Training in progress, epoch 1

adapter_config.json

@@ -13,7 +13,7 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 32,
+  "lora_alpha": 256,
   "lora_dropout": 0.1,
   "megatron_config": null,
   "megatron_core": "megatron.core",
@@ -21,14 +21,14 @@
     "ln_f"
   ],
   "peft_type": "LORA",
-  "r": 32,
+  "r": 256,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
-    "k_proj",
     "q_proj",
-    "o_proj"
+    "v_proj",
+    "o_proj",
+    "k_proj"
   ],
   "task_type": null,
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c3af0be7d072ccadf2fa0728c23fc419485f4f0d7dba01ef06a581f5242d7627
-size 109086416
+oid sha256:61d476f41a6679068897bf00d28dddbd750e9e34bc158c6c17c49950a0cf3306
+size 872450448

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2ec68deb6f4c2c61fe1806c17cdbce09cb365c37da9457f249f40d427be332a7
+oid sha256:b4420c811a7e7092445ec2f0c48f731755df0493c799e4d76a24243cc698c481
 size 5048