Model save

README.md

@@ -2,14 +2,10 @@
 license: mit
 library_name: peft
 tags:
-- alignment-handbook
-- generated_from_trainer
 - trl
 - dpo
 - generated_from_trainer
 base_model: microsoft/phi-2
-datasets:
-- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: phi-2-gpo-test-iter-0
   results: []
@@ -20,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # phi-2-gpo-test-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

@@ -19,9 +19,9 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
-    "q_proj",
     "k_proj",
+    "q_proj",
+    "v_proj",
     "dense"
   ],
   "task_type": "CAUSAL_LM"

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5aefc8771d1c715d0dd7ba3921091cb643f9b864456e67d5d9a7eccc5b4a3338
+oid sha256:c99237b61ddbcc12b02ca16d9318ff7a9125e77add338ae7faccac130d33f3ad
 size 41977616

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.0,
-    "train_loss": 0.0005297146562952548,
-    "train_runtime": 27.8881,
+    "train_loss": 0.00012263256940059364,
+    "train_runtime": 3.6964,
     "train_samples": 30567,
-    "train_samples_per_second": 2.223,
-    "train_steps_per_second": 0.143
+    "train_samples_per_second": 2.164,
+    "train_steps_per_second": 0.541
 }

runs/Mar20_14-05-40_Braylon/events.out.tfevents.1710904006.Braylon.7128.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea01b700538735b2e7533c0dcd9eafc8c9865f23aa1ae3f0e99c8569a40b125f
+size 5787

runs/Mar20_14-08-12_Braylon/events.out.tfevents.1710904158.Braylon.7774.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1022f0befdf79d39c4626475e7c066359d04bb4a1b2baa190cb79eb343c63668
+size 5787

runs/Mar20_14-10-43_Braylon/events.out.tfevents.1710904309.Braylon.8372.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a07d931ca10021e49380ebdd93e7da2cfcf1f3746a62fcec27ab6695f8e11711
+size 5787

runs/Mar20_14-14-21_Braylon/events.out.tfevents.1710904523.Braylon.9369.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:680cb8093a177f757eb1ba57eb5bafed5413022d6cf65ed5f94cee1552b98f27
+size 5781

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.0,
-    "train_loss": 0.0005297146562952548,
-    "train_runtime": 27.8881,
+    "train_loss": 0.00012263256940059364,
+    "train_runtime": 3.6964,
     "train_samples": 30567,
-    "train_samples_per_second": 2.223,
-    "train_steps_per_second": 0.143
+    "train_samples_per_second": 2.164,
+    "train_steps_per_second": 0.541
 }

trainer_state.json

@@ -3,19 +3,19 @@
   "best_model_checkpoint": null,
   "epoch": 2.0,
   "eval_steps": 100,
-  "global_step": 4,
+  "global_step": 2,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.5,
+      "epoch": 1.0,
       "learning_rate": 5e-06,
-      "logits/chosen": 0.8444626331329346,
-      "logits/rejected": 0.9047268629074097,
-      "logps/chosen": -269.189208984375,
-      "logps/rejected": -232.40396118164062,
-      "loss": 0.0005,
+      "logits/chosen": 0.7793408036231995,
+      "logits/rejected": 0.853427529335022,
+      "logps/chosen": -255.33981323242188,
+      "logps/rejected": -306.3643798828125,
+      "loss": 0.0001,
       "rewards/accuracies": 0.0,
       "rewards/chosen": 0.0,
       "rewards/margins": 0.0,
@@ -24,16 +24,16 @@
     },
     {
       "epoch": 2.0,
-      "step": 4,
+      "step": 2,
       "total_flos": 0.0,
-      "train_loss": 0.0005297146562952548,
-      "train_runtime": 27.8881,
-      "train_samples_per_second": 2.223,
-      "train_steps_per_second": 0.143
+      "train_loss": 0.00012263256940059364,
+      "train_runtime": 3.6964,
+      "train_samples_per_second": 2.164,
+      "train_steps_per_second": 0.541
     }
   ],
   "logging_steps": 10,
-  "max_steps": 4,
+  "max_steps": 2,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 2,
   "save_steps": 100,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f8d486432f7edca2ff053d45db771fb36af9ff4664f14de6c9bdd48da8667112
+oid sha256:9d08e7fc0828d4977b610287315f5212d8fe2e8b722629dba12249180a3d141d
 size 5816