Model save

README.md

@@ -2,13 +2,9 @@
 license: mit
 library_name: peft
 tags:
-- alignment-handbook
-- generated_from_trainer
 - trl
 - dpo
 - generated_from_trainer
-datasets:
-- HuggingFaceH4/ultrafeedback_binarized
 base_model: microsoft/phi-2
 model-index:
 - name: phi-2-gpo-iter-0
@@ -20,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # phi-2-gpo-iter-0
 
-This model is a fine-tuned version of [lole25/phi-2-sft-ultrachat-lora](https://huggingface.co/lole25/phi-2-sft-ultrachat-lora) on the HuggingFaceH4/ultrafeedback_binarized dataset.
+This model is a fine-tuned version of [microsoft/phi-2](https://huggingface.co/microsoft/phi-2) on the None dataset.
 
 ## Model description
 

adapter_config.json

@@ -20,9 +20,9 @@
   "revision": null,
   "target_modules": [
     "dense",
-    "k_proj",
+    "q_proj",
     "v_proj",
-    "q_proj"
+    "k_proj"
   ],
   "task_type": "CAUSAL_LM"
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:da885cfcacffb590b8bf76f9f633994f048e47a5c39f1b421d6c7f393e16c911
+oid sha256:f144d13fd4cedf5df77ead36d7d376e79ec3dbe6c14042eae2d43dc8d35030b5
 size 41977616

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.98,
-    "train_loss": 0.010236877437320447,
-    "train_runtime": 1512.0271,
+    "epoch": 1.99,
+    "train_loss": 0.001963847136545566,
+    "train_runtime": 279.1577,
     "train_samples": 30567,
-    "train_samples_per_second": 1.323,
-    "train_steps_per_second": 0.082
+    "train_samples_per_second": 7.164,
+    "train_steps_per_second": 0.444
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.98,
-    "train_loss": 0.010236877437320447,
-    "train_runtime": 1512.0271,
+    "epoch": 1.99,
+    "train_loss": 0.001963847136545566,
+    "train_runtime": 279.1577,
     "train_samples": 30567,
-    "train_samples_per_second": 1.323,
-    "train_steps_per_second": 0.082
+    "train_samples_per_second": 7.164,
+    "train_steps_per_second": 0.444
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.984,
+  "epoch": 1.992,
   "eval_steps": 100,
   "global_step": 124,
   "is_hyper_param_search": false,
@@ -163,41 +163,41 @@
       "step": 100
     },
     {
-      "epoch": 1.76,
+      "epoch": 1.77,
       "learning_rate": 1.937002879188285e-07,
-      "logits/chosen": 0.8223294019699097,
-      "logits/rejected": 0.8318236470222473,
-      "logps/chosen": -167.8646697998047,
-      "logps/rejected": -155.11688232421875,
-      "loss": 0.0103,
-      "rewards/accuracies": 0.32499998807907104,
-      "rewards/chosen": -0.001379788271151483,
-      "rewards/margins": -0.0001792705152183771,
-      "rewards/rejected": -0.0012005178723484278,
+      "logits/chosen": 0.9077476263046265,
+      "logits/rejected": 0.8532902598381042,
+      "logps/chosen": -142.3281707763672,
+      "logps/rejected": -133.73294067382812,
+      "loss": 0.0102,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.0010257468093186617,
+      "rewards/margins": 0.00028010181267745793,
+      "rewards/rejected": -0.0013058485928922892,
       "step": 110
     },
     {
-      "epoch": 1.92,
+      "epoch": 1.93,
       "learning_rate": 1.6003680950742728e-08,
-      "logits/chosen": 0.8208626508712769,
-      "logits/rejected": 0.8026873469352722,
-      "logps/chosen": -152.2623748779297,
-      "logps/rejected": -142.6562957763672,
+      "logits/chosen": 0.9056937098503113,
+      "logits/rejected": 0.8917877078056335,
+      "logps/chosen": -147.62950134277344,
+      "logps/rejected": -135.589599609375,
       "loss": 0.0102,
-      "rewards/accuracies": 0.41874998807907104,
-      "rewards/chosen": -0.0007717030239291489,
-      "rewards/margins": 3.4128395782317966e-05,
-      "rewards/rejected": -0.000805831397883594,
+      "rewards/accuracies": 0.45625001192092896,
+      "rewards/chosen": -0.00017889760783873498,
+      "rewards/margins": 0.0004578630905598402,
+      "rewards/rejected": -0.0006367607275024056,
       "step": 120
     },
     {
-      "epoch": 1.98,
+      "epoch": 1.99,
       "step": 124,
       "total_flos": 0.0,
-      "train_loss": 0.010236877437320447,
-      "train_runtime": 1512.0271,
-      "train_samples_per_second": 1.323,
-      "train_steps_per_second": 0.082
+      "train_loss": 0.001963847136545566,
+      "train_runtime": 279.1577,
+      "train_samples_per_second": 7.164,
+      "train_steps_per_second": 0.444
     }
   ],
   "logging_steps": 10,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ad796a191a97e0c1a72305d8c576efec65712b90fcf4d44483703b60c06b87a4
+oid sha256:65d0b7c47a66f6abb907b621b64d73257ac8eedd85947a8ad6ec086e6a803cf7
 size 5816