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:f3c111696910d0898e6023003d924379e2bdb7972e353309d0823b89f355a50e
-size 109086416
+oid sha256:a697ce50dcf8b69e6d588b597ceefcac3ddd4d9a1a90ec5b4597e54e831e9154
+size 872450448

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:293c77db6ba2c1b2bfc98d11a5b861c3f73d7577a0470da7170a0064e87f58fb
+oid sha256:d919a41c59d71c74dc57b5639615e26245002961b97871cd950e79fd3a8f7a67
 size 5048