Training in progress, step 400

adapter_config.json

@@ -14,7 +14,7 @@
   "loftq_config": {},
   "lora_alpha": 128,
   "lora_bias": false,
-  "lora_dropout": 0.05,
+  "lora_dropout": 0.01,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": null,
@@ -23,13 +23,11 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "o_proj",
     "v_proj",
     "q_proj",
-    "down_proj",
     "k_proj",
-    "up_proj",
-    "gate_proj"
+    "gate_proj",
+    "o_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2e88accc8844973f0fab71f4d7137f88d558b66ae93f0212c4a22671790be58b
-size 671149168
+oid sha256:b6a88725f365a90c9fce603931464ad59db50de2c384b790191cbf07fc091365
+size 369142184