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": [
-    "k_proj",
-    "v_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 sha256:2612d5383841d53b282a849cd02b894b5aa3bc26ce03402f8355a18604b293f0
-size 109086416
+oid sha256:2a7d1e3217e09e9fdef982d760a9b10553f5fb8c49ff63b389f85ee9d953b8b5
+size 872450448

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:376e2b310902a4b6d757829e5e30b4e35a56309224a5b7a78d15e9480b07e1ba
+oid sha256:4c28e93613b08ddbceb17a1dd1d92941fd5d0c737b7c2835326ae2a99f0e4191
 size 5048