Training in progress, step 1000

adapter_config.json

@@ -20,45 +20,51 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "model.layers.21.self_attention.kv_proj",
-    "model.lstm_gates.1.forget_gate",
-    "model.layers.23.self_attention.kv_proj",
-    "model.layers.16.self_attention.kv_proj",
-    "model.layers.13.self_attention.kv_proj",
-    "model.layers.17.self_attention.kv_proj",
-    "model.layers.1.mlp.router.layer",
-    "model.layers.19.self_attention.kv_proj",
-    "model.lstm_gates.0.forget_gate",
-    "model.layers.18.mlp.router.layer",
-    "model.layers.1.mlp.router.perturbation",
-    "model.layers.20.self_attention.kv_proj",
-    "model.lstm_gates.0.input_gate",
-    "model.layers.10.self_attention.kv_proj",
-    "model.layers.1.self_attention.kv_proj",
-    "model.layers.22.self_attention.kv_proj",
-    "model.layers.14.self_attention.kv_proj",
-    "model.layers.18.self_attention.experts.router.layer",
-    "model.lstm_gates.1.input_gate",
-    "model.lstm_gates.0.output_gate",
-    "model.layers.6.self_attention.kv_proj",
+    "model.layers.19.self_attention.experts.router.layer",
     "model.layers.2.self_attention.kv_proj",
+    "model.layers.19.mlp.router.layer",
+    "model.lstm_gates.0.output_gate",
+    "model.layers.8.self_attention.kv_proj",
     "model.lstm_gates.1.transform",
+    "model.lstm_gates.2.output_gate",
+    "model.lstm_gates.2.forget_gate",
+    "model.lstm_gates.0.forget_gate",
+    "model.lstm_gates.1.output_gate",
+    "model.lstm_gates.2.transform",
+    "model.lstm_gates.0.transform",
+    "model.layers.17.self_attention.kv_proj",
+    "model.layers.20.self_attention.experts.router.layer",
+    "model.layers.16.self_attention.kv_proj",
+    "model.layers.1.self_attention.experts.router.layer",
+    "model.layers.18.self_attention.kv_proj",
     "model.layers.0.self_attention.kv_proj",
+    "model.layers.14.self_attention.kv_proj",
+    "model.lstm_gates.1.input_gate",
+    "model.layers.7.self_attention.kv_proj",
     "model.layers.11.self_attention.kv_proj",
-    "model.lstm_gates.1.output_gate",
-    "model.layers.3.self_attention.kv_proj",
+    "model.layers.10.self_attention.kv_proj",
+    "model.layers.22.self_attention.kv_proj",
     "model.layers.9.self_attention.kv_proj",
+    "model.lstm_gates.0.input_gate",
+    "model.layers.21.self_attention.kv_proj",
+    "model.layers.13.self_attention.kv_proj",
     "model.layers.15.self_attention.kv_proj",
-    "model.layers.19.mlp.router.layer",
-    "model.layers.4.self_attention.kv_proj",
-    "model.layers.18.self_attention.kv_proj",
-    "model.layers.19.self_attention.experts.router.layer",
-    "model.layers.7.self_attention.kv_proj",
+    "model.layers.23.self_attention.kv_proj",
+    "model.layers.19.self_attention.kv_proj",
+    "model.lstm_gates.2.input_gate",
+    "model.layers.1.mlp.router.perturbation",
+    "model.layers.1.mlp.router.layer",
     "model.layers.5.self_attention.kv_proj",
-    "model.layers.1.self_attention.experts.router.layer",
+    "model.layers.6.self_attention.kv_proj",
+    "model.lstm_gates.1.forget_gate",
+    "model.layers.4.self_attention.kv_proj",
+    "model.layers.1.self_attention.kv_proj",
     "model.layers.12.self_attention.kv_proj",
-    "model.lstm_gates.0.transform",
-    "model.layers.8.self_attention.kv_proj"
+    "model.layers.20.self_attention.kv_proj",
+    "model.layers.21.self_attention.experts.router.layer",
+    "model.layers.3.self_attention.kv_proj",
+    "model.layers.21.mlp.router.layer",
+    "model.layers.20.mlp.router.layer"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a3180dde5b242a076dbaa1a3d3a4b22e8c70d6da569138f24e8f9b0fd942392b
-size 6303624
+oid sha256:ded2926b6b32fd0682fb44888d3ebdc37da2b930dc27be85394d6db947b5153a
+size 6961088

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1500b2d7c5faffbdf54bd0ff84aa5933c91b72fa6ea95d39011921c13619bc60
+oid sha256:9cff8203b4dc3d3e94a95b8019ebd13fa2bda3b26e55f5c21bd9e277b8bee023
 size 5176