.transformer.layers.22.mlp.fc1": 64, "base_model.model.vision_model.transformer.layers.22.mlp.fc2": 16, "base_model.model.vision_model.transformer.layers.23.self_attn.q_proj": 4, "base_model.model.vision_model.transformer.layers.23.self_attn.k_proj": 4, "base_model.model.vision_model.transformer.layers.23.self_attn.v_proj": 8, "base_model.model.vision_model.transformer.layers.23.self_attn.o_proj": 16, "base_model.model.vision_model.transformer.layers.23.mlp.fc1": 16, "base_model.model.vision_model.transformer.layers.23.mlp.fc2": 16, "base_model.model.vision_model.transformer.layers.24.self_attn.q_proj": 16, "base_model.model.vision_model.transformer.layers.24.self_attn.k_proj": 8, "base_model.model.vision_model.transformer.layers.24.self_attn.v_proj": 64, "base_model.model.vision_model.transformer.layers.24.self_attn.o_proj": 4, "base_model.model.vision_model.transformer.layers.24.mlp.fc1": 16, "base_model.model.vision_model.transformer.layers.24.mlp.fc2": 16, "base_model.model.vision_model.transformer.layers.25.self_attn.q_proj": 4, "base_model.model.vision_model.transformer.layers.25.self_attn.k_proj": 4, "base_model.model.vision_model.transformer.layers.25.self_attn.v_proj": 8, "base_model.model.vision_model.transformer.layers.25.self_attn.o_proj": 32, "base_model.model.vision_model.transformer.layers.25.mlp.fc1": 16, "base_model.model.vision_model.transformer.layers.25.mlp.fc2": 32, "base_model.model.vision_model.transformer.layers.26.self_attn.q_proj": 8, "base_model.model.vision_model.transformer.layers.26.self_attn.k_proj": 64, "base_model.model.vision_model.transformer.layers.26.self_attn.v_proj": 4, "base_model.model.vision_model.transformer.layers.26.self_attn.o_proj": 8, "base_model.model.vision_model.transformer.layers.26.mlp.fc1": 64, "base_model.model.vision_model.transformer.layers.26.mlp.fc2": 4, "base_model.model.vision_model.transformer.layers.27.self_attn.q_proj": 8, "base_model.model.vision_model.transformer.layers.27.self_attn.k_proj": 64, "base_model.model.vision_model.transformer.layers.27.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.27.self_attn.o_proj": 64, "base_model.model.vision_model.transformer.layers.27.mlp.fc1": 4, "base_model.model.vision_model.transformer.layers.27.mlp.fc2": 32, "base_model.model.vision_model.transformer.layers.28.self_attn.q_proj": 4, "base_model.model.vision_model.transformer.layers.28.self_attn.k_proj": 64, "base_model.model.vision_model.transformer.layers.28.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.28.self_attn.o_proj": 16, "base_model.model.vision_model.transformer.layers.28.mlp.fc1": 16, "base_model.model.vision_model.transformer.layers.28.mlp.fc2": 16, "base_model.model.vision_model.transformer.layers.29.self_attn.q_proj": 16, "base_model.model.vision_model.transformer.layers.29.self_attn.k_proj": 16, "base_model.model.vision_model.transformer.layers.29.self_attn.v_proj": 16, "base_model.model.vision_model.transformer.layers.29.self_attn.o_proj": 4, "base_model.model.vision_model.transformer.layers.29.mlp.fc1": 8, "base_model.model.vision_model.transformer.layers.29.mlp.fc2": 32, "base_model.model.vision_model.transformer.layers.30.self_attn.q_proj": 8, "base_model.model.vision_model.transformer.layers.30.self_attn.k_proj": 8, "base_model.model.vision_model.transformer.layers.30.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.30.self_attn.o_proj": 4, "base_model.model.vision_model.transformer.layers.30.mlp.fc1": 16, "base_model.model.vision_model.transformer.layers.30.mlp.fc2": 4, "base_model.model.vision_model.transformer.layers.31.self_attn.q_proj": 64, "base_model.model.vision_model.transformer.layers.31.self_attn.k_proj": 4, "base_model.model.vision_model.transformer.layers.31.self_attn.v_proj": 32, "base_model.model.vision_model.transformer.layers.31.self_attn.o_proj": 64, "base_model.model.vision_model.transformer.layers.31.mlp.fc1": 64, "base_model.model.vision_model.transformer.layers.31.mlp.fc2": 32, "base_model.model.vision_model.global_transformer.layers.0.self_attn.q_proj": 64, "base_model.model.vision_model.global_transformer.layers.0.self_attn.k_proj": 4, "base_model.model.vision_model.global_transformer.layers.0.self_attn.v_proj": 32, "base_model.model.vision_model.global_transformer.layers.0.self_attn.o_proj": 32, "base_model.model.vision_model.global_transformer.layers.0.mlp.fc1": 16, "base_model.model.vision_model.global_transformer.layers.0.mlp.fc2": 16, "base_model.model.vision_model.global_transformer.layers.1.self_attn.q_proj": 4, "base_model.model.vision_model.global_transformer.layers.1.self_attn.k_proj": 16, "base_model.model.vision_model.global_transformer.layers.1.self_attn.v_proj": 8, "base_model.model.vision_model.global_transformer.layers.1.self_attn.o_proj": 8, "base_model.model.vision_model.global_transformer.layers.1.mlp.fc1": 4, "base_model.model.vision_model.global_transformer.layers.1.mlp.fc2": 32, "base_model.model.vision_model.global_transformer.layers.2.self_attn.q_proj": 64, "base_model.model.vision_model.global_transformer.layers.2.self_attn.k_proj": 8, "base_model.model.vision_model.global_transformer.layers.2.self_attn.v_proj": 32, "base_model.model.vision_model.global_transformer.layers.2.self_attn.o_proj": 8, "base_model.model.vision_model.global_transformer.layers.2.mlp.fc1": 16, "base_model.model.vision_model.global_transformer.layers.2.mlp.fc2": 4, "base_model.model.vision_model.global_transformer.layers.3.self_attn.q_proj": 32, "base_model.model.vision_model.global_transformer.layers.3.self_attn.k_proj": 16, "base_model.model.vision_model.global_transformer.layers.3.self_attn.v_proj": 32, "base_model.model.vision_model.global_transformer.layers.3.self_attn.o_proj": 4, "base_model.model.vision_model.global_transformer.layers.3.mlp.fc1": 64, "base_model.model.vision_model.global_transformer.layers.3.mlp.fc2": 16, "base_model.model.vision_model.global_transformer.layers.4.self_attn.q_proj": 64, "base_model.model.vision_model.global_transformer.layers.4.self_attn.k_proj": 64, "base_model.model.vision_model.global_transformer.layers.4.self_attn.v_proj": 8, "base_model.model.vision_model.global_transformer.layers.4.self_attn.o_proj": 4, "base_model.model.vision_model.global_transformer.layers.4.mlp.fc1": 64, "base_model.model.vision_model.global_transformer.layers.4.mlp.fc2": 16, "base_model.model.vision_model.global_transformer.layers.5.self_attn.q_proj": 16, "base_model.model.vision_model.global_transformer.layers.5.self_attn.k_proj": 32, "base_model.model.vision_model.global_transformer.layers.5.self_attn.v_proj": 16, "base_model.model.vision_model.global_transformer.layers.5.self_attn.o_proj": 4, "base_model.model.vision_model.global_transformer.layers.5.mlp.fc1": 16, "base_model.model.vision_model.global_transformer.layers.5.mlp.fc2": 8, "base_model.model.vision_model.global_transformer.layers.6.self_attn.q_proj": 8, "base_model.model.vision_model.global_transformer.layers.6.self_attn.k_proj": 8, "base_model.model.vision_model.global_transformer.layers.6.self_attn.v_proj": 64, "base_model.model.vision_model.global_transformer.layers.6.self_attn.o_proj": 64, "base_model.model.vision_model.global_transformer.layers.6.mlp.fc1": 32, "base_model.model.vision_model.global_transformer.layers.6.mlp.fc2": 4, "base_model.model.vision_model.global_transformer.layers.7.self_attn.q_proj": 16, "base_model.model.vision_model.global_transformer.layers.7.self_attn.k_proj": 64, "base_model.model.vision_model.global_transformer.layers.7.self_attn.v_proj": 32, "base_model.model.vision_model.global_transformer.layers.7.self_attn.o_proj": 64, "base_model.model.vision_model.global_transformer.layers.7.mlp.fc1": 64, "base_model.model.vision_model.global_transformer.layers.7.mlp.fc2": 64, "base_model.model.language_model.model.layers.0.self_attn.q_proj": 64, "base_model.model.language_model.model.layers.0.self_attn.k_proj": 8, "base_model.model.language_model.model.layers.0.self_attn.v_proj": 16, "base_model.model.language_model.model.layers.0.self_attn.o_proj": 16, "base_model.model.language_model.model.layers.0.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.0.mlp.up_proj": 16, "base_model.model.language_model.model.layers.0.mlp.down_proj": 32, "base_model.model.language_model.model.layers.1.self_attn.q_proj": 4, "base_model.model.language_model.model.layers.1.self_attn.k_proj": 32, "base_model.model.language_model.model.layers.1.self_attn.v_proj": 8, "base_model.model.language_model.model.layers.1.self_attn.o_proj": 64, "base_model.model.language_model.model.layers.1.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.1.mlp.up_proj": 32, "base_model.model.language_model.model.layers.1.mlp.down_proj": 16, "base_model.model.language_model.model.layers.2.self_attn.q_proj": 4, "base_model.model.language_model.model.layers.2.self_attn.k_proj": 16, "base_model.model.language_model.model.layers.2.self_attn.v_proj": 4, "base_model.model.language_model.model.layers.2.self_attn.o_proj": 8, "base_model.model.language_model.model.layers.2.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.2.mlp.up_proj": 4, "base_model.model.language_model.model.layers.2.mlp.down_proj": 4, "base_model.model.language_model.model.layers.3.cross_attn.q_proj": 16, "base_model.model.language_model.model.layers.3.cross_attn.k_proj": 16, "base_model.model.language_model.model.layers.3.cross_attn.v_proj": 32, "base_model.model.language_model.model.layers.3.cross_attn.o_proj": 32, "base_model.model.language_model.model.layers.3.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.3.mlp.up_proj": 32, "base_model.model.language_model.model.layers.3.mlp.down_proj": 4, "base_model.model.language_model.model.layers.4.self_attn.q_proj": 64, "base_model.model.language_model.model.layers.4.self_attn.k_proj": 4, "base_model.model.language_model.model.layers.4.self_attn.v_proj": 64, "base_model.model.language_model.model.layers.4.self_attn.o_proj": 32, "base_model.model.language_model.model.layers.4.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.4.mlp.up_proj": 16, "base_model.model.language_model.model.layers.4.mlp.down_proj": 8, "base_model.model.language_model.model.layers.5.self_attn.q_proj": 8, "base_model.model.language_model.model.layers.5.self_attn.k_proj": 64, "base_model.model.language_model.model.layers.5.self_attn.v_proj": 16, "base_model.model.language_model.model.layers.5.self_attn.o_proj": 4, "base_model.model.language_model.model.layers.5.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.5.mlp.up_proj": 32, "base_model.model.language_model.model.layers.5.mlp.down_proj": 4, "base_model.model.language_model.model.layers.6.self_attn.q_proj": 16, "base_model.model.language_model.model.layers.6.self_attn.k_proj": 64, "base_model.model.language_model.model.layers.6.self_attn.v_proj": 16, "base_model.model.language_model.model.layers.6.self_attn.o_proj": 16, "base_model.model.language_model.model.layers.6.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.6.mlp.up_proj": 4, "base_model.model.language_model.model.layers.6.mlp.down_proj": 4, "base_model.model.language_model.model.layers.7.self_attn.q_proj": 4, "base_model.model.language_model.model.layers.7.self_attn.k_proj": 16, "base_model.model.language_model.model.layers.7.self_attn.v_proj": 4, "base_model.model.language_model.model.layers.7.self_attn.o_proj": 64, "base_model.model.language_model.model.layers.7.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.7.mlp.up_proj": 4, "base_model.model.language_model.model.layers.7.mlp.down_proj": 64, "base_model.model.language_model.model.layers.8.cross_attn.q_proj": 8, "base_model.model.language_model.model.layers.8.cross_attn.k_proj": 16, "base_model.model.language_model.model.layers.8.cross_attn.v_proj": 16, "base_model.model.language_model.model.layers.8.cross_attn.o_proj": 64, "base_model.model.language_model.model.layers.8.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.8.mlp.up_proj": 64, "base_model.model.language_model.model.layers.8.mlp.down_proj": 16, "base_model.model.language_model.model.layers.9.self_attn.q_proj": 32, "base_model.model.language_model.model.layers.9.self_attn.k_proj": 8, "base_model.model.language_model.model.layers.9.self_attn.v_proj": 8, "base_model.model.language_model.model.layers.9.self_attn.o_proj": 8, "base_model.model.language_model.model.layers.9.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.9.mlp.up_proj": 64, "base_model.model.language_model.model.layers.9.mlp.down_proj": 4, "base_model.model.language_model.model.layers.10.self_attn.q_proj": 16, "base_model.model.language_model.model.layers.10.self_attn.k_proj": 4, "base_model.model.language_model.model.layers.10.self_attn.v_proj": 8, "base_model.model.language_model.model.layers.10.self_attn.o_proj": 16, "base_model.model.language_model.model.layers.10.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.10.mlp.up_proj": 8, "base_model.model.language_model.model.layers.10.mlp.down_proj": 32, "base_model.model.language_model.model.layers.11.self_attn.q_proj": 64, "base_model.model.language_model.model.layers.11.self_attn.k_proj": 64, "base_model.model.language_model.model.layers.11.self_attn.v_proj": 4, "base_model.model.language_model.model.layers.11.self_attn.o_proj": 32, "base_model.model.language_model.model.layers.11.mlp.gate_proj": 16, "base_model.model.language_model.model.layers.11.mlp.up_proj": 32, "base_model.model.language_model.model.layers.11.mlp.down_proj": 32, "base_model.model.language_model.model.layers.12.self_attn.q_proj": 16, "base_model.model.language_model.model.layers.12.self_attn.k_proj": 4, "base_model.model.language_model.model.layers.12.self_attn.v_proj": 8, "base_model.model.language_model.model.layers.12.self_attn.o_proj": 16, "base_model.model.language_model.model.layers.12.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.12.mlp.up_proj": 16, "base_model.model.language_model.model.layers.12.mlp.down_proj": 16, "base_model.model.language_model.model.layers.13.cross_attn.q_proj": 16, "base_model.model.language_model.model.layers.13.cross_attn.k_proj": 8, "base_model.model.language_model.model.layers.13.cross_attn.v_proj": 8, "base_model.model.language_model.model.layers.13.cross_attn.o_proj": 8, "base_model.model.language_model.model.layers.13.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.13.mlp.up_proj": 32, "base_model.model.language_model.model.layers.13.mlp.down_proj": 8, "base_model.model.language_model.model.layers.14.self_attn.q_proj": 16, "base_model.model.language_model.model.layers.14.self_attn.k_proj": 64, "base_model.model.language_model.model.layers.14.self_attn.v_proj": 8, "base_model.model.language_model.model.layers.14.self_attn.o_proj": 8, "base_model.model.language_model.model.layers.14.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.14.mlp.up_proj": 32, "base_model.model.language_model.model.layers.14.mlp.down_proj": 64, "base_model.model.language_model.model.layers.15.self_attn.q_proj": 16, "base_model.model.language_model.model.layers.15.self_attn.k_proj": 64, "base_model.model.language_model.model.layers.15.self_attn.v_proj": 64, "base_model.model.language_model.model.layers.15.self_attn.o_proj": 16, "base_model.model.language_model.model.layers.15.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.15.mlp.up_proj": 32, "base_model.model.language_model.model.layers.15.mlp.down_proj": 32, "base_model.model.language_model.model.layers.16.self_attn.q_proj": 64, "base_model.model.language_model.model.layers.16.self_attn.k_proj": 4, "base_model.model.language_model.model.layers.16.self_attn.v_proj": 8, "base_model.model.language_model.model.layers.16.self_attn.o_proj": 8, "base_model.model.language_model.model.layers.16.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.16.mlp.up_proj": 32, "base_model.model.language_model.model.layers.16.mlp.down_proj": 4, "base_model.model.language_model.model.layers.17.self_attn.q_proj": 64, "base_model.model.language_model.model.layers.17.self_attn.k_proj": 16, "base_model.model.language_model.model.layers.17.self_attn.v_proj": 4, "base_model.model.language_model.model.layers.17.self_attn.o_proj": 32, "base_model.model.language_model.model.layers.17.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.17.mlp.up_proj": 32, "base_model.model.language_model.model.layers.17.mlp.down_proj": 16, "base_model.model.language_model.model.layers.18.cross_attn.q_proj": 16, "base_model.model.language_model.model.layers.18.cross_attn.k_proj": 8, "base_model.model.language_model.model.layers.18.cross_attn.v_proj": 64, "base_model.model.language_model.model.layers.18.cross_attn.o_proj": 32, "base_model.model.language_model.model.layers.18.mlp.gate_proj": 16, "base_model.model.language_model.model.layers.18.mlp.up_proj": 16, "base_model.model.language_model.model.layers.18.mlp.down_proj": 4, "base_model.model.language_model.model.layers.19.self_attn.q_proj": 16, "base_model.model.language_model.model.layers.19.self_attn.k_proj": 4, "base_model.model.language_model.model.layers.19.self_attn.v_proj": 16, "base_model.model.language_model.model.layers.19.self_attn.o_proj": 4, "base_model.model.language_model.model.layers.19.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.19.mlp.up_proj": 32, "base_model.model.language_model.model.layers.19.mlp.down_proj": 8, "base_model.model.language_model.model.layers.20.self_attn.q_proj": 16, "base_model.model.language_model.model.layers.20.self_attn.k_proj": 16, "base_model.model.language_model.model.layers.20.self_attn.v_proj": 64, "base_model.model.language_model.model.layers.20.self_attn.o_proj": 8, "base_model.model.language_model.model.layers.20.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.20.mlp.up_proj": 64, "base_model.model.language_model.model.layers.20.mlp.down_proj": 32, "base_model.model.language_model.model.layers.21.self_attn.q_proj": 32, "base_model.model.language_model.model.layers.21.self_attn.k_proj": 16, "base_model.model.language_model.model.layers.21.self_attn.v_proj": 8, "base_model.model.language_model.model.layers.21.self_attn.o_proj": 64, "base_model.model.language_model.model.layers.21.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.21.mlp.up_proj": 16, "base_model.model.language_model.model.layers.21.mlp.down_proj": 32, "base_model.model.language_model.model.layers.22.self_attn.q_proj": 32, "base_model.model.language_model.model.layers.22.self_attn.k_proj": 4, "base_model.model.language_model.model.layers.22.self_attn.v_proj": 8, "base_model.model.language_model.model.layers.22.self_attn.o_proj": 8, "base_model.model.language_model.model.layers.22.mlp.gate_proj": 16, "base_model.model.language_model.model.layers.22.mlp.up_proj": 16, "base_model.model.language_model.model.layers.22.mlp.down_proj": 64, "base_model.model.language_model.model.layers.23.cross_attn.q_proj": 16, "base_model.model.language_model.model.layers.23.cross_attn.k_proj": 4, "base_model.model.language_model.model.layers.23.cross_attn.v_proj": 4, "base_model.model.language_model.model.layers.23.cross_attn.o_proj": 64, "base_model.model.language_model.model.layers.23.mlp.gate_proj": 16, "base_model.model.language_model.model.layers.23.mlp.up_proj": 64, "base_model.model.language_model.model.layers.23.mlp.down_proj": 64, "base_model.model.language_model.model.layers.24.self_attn.q_proj": 8, "base_model.model.language_model.model.layers.24.self_attn.k_proj": 64, "base_model.model.language_model.model.layers.24.self_attn.v_proj": 64, "base_model.model.language_model.model.layers.24.self_attn.o_proj": 64, "base_model.model.language_model.model.layers.24.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.24.mlp.up_proj": 8, "base_model.model.language_model.model.layers.24.mlp.down_proj": 64, "base_model.model.language_model.model.layers.25.self_attn.q_proj": 16, "base_model.model.language_model.model.layers.25.self_attn.k_proj": 64, "base_model.model.language_model.model.layers.25.self_attn.v_proj": 4, "base_model.model.language_model.model.layers.25.self_attn.o_proj": 8, "base_model.model.language_model.model.layers.25.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.25.mlp.up_proj": 16, "base_model.model.language_model.model.layers.25.mlp.down_proj": 32, "base_model.model.language_model.model.layers.26.self_attn.q_proj": 32, "base_model.model.language_model.model.layers.26.self_attn.k_proj": 32, "base_model.model.language_model.model.layers.26.self_attn.v_proj": 4, "base_model.model.language_model.model.layers.26.self_attn.o_proj": 8, "base_model.model.language_model.model.layers.26.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.26.mlp.up_proj": 64, "base_model.model.language_model.model.layers.26.mlp.down_proj": 64, "base_model.model.language_model.model.layers.27.self_attn.q_proj": 64, "base_model.model.language_model.model.layers.27.self_attn.k_proj": 32, "base_model.model.language_model.model.layers.27.self_attn.v_proj": 8, "base_model.model.language_model.model.layers.27.self_attn.o_proj": 4, "base_model.model.language_model.model.layers.27.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.27.mlp.up_proj": 64, "base_model.model.language_model.model.layers.27.mlp.down_proj": 8, "base_model.model.language_model.model.layers.28.cross_attn.q_proj": 4, "base_model.model.language_model.model.layers.28.cross_attn.k_proj": 32, "base_model.model.language_model.model.layers.28.cross_attn.v_proj": 4, "base_model.model.language_model.model.layers.28.cross_attn.o_proj": 16, "base_model.model.language_model.model.layers.28.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.28.mlp.up_proj": 32, "base_model.model.language_model.model.layers.28.mlp.down_proj": 32, "base_model.model.language_model.model.layers.29.self_attn.q_proj": 8, "base_model.model.language_model.model.layers.29.self_attn.k_proj": 64, "base_model.model.language_model.model.layers.29.self_attn.v_proj": 8, "base_model.model.language_model.model.layers.29.self_attn.o_proj": 4, "base_model.model.language_model.model.layers.29.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.29.mlp.up_proj": 32, "base_model.model.language_model.model.layers.29.mlp.down_proj": 32, "base_model.model.language_model.model.layers.30.self_attn.q_proj": 4, "base_model.model.language_model.model.layers.30.self_attn.k_proj": 8, "base_model.model.language_model.model.layers.30.self_attn.v_proj": 16, "base_model.model.language_model.model.layers.30.self_attn.o_proj": 16, "base_model.model.language_model.model.layers.30.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.30.mlp.up_proj": 32, "base_model.model.language_model.model.layers.30.mlp.down_proj": 4, "base_model.model.language_model.model.layers.31.self_attn.q_proj": 8, "base_model.model.language_model.model.layers.31.self_attn.k_proj": 4, "base_model.model.language_model.model.layers.31.self_attn.v_proj": 32, "base_model.model.language_model.model.layers.31.self_attn.o_proj": 16, "base_model.model.language_model.model.layers.31.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.31.mlp.up_proj": 16, "base_model.model.language_model.model.layers.31.mlp.down_proj": 8, "base_model.model.language_model.model.layers.32.self_attn.q_proj": 8, "base_model.model.language_model.model.layers.32.self_attn.k_proj": 64, "base_model.model.language_model.model.layers.32.self_attn.v_proj": 32, "base_model.model.language_model.model.layers.32.self_attn.o_proj": 16, "base_model.model.language_model.model.layers.32.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.32.mlp.up_proj": 16, "base_model.model.language_model.model.layers.32.mlp.down_proj": 64, "base_model.model.language_model.model.layers.33.cross_attn.q_proj": 16, "base_model.model.language_model.model.layers.33.cross_attn.k_proj": 4, "base_model.model.language_model.model.layers.33.cross_attn.v_proj": 32, "base_model.model.language_model.model.layers.33.cross_attn.o_proj": 32, "base_model.model.language_model.model.layers.33.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.33.mlp.up_proj": 16, "base_model.model.language_model.model.layers.33.mlp.down_proj": 32, "base_model.model.language_model.model.layers.34.self_attn.q_proj": 16, "base_model.model.language_model.model.layers.34.self_attn.k_proj": 64, "base_model.model.language_model.model.layers.34.self_attn.v_proj": 64, "base_model.model.language_model.model.layers.34.self_attn.o_proj": 8, "base_model.model.language_model.model.layers.34.mlp.gate_proj": 32, "base_model.model.language_model.model.layers.34.mlp.up_proj": 64, "base_model.model.language_model.model.layers.34.mlp.down_proj": 4, "base_model.model.language_model.model.layers.35.self_attn.q_proj": 16, "base_model.model.language_model.model.layers.35.self_attn.k_proj": 8, "base_model.model.language_model.model.layers.35.self_attn.v_proj": 4, "base_model.model.language_model.model.layers.35.self_attn.o_proj": 64, "base_model.model.language_model.model.layers.35.mlp.gate_proj": 64, "base_model.model.language_model.model.layers.35.mlp.up_proj": 64, "base_model.model.language_model.model.layers.35.mlp.down_proj": 32, "base_model.model.language_model.model.layers.36.self_attn.q_proj": 64, "base_model.model.language_model.model.layers.36.self_attn.k_proj": 16, "base_model.model.language_model.model.layers.36.self_attn.v_proj": 16, "base_model.model.language_model.model.layers.36.self_attn.o_proj": 8, "base_model.model.language_model.model.layers.36.mlp.gate_proj": 4, "base_model.model.language_model.model.layers.36.mlp.up_proj": 8, "base_model.model.language_model.model.layers.36.mlp.down_proj": 8, "base_model.model.language_model.model.layers.37.self_attn.q_proj": 16, "base_model.model.language_model.model.layers.37.self_attn.k_proj": 16, "base_model.model.language_model.model.layers.37.self_attn.v_proj": 4, "base_model.model.language_model.model.layers.37.self_attn.o_proj": 8, "base_model.model.language_model.model.layers.37.mlp.gate_proj": 8, "base_model.model.language_model.model.layers.37.mlp.up_proj": 8, "base_model.model.language_model.model.layers.37.mlp.down_proj": 32, "base_model.model.language_model.model.layers.38.cross_attn.q_proj": 16, "base_model.model.language_model.model.layers.38.cross_attn.k_proj": 4, "base_model.model.language_model.model.layers.38.cross_attn.v_proj": 8, "base_model.model.language_model.model.layers.38.cross_attn.o_proj": 8, "base_model.model.language_model.model.layers.38.mlp.gate_proj": 16, "base_model.model.language_model.model.layers.38.mlp.up_proj": 16, "base_model.model.language_model.model.layers.38.mlp.down_proj": 16, "base_model.model.language_model.model.layers.39.self_attn.q_proj": 64, "base_model.model.language_model.model.layers.39.self_attn.k_proj": 64, "base_model.model.language_model.model.layers.39.self_attn.v_proj": 64, "base_model.model.language_model.model.layers.39.self_attn.o_proj": 64, "base_model.model.language_model.model.layers.39.mlp.gate_proj": 16, "base_model.model.language_model.model.layers.39.mlp.up_proj": 16, "base_model.model.language_model.model.layers.39.mlp.down_proj": 4}