Training in progress, step 100

adapter_config.json

@@ -23,374 +23,374 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "language_model.model.layers.23.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.2.mlp.fc1",
-    "language_model.model.layers.3.mlp.up_proj",
-    "language_model.model.layers.24.mlp.gate_proj",
-    "language_model.model.layers.1.mlp.down_proj",
+    "language_model.model.layers.26.self_attn.o_proj",
+    "language_model.model.layers.22.self_attn.o_proj",
+    "vision_tower.vision_model.encoder.layers.20.mlp.fc1",
+    "vision_tower.vision_model.encoder.layers.6.mlp.fc2",
+    "language_model.model.layers.4.mlp.gate_proj",
+    "language_model.model.layers.20.self_attn.o_proj",
+    "language_model.model.layers.10.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.out_proj",
+    "language_model.model.layers.27.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.k_proj",
+    "language_model.model.layers.19.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.out_proj",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.1.mlp.fc1",
+    "language_model.model.layers.2.self_attn.k_proj",
+    "language_model.model.layers.14.self_attn.q_proj",
     "language_model.model.layers.19.self_attn.q_proj",
-    "language_model.model.layers.21.mlp.up_proj",
-    "language_model.model.layers.0.self_attn.k_proj",
-    "language_model.model.layers.11.mlp.down_proj",
-    "language_model.model.layers.10.self_attn.o_proj",
-    "vision_tower.vision_model.encoder.layers.21.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.8.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.22.self_attn.q_proj",
-    "language_model.model.layers.23.mlp.down_proj",
-    "language_model.model.layers.12.self_attn.k_proj",
-    "language_model.model.layers.15.mlp.down_proj",
-    "language_model.model.layers.2.self_attn.o_proj",
-    "language_model.model.layers.5.self_attn.k_proj",
-    "language_model.model.layers.25.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.0.self_attn.out_proj",
-    "vision_tower.vision_model.encoder.layers.2.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.23.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.k_proj",
+    "language_model.model.layers.16.self_attn.k_proj",
+    "language_model.model.layers.9.self_attn.o_proj",
+    "language_model.model.layers.10.self_attn.v_proj",
+    "language_model.model.layers.4.mlp.down_proj",
     "language_model.model.layers.18.self_attn.q_proj",
+    "language_model.model.layers.26.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.v_proj",
+    "language_model.model.layers.22.mlp.down_proj",
+    "language_model.model.layers.20.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.14.mlp.fc2",
+    "language_model.model.layers.14.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.0.mlp.fc1",
+    "language_model.model.layers.26.mlp.down_proj",
+    "language_model.model.layers.5.mlp.up_proj",
+    "language_model.model.layers.11.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.7.self_attn.v_proj",
+    "language_model.model.layers.30.self_attn.o_proj",
+    "language_model.model.layers.21.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.q_proj",
+    "language_model.model.layers.0.mlp.down_proj",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.5.self_attn.q_proj",
+    "language_model.model.layers.15.self_attn.v_proj",
+    "language_model.model.layers.11.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.3.mlp.fc1",
+    "language_model.model.layers.18.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.v_proj",
+    "language_model.model.layers.25.self_attn.o_proj",
+    "language_model.model.layers.4.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.q_proj",
+    "language_model.model.layers.14.self_attn.o_proj",
+    "language_model.model.layers.28.self_attn.o_proj",
+    "vision_tower.vision_model.encoder.layers.11.self_attn.v_proj",
+    "language_model.model.layers.15.self_attn.k_proj",
+    "language_model.model.layers.20.mlp.down_proj",
+    "vision_tower.vision_model.encoder.layers.18.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.17.mlp.fc1",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.q_proj",
+    "language_model.model.layers.23.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.15.mlp.fc2",
     "vision_tower.vision_model.encoder.layers.14.mlp.fc1",
-    "language_model.model.layers.14.mlp.down_proj",
-    "language_model.model.layers.9.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.1.mlp.fc1",
-    "language_model.model.layers.20.self_attn.q_proj",
-    "language_model.model.layers.7.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.12.self_attn.out_proj",
+    "language_model.model.layers.5.mlp.down_proj",
+    "language_model.model.layers.29.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.18.mlp.fc1",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.out_proj",
     "language_model.model.layers.5.self_attn.v_proj",
-    "language_model.model.layers.22.self_attn.q_proj",
+    "language_model.model.layers.28.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.11.mlp.fc2",
+    "language_model.model.layers.20.mlp.gate_proj",
+    "language_model.model.layers.28.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.1.mlp.fc2",
+    "language_model.model.layers.17.self_attn.o_proj",
+    "language_model.model.layers.30.self_attn.q_proj",
+    "language_model.model.layers.27.mlp.gate_proj",
+    "language_model.model.layers.11.mlp.gate_proj",
+    "language_model.model.layers.6.self_attn.o_proj",
+    "language_model.model.layers.26.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.5.self_attn.v_proj",
+    "language_model.model.layers.29.mlp.gate_proj",
+    "language_model.model.layers.31.mlp.gate_proj",
+    "language_model.model.layers.23.self_attn.v_proj",
+    "language_model.model.layers.16.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.9.mlp.fc2",
     "language_model.model.layers.4.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.19.self_attn.q_proj",
+    "language_model.model.layers.16.mlp.down_proj",
+    "language_model.model.layers.11.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.out_proj",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.out_proj",
+    "language_model.model.layers.25.mlp.gate_proj",
+    "language_model.model.layers.1.self_attn.k_proj",
+    "language_model.model.layers.9.self_attn.q_proj",
+    "language_model.model.layers.14.mlp.gate_proj",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.v_proj",
+    "language_model.model.layers.12.mlp.up_proj",
+    "language_model.model.layers.27.self_attn.o_proj",
+    "language_model.model.layers.9.mlp.gate_proj",
+    "vision_tower.vision_model.encoder.layers.11.mlp.fc1",
+    "language_model.model.layers.29.mlp.down_proj",
+    "language_model.model.layers.20.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.7.self_attn.out_proj",
+    "language_model.model.layers.27.self_attn.q_proj",
+    "language_model.model.layers.13.self_attn.o_proj",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.q_proj",
+    "language_model.model.layers.21.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.13.self_attn.out_proj",
     "vision_tower.vision_model.encoder.layers.20.self_attn.q_proj",
+    "language_model.model.layers.8.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.16.mlp.fc1",
     "language_model.model.layers.28.mlp.down_proj",
-    "language_model.model.layers.10.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.22.self_attn.out_proj",
-    "vision_tower.vision_model.encoder.layers.5.self_attn.out_proj",
-    "vision_tower.vision_model.encoder.layers.4.self_attn.q_proj",
-    "language_model.model.layers.28.self_attn.v_proj",
-    "language_model.model.layers.18.mlp.down_proj",
-    "vision_tower.vision_model.encoder.layers.17.self_attn.v_proj",
-    "language_model.model.layers.2.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.5.mlp.fc1",
-    "language_model.model.layers.12.self_attn.o_proj",
-    "language_model.model.layers.7.self_attn.o_proj",
-    "language_model.model.layers.3.mlp.gate_proj",
-    "language_model.model.layers.27.mlp.up_proj",
-    "language_model.model.layers.8.mlp.up_proj",
-    "language_model.model.layers.16.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.12.self_attn.q_proj",
+    "language_model.model.layers.31.self_attn.o_proj",
+    "language_model.model.layers.1.mlp.down_proj",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.q_proj",
+    "language_model.model.layers.15.mlp.gate_proj",
+    "language_model.model.layers.15.mlp.down_proj",
+    "vision_tower.vision_model.encoder.layers.8.mlp.fc2",
+    "language_model.model.layers.8.mlp.gate_proj",
+    "language_model.model.layers.0.mlp.gate_proj",
+    "language_model.model.layers.3.self_attn.q_proj",
     "language_model.model.layers.24.self_attn.q_proj",
-    "language_model.model.layers.14.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.19.mlp.fc2",
     "vision_tower.vision_model.encoder.layers.12.mlp.fc1",
-    "vision_tower.vision_model.encoder.layers.16.self_attn.out_proj",
-    "vision_tower.vision_model.encoder.layers.1.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.18.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.4.self_attn.out_proj",
-    "language_model.model.layers.26.self_attn.k_proj",
-    "language_model.model.layers.0.self_attn.v_proj",
+    "language_model.model.layers.16.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.6.self_attn.q_proj",
+    "language_model.model.layers.31.self_attn.v_proj",
+    "language_model.model.layers.3.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.9.mlp.fc1",
+    "language_model.model.layers.10.self_attn.q_proj",
+    "language_model.model.layers.25.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.7.mlp.fc2",
+    "vision_tower.vision_model.encoder.layers.20.mlp.fc2",
+    "language_model.model.layers.8.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.q_proj",
+    "language_model.model.layers.21.mlp.down_proj",
+    "language_model.model.layers.26.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.6.self_attn.k_proj",
+    "language_model.model.layers.12.mlp.down_proj",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.out_proj",
+    "language_model.model.layers.13.mlp.up_proj",
+    "language_model.model.layers.7.mlp.down_proj",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.out_proj",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.out_proj",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.v_proj",
     "language_model.model.layers.5.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.out_proj",
+    "language_model.model.layers.27.mlp.down_proj",
+    "language_model.model.layers.2.mlp.down_proj",
+    "language_model.model.layers.19.mlp.gate_proj",
+    "language_model.model.layers.22.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.k_proj",
+    "language_model.model.layers.26.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.13.mlp.fc1",
+    "language_model.model.layers.23.self_attn.o_proj",
+    "vision_tower.vision_model.encoder.layers.21.mlp.fc2",
+    "language_model.model.layers.14.self_attn.v_proj",
+    "language_model.model.layers.4.self_attn.o_proj",
     "language_model.model.layers.17.self_attn.v_proj",
-    "language_model.model.layers.31.self_attn.o_proj",
-    "language_model.model.layers.5.mlp.gate_proj",
-    "language_model.model.layers.23.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.12.self_attn.v_proj",
+    "language_model.model.layers.8.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.5.mlp.fc1",
+    "language_model.model.layers.10.self_attn.o_proj",
+    "language_model.model.layers.18.self_attn.k_proj",
     "language_model.model.layers.18.mlp.gate_proj",
-    "language_model.model.layers.8.self_attn.k_proj",
-    "language_model.model.layers.31.self_attn.k_proj",
-    "language_model.model.layers.24.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.19.self_attn.out_proj",
-    "vision_tower.vision_model.encoder.layers.15.mlp.fc1",
-    "language_model.model.layers.20.self_attn.k_proj",
-    "language_model.model.layers.9.mlp.gate_proj",
-    "language_model.model.layers.3.self_attn.q_proj",
-    "language_model.model.layers.30.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.3.self_attn.out_proj",
-    "language_model.model.layers.0.mlp.gate_proj",
-    "language_model.model.layers.6.mlp.gate_proj",
-    "language_model.model.layers.16.self_attn.o_proj",
-    "language_model.model.layers.29.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.15.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.15.self_attn.k_proj",
-    "language_model.model.layers.7.mlp.up_proj",
-    "language_model.model.layers.3.mlp.down_proj",
-    "language_model.model.layers.16.self_attn.k_proj",
-    "language_model.model.layers.22.mlp.up_proj",
-    "language_model.model.layers.8.mlp.gate_proj",
-    "language_model.model.layers.9.self_attn.o_proj",
-    "vision_tower.vision_model.encoder.layers.8.self_attn.out_proj",
-    "language_model.model.layers.24.self_attn.k_proj",
-    "language_model.model.layers.25.mlp.down_proj",
-    "language_model.model.layers.2.self_attn.k_proj",
+    "language_model.model.layers.29.self_attn.o_proj",
+    "language_model.model.layers.3.mlp.gate_proj",
+    "language_model.model.layers.9.mlp.up_proj",
+    "language_model.model.layers.3.self_attn.o_proj",
+    "language_model.model.layers.21.self_attn.o_proj",
+    "language_model.model.layers.15.self_attn.o_proj",
+    "language_model.model.layers.0.self_attn.o_proj",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.20.self_attn.v_proj",
     "vision_tower.vision_model.encoder.layers.11.self_attn.out_proj",
-    "vision_tower.vision_model.encoder.layers.23.self_attn.out_proj",
-    "vision_tower.vision_model.encoder.layers.8.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.6.mlp.fc1",
-    "language_model.model.layers.1.self_attn.o_proj",
-    "vision_tower.vision_model.encoder.layers.0.self_attn.q_proj",
-    "language_model.model.layers.11.self_attn.o_proj",
-    "language_model.model.layers.31.mlp.down_proj",
-    "vision_tower.vision_model.encoder.layers.23.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.22.mlp.fc1",
-    "vision_tower.vision_model.encoder.layers.14.self_attn.v_proj",
-    "language_model.model.layers.20.mlp.gate_proj",
-    "language_model.model.layers.21.self_attn.k_proj",
-    "language_model.model.layers.16.mlp.gate_proj",
-    "language_model.model.layers.1.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.4.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.21.mlp.fc1",
-    "vision_tower.vision_model.encoder.layers.9.mlp.fc1",
-    "vision_tower.vision_model.encoder.layers.11.self_attn.v_proj",
-    "language_model.model.layers.22.mlp.gate_proj",
-    "language_model.model.layers.2.mlp.down_proj",
-    "language_model.model.layers.21.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.23.self_attn.q_proj",
+    "language_model.model.layers.0.self_attn.k_proj",
+    "language_model.model.layers.25.self_attn.v_proj",
+    "language_model.model.layers.0.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.out_proj",
+    "language_model.model.layers.27.mlp.up_proj",
+    "language_model.model.layers.14.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.5.mlp.fc2",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.k_proj",
+    "language_model.model.layers.21.mlp.gate_proj",
+    "language_model.model.layers.22.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.18.self_attn.out_proj",
+    "vision_tower.vision_model.encoder.layers.10.mlp.fc1",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.out_proj",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.v_proj",
+    "language_model.model.layers.23.mlp.up_proj",
+    "language_model.model.layers.8.mlp.down_proj",
+    "language_model.model.layers.17.mlp.gate_proj",
+    "language_model.model.layers.24.mlp.gate_proj",
+    "language_model.model.layers.10.mlp.down_proj",
+    "language_model.model.layers.31.mlp.down_proj",
     "vision_tower.vision_model.encoder.layers.23.mlp.fc1",
-    "language_model.model.layers.3.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.18.mlp.fc1",
+    "language_model.model.layers.19.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.k_proj",
+    "language_model.model.layers.12.self_attn.q_proj",
+    "language_model.model.layers.2.self_attn.q_proj",
+    "language_model.model.layers.0.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.20.self_attn.out_proj",
+    "language_model.model.layers.19.self_attn.o_proj",
+    "language_model.model.layers.25.mlp.down_proj",
+    "language_model.model.layers.17.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.11.self_attn.k_proj",
+    "language_model.model.layers.22.mlp.gate_proj",
+    "language_model.model.layers.30.mlp.down_proj",
+    "language_model.model.layers.15.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.0.mlp.fc2",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.k_proj",
+    "language_model.model.layers.7.mlp.gate_proj",
     "vision_tower.vision_model.encoder.layers.18.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.10.mlp.fc2",
-    "language_model.model.layers.29.mlp.up_proj",
-    "language_model.model.layers.27.self_attn.o_proj",
-    "language_model.model.layers.8.self_attn.v_proj",
-    "language_model.model.layers.25.self_attn.o_proj",
-    "language_model.model.layers.26.mlp.down_proj",
-    "vision_tower.vision_model.encoder.layers.22.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.21.self_attn.v_proj",
+    "language_model.model.layers.12.self_attn.o_proj",
+    "language_model.model.layers.1.self_attn.q_proj",
     "language_model.model.layers.26.mlp.gate_proj",
-    "language_model.model.layers.13.mlp.down_proj",
-    "language_model.model.layers.20.mlp.up_proj",
-    "language_model.model.layers.19.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.16.mlp.fc1",
-    "vision_tower.vision_model.encoder.layers.4.self_attn.v_proj",
-    "language_model.model.layers.9.self_attn.q_proj",
-    "language_model.model.layers.15.self_attn.q_proj",
-    "language_model.model.layers.15.self_attn.o_proj",
-    "language_model.model.layers.28.self_attn.o_proj",
-    "vision_tower.vision_model.encoder.layers.12.mlp.fc2",
-    "language_model.model.layers.14.self_attn.v_proj",
-    "language_model.model.layers.23.mlp.gate_proj",
-    "vision_tower.vision_model.encoder.layers.18.self_attn.q_proj",
-    "language_model.model.layers.23.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.11.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.6.self_attn.k_proj",
-    "language_model.model.layers.20.mlp.down_proj",
-    "language_model.model.layers.5.self_attn.o_proj",
-    "language_model.model.layers.25.mlp.gate_proj",
-    "language_model.model.layers.18.mlp.up_proj",
-    "language_model.model.layers.14.mlp.up_proj",
-    "language_model.model.layers.12.mlp.gate_proj",
+    "language_model.model.layers.17.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.23.mlp.fc2",
+    "language_model.model.layers.5.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.19.mlp.fc1",
+    "language_model.model.layers.1.self_attn.v_proj",
+    "language_model.model.layers.17.mlp.down_proj",
+    "language_model.model.layers.11.mlp.down_proj",
+    "language_model.model.layers.6.mlp.gate_proj",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.out_proj",
+    "language_model.model.layers.31.self_attn.q_proj",
+    "language_model.model.layers.8.self_attn.o_proj",
+    "vision_tower.vision_model.encoder.layers.6.self_attn.out_proj",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.out_proj",
+    "language_model.model.layers.17.self_attn.k_proj",
+    "language_model.model.layers.31.mlp.up_proj",
+    "language_model.model.layers.16.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.10.mlp.fc2",
     "language_model.model.layers.0.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.0.mlp.fc2",
-    "language_model.model.layers.5.mlp.down_proj",
-    "language_model.model.layers.17.mlp.gate_proj",
-    "vision_tower.vision_model.encoder.layers.4.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.2.self_attn.out_proj",
+    "language_model.model.layers.21.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.15.mlp.fc1",
+    "language_model.model.layers.12.mlp.gate_proj",
+    "vision_tower.vision_model.encoder.layers.16.mlp.fc2",
+    "vision_tower.vision_model.encoder.layers.2.mlp.fc2",
+    "language_model.model.layers.30.mlp.gate_proj",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.k_proj",
     "language_model.model.layers.25.mlp.up_proj",
-    "language_model.model.layers.22.mlp.down_proj",
-    "language_model.model.layers.28.self_attn.q_proj",
-    "language_model.model.layers.21.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.4.mlp.fc1",
-    "vision_tower.vision_model.encoder.layers.0.mlp.fc1",
-    "vision_tower.vision_model.encoder.layers.1.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.19.mlp.fc1",
-    "language_model.model.layers.30.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.13.mlp.fc1",
-    "language_model.model.layers.13.mlp.gate_proj",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.q_proj",
+    "language_model.model.layers.3.mlp.down_proj",
+    "vision_tower.vision_model.encoder.layers.5.self_attn.out_proj",
     "language_model.model.layers.30.self_attn.k_proj",
-    "language_model.model.layers.12.self_attn.q_proj",
-    "language_model.model.layers.25.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.20.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.13.self_attn.out_proj",
-    "vision_tower.vision_model.encoder.layers.10.self_attn.k_proj",
-    "language_model.model.layers.26.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.13.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.15.self_attn.v_proj",
-    "language_model.model.layers.15.self_attn.k_proj",
-    "language_model.model.layers.10.self_attn.v_proj",
-    "language_model.model.layers.8.self_attn.o_proj",
-    "language_model.model.layers.29.mlp.down_proj",
-    "vision_tower.vision_model.encoder.layers.19.self_attn.k_proj",
-    "language_model.model.layers.8.mlp.down_proj",
-    "vision_tower.vision_model.encoder.layers.8.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.17.mlp.fc2",
-    "language_model.model.layers.23.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.7.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.20.self_attn.out_proj",
-    "language_model.model.layers.14.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.21.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.1.self_attn.out_proj",
-    "language_model.model.layers.6.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.3.mlp.fc1",
-    "vision_tower.vision_model.encoder.layers.15.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.8.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.11.self_attn.k_proj",
-    "language_model.model.layers.17.self_attn.o_proj",
+    "language_model.model.layers.2.self_attn.v_proj",
+    "language_model.model.layers.13.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.2.mlp.fc1",
+    "vision_tower.vision_model.encoder.layers.18.self_attn.q_proj",
+    "language_model.model.layers.30.self_attn.v_proj",
     "vision_tower.vision_model.encoder.layers.7.mlp.fc1",
-    "vision_tower.vision_model.encoder.layers.6.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.9.self_attn.q_proj",
-    "language_model.model.layers.28.mlp.gate_proj",
-    "language_model.model.layers.0.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.18.self_attn.out_proj",
-    "vision_tower.vision_model.encoder.layers.9.self_attn.out_proj",
-    "language_model.model.layers.5.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.3.self_attn.q_proj",
-    "language_model.model.layers.9.self_attn.v_proj",
-    "language_model.model.layers.29.mlp.gate_proj",
     "language_model.model.layers.9.mlp.down_proj",
-    "vision_tower.vision_model.encoder.layers.15.self_attn.out_proj",
-    "language_model.model.layers.7.self_attn.q_proj",
-    "language_model.model.layers.21.mlp.gate_proj",
-    "language_model.model.layers.7.self_attn.k_proj",
-    "language_model.model.layers.18.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.17.self_attn.k_proj",
-    "language_model.model.layers.19.self_attn.o_proj",
-    "language_model.model.layers.4.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.21.self_attn.out_proj",
-    "language_model.model.layers.2.mlp.gate_proj",
-    "language_model.model.layers.10.mlp.up_proj",
-    "language_model.model.layers.6.self_attn.q_proj",
-    "language_model.model.layers.30.self_attn.q_proj",
-    "language_model.model.layers.9.self_attn.k_proj",
-    "language_model.model.layers.27.self_attn.q_proj",
-    "language_model.model.layers.6.self_attn.v_proj",
-    "language_model.model.layers.16.mlp.up_proj",
-    "language_model.model.layers.21.mlp.down_proj",
-    "language_model.model.layers.4.self_attn.o_proj",
+    "language_model.model.layers.12.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.out_proj",
+    "language_model.model.layers.11.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.17.mlp.fc2",
+    "language_model.model.layers.10.mlp.gate_proj",
+    "vision_tower.vision_model.encoder.layers.4.mlp.fc2",
+    "vision_tower.vision_model.encoder.layers.13.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.20.self_attn.k_proj",
+    "language_model.model.layers.23.self_attn.k_proj",
+    "language_model.model.layers.5.self_attn.o_proj",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.k_proj",
+    "language_model.model.layers.20.self_attn.q_proj",
+    "language_model.model.layers.1.self_attn.o_proj",
+    "language_model.model.layers.24.self_attn.o_proj",
     "vision_tower.vision_model.encoder.layers.13.mlp.fc2",
-    "language_model.model.layers.11.mlp.gate_proj",
-    "vision_tower.vision_model.encoder.layers.0.self_attn.k_proj",
-    "language_model.model.layers.28.mlp.up_proj",
-    "language_model.model.layers.31.mlp.gate_proj",
-    "vision_tower.vision_model.encoder.layers.8.mlp.fc1",
-    "vision_tower.vision_model.encoder.layers.6.self_attn.out_proj",
-    "vision_tower.vision_model.encoder.layers.5.self_attn.v_proj",
-    "language_model.model.layers.20.self_attn.o_proj",
-    "language_model.model.layers.31.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.7.self_attn.out_proj",
-    "language_model.model.layers.29.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.14.self_attn.out_proj",
-    "vision_tower.vision_model.encoder.layers.9.self_attn.v_proj",
-    "language_model.model.layers.13.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.16.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.11.mlp.fc1",
+    "language_model.model.layers.6.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.q_proj",
+    "language_model.model.layers.24.self_attn.v_proj",
     "language_model.model.layers.2.mlp.up_proj",
-    "language_model.model.layers.17.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.10.mlp.fc1",
-    "language_model.model.layers.3.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.1.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.9.mlp.fc2",
     "vision_tower.vision_model.encoder.layers.13.self_attn.v_proj",
-    "language_model.model.layers.22.self_attn.v_proj",
-    "language_model.model.layers.0.mlp.down_proj",
-    "language_model.model.layers.16.mlp.down_proj",
-    "vision_tower.vision_model.encoder.layers.7.mlp.fc2",
-    "language_model.model.layers.4.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.3.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.19.mlp.fc2",
-    "language_model.model.layers.20.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.19.self_attn.v_proj",
-    "language_model.model.layers.1.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.q_proj",
     "vision_tower.vision_model.encoder.layers.7.self_attn.k_proj",
-    "language_model.model.layers.18.self_attn.o_proj",
-    "language_model.model.layers.27.self_attn.k_proj",
+    "language_model.model.layers.24.mlp.down_proj",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.v_proj",
+    "language_model.model.layers.13.self_attn.v_proj",
+    "language_model.model.layers.3.self_attn.k_proj",
+    "language_model.model.layers.28.mlp.gate_proj",
+    "language_model.model.layers.11.self_attn.o_proj",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.v_proj",
+    "language_model.model.layers.7.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.12.mlp.fc2",
+    "language_model.model.layers.3.self_attn.v_proj",
+    "language_model.model.layers.12.self_attn.v_proj",
+    "language_model.model.layers.27.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.k_proj",
+    "language_model.model.layers.7.self_attn.o_proj",
+    "language_model.model.layers.14.mlp.down_proj",
+    "vision_tower.vision_model.encoder.layers.8.mlp.fc1",
+    "language_model.model.layers.6.mlp.up_proj",
+    "language_model.model.layers.2.self_attn.o_proj",
+    "language_model.model.layers.20.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.q_proj",
     "vision_tower.vision_model.encoder.layers.5.self_attn.k_proj",
-    "language_model.model.layers.6.self_attn.k_proj",
-    "language_model.model.layers.22.self_attn.o_proj",
-    "vision_tower.vision_model.encoder.layers.17.self_attn.out_proj",
-    "language_model.model.layers.19.mlp.up_proj",
-    "language_model.model.layers.15.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.12.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.14.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.14.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.17.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.v_proj",
     "language_model.model.layers.22.self_attn.k_proj",
-    "language_model.model.layers.10.mlp.down_proj",
-    "vision_tower.vision_model.encoder.layers.14.mlp.fc2",
-    "language_model.model.layers.14.mlp.gate_proj",
-    "language_model.model.layers.25.self_attn.q_proj",
-    "language_model.model.layers.29.self_attn.q_proj",
-    "language_model.model.layers.6.mlp.down_proj",
-    "vision_tower.vision_model.encoder.layers.16.self_attn.q_proj",
-    "language_model.model.layers.2.self_attn.q_proj",
-    "language_model.model.layers.17.mlp.up_proj",
-    "language_model.model.layers.26.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.11.mlp.fc2",
-    "language_model.model.layers.26.self_attn.o_proj",
-    "vision_tower.vision_model.encoder.layers.9.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.0.self_attn.v_proj",
-    "language_model.model.layers.24.mlp.down_proj",
-    "language_model.model.layers.7.mlp.down_proj",
-    "language_model.model.layers.14.self_attn.o_proj",
-    "language_model.model.layers.18.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.2.self_attn.q_proj",
-    "language_model.model.layers.10.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.10.self_attn.out_proj",
-    "language_model.model.layers.7.mlp.gate_proj",
-    "language_model.model.layers.31.self_attn.q_proj",
-    "language_model.model.layers.6.self_attn.o_proj",
-    "vision_tower.vision_model.encoder.layers.2.mlp.fc2",
-    "language_model.model.layers.12.mlp.down_proj",
+    "language_model.model.layers.23.mlp.down_proj",
+    "language_model.model.layers.18.mlp.down_proj",
+    "language_model.model.layers.2.mlp.gate_proj",
+    "language_model.model.layers.29.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.13.self_attn.q_proj",
+    "language_model.model.layers.16.mlp.gate_proj",
+    "language_model.model.layers.4.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.22.mlp.fc1",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.out_proj",
+    "language_model.model.layers.23.mlp.gate_proj",
+    "language_model.model.layers.7.self_attn.q_proj",
+    "language_model.model.layers.19.mlp.down_proj",
+    "language_model.model.layers.13.mlp.gate_proj",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.out_proj",
+    "language_model.model.layers.8.mlp.up_proj",
     "vision_tower.vision_model.encoder.layers.6.self_attn.v_proj",
-    "language_model.model.layers.12.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.5.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.7.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.16.self_attn.k_proj",
-    "language_model.model.layers.30.self_attn.o_proj",
-    "vision_tower.vision_model.encoder.layers.6.self_attn.q_proj",
-    "language_model.model.layers.4.self_attn.k_proj",
-    "language_model.model.layers.31.self_attn.v_proj",
-    "language_model.model.layers.16.self_attn.q_proj",
-    "language_model.model.layers.11.self_attn.q_proj",
-    "language_model.model.layers.13.self_attn.o_proj",
-    "language_model.model.layers.30.mlp.gate_proj",
-    "vision_tower.vision_model.encoder.layers.20.mlp.fc1",
-    "language_model.model.layers.1.mlp.gate_proj",
-    "language_model.model.layers.17.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.21.self_attn.k_proj",
-    "language_model.model.layers.11.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.5.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.1.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.3.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.16.self_attn.v_proj",
-    "language_model.model.layers.23.self_attn.o_proj",
-    "language_model.model.layers.29.self_attn.o_proj",
-    "language_model.model.layers.21.self_attn.o_proj",
-    "language_model.model.layers.11.self_attn.v_proj",
-    "vision_tower.vision_model.encoder.layers.2.self_attn.k_proj",
-    "language_model.model.layers.1.self_attn.q_proj",
-    "language_model.model.layers.19.self_attn.v_proj",
-    "language_model.model.layers.1.mlp.up_proj",
-    "vision_tower.vision_model.encoder.layers.10.self_attn.v_proj",
-    "language_model.model.layers.30.mlp.down_proj",
-    "language_model.model.layers.26.mlp.up_proj",
-    "language_model.model.layers.27.self_attn.v_proj",
-    "language_model.model.layers.13.mlp.up_proj",
-    "language_model.model.layers.24.self_attn.o_proj",
-    "language_model.model.layers.4.mlp.down_proj",
-    "language_model.model.layers.15.self_attn.v_proj",
-    "language_model.model.layers.11.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.17.mlp.fc1",
-    "language_model.model.layers.10.mlp.gate_proj",
-    "language_model.model.layers.24.self_attn.v_proj",
-    "language_model.model.layers.27.mlp.gate_proj",
-    "vision_tower.vision_model.encoder.layers.20.self_attn.k_proj",
-    "language_model.model.layers.27.mlp.down_proj",
-    "language_model.model.layers.17.mlp.down_proj",
+    "language_model.model.layers.6.self_attn.q_proj",
     "vision_tower.vision_model.encoder.layers.22.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.20.mlp.fc2",
-    "vision_tower.vision_model.encoder.layers.18.self_attn.k_proj",
-    "language_model.model.layers.19.mlp.down_proj",
-    "vision_tower.vision_model.encoder.layers.22.self_attn.k_proj",
-    "language_model.model.layers.3.self_attn.o_proj",
+    "language_model.model.layers.28.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.18.mlp.fc2",
+    "language_model.model.layers.5.mlp.gate_proj",
+    "language_model.model.layers.25.self_attn.q_proj",
+    "language_model.model.layers.24.self_attn.k_proj",
+    "language_model.model.layers.31.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.4.mlp.fc1",
+    "language_model.model.layers.9.self_attn.v_proj",
+    "language_model.model.layers.19.self_attn.v_proj",
     "language_model.model.layers.13.self_attn.k_proj",
-    "language_model.model.layers.4.mlp.gate_proj",
-    "language_model.model.layers.12.self_attn.v_proj",
-    "language_model.model.layers.0.self_attn.o_proj",
-    "language_model.model.layers.13.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.3.self_attn.v_proj",
-    "language_model.model.layers.15.mlp.gate_proj",
-    "language_model.model.layers.19.mlp.gate_proj",
-    "vision_tower.vision_model.encoder.layers.23.mlp.fc2",
-    "language_model.model.layers.8.self_attn.q_proj",
-    "language_model.model.layers.28.self_attn.k_proj",
-    "vision_tower.vision_model.encoder.layers.10.self_attn.q_proj",
-    "vision_tower.vision_model.encoder.layers.13.self_attn.q_proj"
+    "vision_tower.vision_model.encoder.layers.21.mlp.fc1",
+    "language_model.model.layers.7.self_attn.v_proj",
+    "language_model.model.layers.30.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.11.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.q_proj",
+    "language_model.model.layers.7.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.6.mlp.fc1",
+    "language_model.model.layers.21.mlp.up_proj",
+    "language_model.model.layers.4.self_attn.k_proj",
+    "language_model.model.layers.28.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.q_proj",
+    "vision_tower.vision_model.encoder.layers.7.self_attn.q_proj",
+    "language_model.model.layers.29.self_attn.v_proj",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.k_proj",
+    "language_model.model.layers.1.mlp.up_proj",
+    "language_model.model.layers.6.mlp.down_proj",
+    "language_model.model.layers.13.mlp.down_proj",
+    "language_model.model.layers.18.self_attn.o_proj",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.q_proj",
+    "language_model.model.layers.29.self_attn.k_proj",
+    "vision_tower.vision_model.encoder.layers.3.mlp.fc2",
+    "language_model.model.layers.16.self_attn.o_proj",
+    "language_model.model.layers.10.mlp.up_proj",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.out_proj",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.k_proj",
+    "language_model.model.layers.9.self_attn.k_proj",
+    "language_model.model.layers.6.self_attn.k_proj",
+    "language_model.model.layers.15.self_attn.q_proj",
+    "language_model.model.layers.1.mlp.gate_proj",
+    "language_model.model.layers.22.self_attn.v_proj",
+    "language_model.model.layers.18.self_attn.v_proj",
+    "language_model.model.layers.24.mlp.up_proj"
   ],
   "task_type": null,
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4563a60061cb86fcd08ba8e63fa0c03a15a4624d3f2ff1b2e8c02f4115394a30
+oid sha256:ff4fcd8d9beeedfc683c5f77680eb6cff0ecfc412d20d5ec49a63b4d69611229
 size 784448728

runs/Sep13_02-42-49_014fada30694/events.out.tfevents.1726195418.014fada30694.460.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d9cf3d789107a84c89fd4a9a5ba277cb4b90c8e39273e6eb96be1da6d923090
+size 6612

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a160aa2cfd6193a46d7c711a655cb3ad2a4493aa0104e59631eacd6a545e0af5
+oid sha256:59f9826e1f8267937483f7be0cce677f94ff7b71d520d5bb9673146fce9e555b
 size 5176