Model save

README.md

@@ -34,12 +34,12 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 2e-05
-- train_batch_size: 11
+- train_batch_size: 1
 - eval_batch_size: 16
 - seed: 42
 - distributed_type: multi-GPU
 - gradient_accumulation_steps: 4
-- total_train_batch_size: 44
+- total_train_batch_size: 4
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - num_epochs: 1

adapter_config.json

@@ -17,9 +17,9 @@
   "revision": null,
   "target_modules": [
     "k_proj",
+    "v_proj",
     "o_proj",
-    "q_proj",
-    "v_proj"
+    "q_proj"
   ],
   "task_type": "CAUSAL_LM"
 }

runs/Feb04_00-47-37_r2d2-devbox-fqvnh/events.out.tfevents.1738630073.r2d2-devbox-fqvnh.22604.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2038b40239f1c9314cbfb2cb17cc634742a37ca05bd635a3370c3db5882d2286
+size 4702

runs/Feb04_01-20-10_r2d2-devbox-fqvnh/events.out.tfevents.1738632026.r2d2-devbox-fqvnh.23252.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbf9bbcd4df87a9eaa4f03287ea293f6e2c4b44c8b8565d8f317734561d11300
+size 4702

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8cb086fa0f352a7cdaf6af7173b7ed288f2f20adf12b97a9cf104524dceb24a8
-size 5816
+oid sha256:7f61e0d477403b0348768c2a5c82297770d6a18e17c7c1da31e26b329c52011d
+size 5880