Training in progress, step 50

adapter_config.json

@@ -11,19 +11,19 @@
   "layers_to_transform": null,
   "loftq_config": {},
   "lora_alpha": 32,
-  "lora_dropout": 0.01,
+  "lora_dropout": 0.025,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": null,
   "peft_type": "LORA",
-  "r": 4,
+  "r": 8,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
-    "k_proj",
     "o_proj",
-    "v_proj"
+    "k_proj",
+    "v_proj",
+    "q_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:c174ea3d56f08d4e2bf2d8a884c39db1e813c9d5935a4fd40fe7a216c74d158c
-size 3705248
+oid sha256:c2488f27e0907f64b86802b503c973a426c62452c6fcea6059cabd9ba3ca3558
+size 7391688

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f758fb48f754a94281022b68ad5f311d08913faf065ec84b1eae43451832721f
+oid sha256:dac4e87813873930e2b3d68a63e8a35c875c8e2879524b4917cfa3eec2b2d6d2
 size 4984