Try lr 13e-5

README.md

@@ -4,8 +4,8 @@ library_name: transformers
 model_name: ilo-toki-gemmax2-28-2b
 tags:
 - generated_from_trainer
-- sft
 - trl
+- sft
 licence: license
 ---
 

adapter_config.json

@@ -25,13 +25,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
+    "up_proj",
     "down_proj",
-    "gate_proj",
-    "o_proj",
-    "k_proj",
     "q_proj",
+    "k_proj",
+    "gate_proj",
     "v_proj",
-    "up_proj"
+    "o_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e41df19783ffeabb30d0d862f5429135a6175176caa48946c94ed3fc67f76cd7
+oid sha256:84fcad9ea0c8652d4bc9523f93d5d5b1e1ff67cf81ff7384cedfbdd5f436e689
 size 41581360

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d2454c5ff0108dc47ce21c6eb9c9bdf16b4cbbc2a7f1e60c21064ab600efef47
+oid sha256:cd8cd14f9e60a4336c9af3fd57767c07145338d922cc9bb4dd9544bc3f2c1789
 size 6161