Model save

README.md

@@ -33,12 +33,12 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 8e-05
-- train_batch_size: 8
+- learning_rate: 0.08
+- train_batch_size: 4
 - eval_batch_size: 8
 - seed: 42
 - gradient_accumulation_steps: 8
-- total_train_batch_size: 64
+- total_train_batch_size: 32
 - optimizer: Use adafactor and the args are:
 No additional optimizer arguments
 - lr_scheduler_type: linear

adapter_config.json

@@ -23,13 +23,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
     "v_proj",
     "up_proj",
-    "gate_proj",
-    "o_proj",
     "down_proj",
-    "k_proj"
+    "gate_proj",
+    "q_proj",
+    "k_proj",
+    "o_proj"
   ],
   "task_type": null,
   "use_dora": true,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4db63ec5d15ab752f9693a081315a1399de882ef9f5a77dc12d79f7ffde73967
+oid sha256:ecf638102a225aab8c00ec0326c4b5775935f387694086df86e7ab1e4d478cf4
 size 51802928

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c8a435bc54cc2e796b340027b68e1927ef59a84897cb59ff040cb73387a501df
+oid sha256:d6e8b373bee51b72a7104fadf549da1b2ae55b4a63838d471917d291e49821b8
 size 5240