Model save

README.md

@@ -34,12 +34,14 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 2e-05
-- train_batch_size: 8
-- eval_batch_size: 16
+- train_batch_size: 4
+- eval_batch_size: 8
 - seed: 42
 - distributed_type: multi-GPU
-- gradient_accumulation_steps: 4
-- total_train_batch_size: 32
+- num_devices: 2
+- gradient_accumulation_steps: 128
+- total_train_batch_size: 1024
+- total_eval_batch_size: 16
 - 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": [
+    "q_proj",
     "o_proj",
-    "v_proj",
     "k_proj",
-    "q_proj"
+    "v_proj"
   ],
   "task_type": "CAUSAL_LM"
 }

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b8e818c4f29f92eacb112a39c779bcbab81dd0a966e50759226c048173df3ec0
-size 4712
+oid sha256:45d2ff3b30e4f1a0abd211c5b1be46bf5b28893e27d8cb8bbbba4503efd3127a
+size 5174

runs/Feb01_23-39-41_r2d2-devbox-l74gc/events.out.tfevents.1738453318.r2d2-devbox-l74gc.883.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:741bee1714c52550b4b540e0a88c1b934b4e2fe46945d6ccfd760452891b028b
+size 4699

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:256c32577978daf1314879c1a40d5d309652be85248145f60cc9352f169e81a8
-size 5880
+oid sha256:83a452bfaf833a0e1a022c8aa6d5cbbfcd12ed965b49a5d3e2bdb03a64568a5d
+size 5816