Training in progress, epoch 1

adapter_config.json

@@ -13,7 +13,7 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 16,
+  "lora_alpha": 128,
   "lora_dropout": 0.1,
   "megatron_config": null,
   "megatron_core": "megatron.core",
@@ -22,12 +22,12 @@
     "qa_outputs.weight"
   ],
   "peft_type": "LORA",
-  "r": 16,
+  "r": 128,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "value",
-    "query"
+    "query",
+    "value"
   ],
   "task_type": null,
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3c41e1906937c6ad63e30cf74509e151d038fbf1cea23263312036b447bebf1e
-size 2366064
+oid sha256:e4cbc1e129d46ba6f9831f5d20f537533197541a05ef01307017993ef0819b1f
+size 18881272

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0d80ff231ba30be219707274e155a0abb6aa27dbf084fcb86b9b1ac1e4f1b020
+oid sha256:dd2ed253e243495f4e97f0036b4a4f6a5630877eebaf33e9ee420f16949594f2
 size 5368