dpo_4rbrxhkg

README.md

@@ -17,15 +17,15 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model was trained from scratch on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6931
-- Rewards/chosen: 0.0
-- Rewards/rejected: 0.0
-- Rewards/accuracies: 0.0
-- Rewards/margins: 0.0
-- Logps/rejected: -32.6347
-- Logps/chosen: -35.1065
-- Logits/rejected: -1.1843
-- Logits/chosen: -1.2320
+- Loss: 0.7021
+- Rewards/chosen: 0.5480
+- Rewards/rejected: 0.4854
+- Rewards/accuracies: 0.1242
+- Rewards/margins: 0.0626
+- Logps/rejected: -26.1813
+- Logps/chosen: -28.6110
+- Logits/rejected: -1.0475
+- Logits/chosen: -1.0554
 
 ## Model description
 
@@ -44,14 +44,14 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 5e-05
+- learning_rate: 0.0001
 - train_batch_size: 1
 - eval_batch_size: 1
 - seed: 0
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_steps: 100
-- training_steps: 2
+- training_steps: 2500
 
 ### Training results
 

adapter_config.json

@@ -16,17 +16,17 @@
   "megatron_core": "megatron.core",
   "modules_to_save": null,
   "peft_type": "LORA",
-  "r": 16,
+  "r": 64,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "out_proj",
-    "q_proj",
-    "fc_in",
     "k_proj",
-    "v_proj",
     "wte",
-    "fc_out"
+    "q_proj",
+    "fc_out",
+    "out_proj",
+    "v_proj",
+    "fc_in"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5ab0fda06fcdd9a8e16d9a236718f37697611b438e210d287c0a0fcbfea98ee7
-size 18900240
+oid sha256:37c713439897aefab1d63b1f1938ec6a3ff714264132976bebe1d9780fd5dfe3
+size 75523504

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d13602293bb6f8bb01001bbab0f60f58b3113aee181e0ee084d00eb02d9a39aa
+oid sha256:3b116eedb32c45e3728095b6935fbdfd4c2aa57a3764d638aa33e5a1d3237a04
 size 5560