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 sha256:3eeb6e32bcbb2735be41cdbb97a55411e9f30e3be1f7fa2d03abdd369c65475c
-size 109086416
+oid sha256:53b02041249d976403bf135f167f0197d44449a138db51129d366ed55223de1d
+size 872450448

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:35519dfa348d7e2f59eeabdc948da0eaefc267bcb28ef557cf45573e8e0d53a2
+oid sha256:f9634cbfc47a0993f4633f8fdd1d502266b24210ed76cdfbcfd53606dc8ef60d
 size 5048