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: 4
-- eval_batch_size: 8
+- train_batch_size: 8
+- eval_batch_size: 16
 - seed: 42
 - distributed_type: multi-GPU
-- gradient_accumulation_steps: 128
-- total_train_batch_size: 512
+- gradient_accumulation_steps: 4
+- total_train_batch_size: 32
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - num_epochs: 1

adapter_config.json

@@ -16,10 +16,10 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "k_proj",
-    "q_proj",
+    "o_proj",
     "v_proj",
-    "o_proj"
+    "k_proj",
+    "q_proj"
   ],
   "task_type": "CAUSAL_LM"
 }

runs/Feb01_16-42-58_r2d2-tom-000-train-lora-t4whq/events.out.tfevents.1738428322.r2d2-tom-000-train-lora-t4whq.96.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fbf1a6f080620a0a6868857852f6d8a3de1595c7e013e051718486078fb9262b
-size 4711
+oid sha256:6b9aba495b02ccda8d60922bc521640398ecb61207a84f1c19cdfc18112183e7
+size 5019

runs/Feb01_23-21-59_r2d2-tom-000b-train-lora-pl9d7/events.out.tfevents.1738452258.r2d2-tom-000b-train-lora-pl9d7.96.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b8e818c4f29f92eacb112a39c779bcbab81dd0a966e50759226c048173df3ec0
+size 4712

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dbd2f1ef498e93094a0ce00de27829921f4592d024d9a6f82dffa6af10ec9bd1
+oid sha256:256c32577978daf1314879c1a40d5d309652be85248145f60cc9352f169e81a8
 size 5880