Model save

README.md

@@ -2,13 +2,9 @@
 license: apache-2.0
 base_model: alignment-handbook/zephyr-7b-sft-full
 tags:
-- alignment-handbook
-- generated_from_trainer
 - trl
 - dpo
 - generated_from_trainer
-datasets:
-- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-dpo-full
   results: []
@@ -19,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-dpo-full
 
-This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the HuggingFaceH4/ultrafeedback_binarized dataset.
+This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the None dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.5028
 - Rewards/chosen: -0.9469

all_results.json

@@ -9,13 +9,13 @@
     "eval_rewards/chosen": -0.9468507170677185,
     "eval_rewards/margins": 0.946345865726471,
     "eval_rewards/rejected": -1.8931965827941895,
-    "eval_runtime": 88.2311,
+    "eval_runtime": 86.4354,
     "eval_samples": 2000,
-    "eval_samples_per_second": 22.668,
-    "eval_steps_per_second": 0.363,
+    "eval_samples_per_second": 23.139,
+    "eval_steps_per_second": 0.37,
     "train_loss": 0.5366686437918052,
-    "train_runtime": 5259.7251,
+    "train_runtime": 5196.7487,
     "train_samples": 61135,
-    "train_samples_per_second": 11.623,
-    "train_steps_per_second": 0.091
+    "train_samples_per_second": 11.764,
+    "train_steps_per_second": 0.092
 }

eval_results.json

@@ -9,8 +9,8 @@
     "eval_rewards/chosen": -0.9468507170677185,
     "eval_rewards/margins": 0.946345865726471,
     "eval_rewards/rejected": -1.8931965827941895,
-    "eval_runtime": 88.2311,
+    "eval_runtime": 86.4354,
     "eval_samples": 2000,
-    "eval_samples_per_second": 22.668,
-    "eval_steps_per_second": 0.363
+    "eval_samples_per_second": 23.139,
+    "eval_steps_per_second": 0.37
 }

runs/Jan09_22-10-39_ip-26-0-161-123/events.out.tfevents.1704838276.ip-26-0-161-123.15622.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0d7dd4e1763b4cf06eb8548c3dd60a78c4deb4f3abc6c8e5106d560fe48e1903
-size 33330
+oid sha256:aa992e831700f770bddff3a401b7041798cbf0c8d045f8b2c3a00e6466847ef0
+size 38122

runs/Jan09_22-10-39_ip-26-0-161-123/events.out.tfevents.1704843559.ip-26-0-161-123.15622.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52db6ce34acd137452951519c58c3a279488a7192f40fc17a7a1f0feee27a8ec
+size 828

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
     "train_loss": 0.5366686437918052,
-    "train_runtime": 5259.7251,
+    "train_runtime": 5196.7487,
     "train_samples": 61135,
-    "train_samples_per_second": 11.623,
-    "train_steps_per_second": 0.091
+    "train_samples_per_second": 11.764,
+    "train_steps_per_second": 0.092
 }

trainer_state.json

@@ -173,9 +173,9 @@
       "eval_rewards/chosen": -0.49532508850097656,
       "eval_rewards/margins": 0.6263692378997803,
       "eval_rewards/rejected": -1.1216944456100464,
-      "eval_runtime": 87.9008,
-      "eval_samples_per_second": 22.753,
-      "eval_steps_per_second": 0.364,
+      "eval_runtime": 85.4149,
+      "eval_samples_per_second": 23.415,
+      "eval_steps_per_second": 0.375,
       "step": 100
     },
     {
@@ -329,9 +329,9 @@
       "eval_rewards/chosen": -0.899510383605957,
       "eval_rewards/margins": 0.8722902536392212,
       "eval_rewards/rejected": -1.7718006372451782,
-      "eval_runtime": 87.889,
-      "eval_samples_per_second": 22.756,
-      "eval_steps_per_second": 0.364,
+      "eval_runtime": 85.7368,
+      "eval_samples_per_second": 23.327,
+      "eval_steps_per_second": 0.373,
       "step": 200
     },
     {
@@ -485,9 +485,9 @@
       "eval_rewards/chosen": -0.7945692539215088,
       "eval_rewards/margins": 0.8339425921440125,
       "eval_rewards/rejected": -1.628511905670166,
-      "eval_runtime": 88.1333,
-      "eval_samples_per_second": 22.693,
-      "eval_steps_per_second": 0.363,
+      "eval_runtime": 86.1066,
+      "eval_samples_per_second": 23.227,
+      "eval_steps_per_second": 0.372,
       "step": 300
     },
     {
@@ -641,9 +641,9 @@
       "eval_rewards/chosen": -0.9776647090911865,
       "eval_rewards/margins": 0.9803228974342346,
       "eval_rewards/rejected": -1.9579875469207764,
-      "eval_runtime": 88.4314,
-      "eval_samples_per_second": 22.616,
-      "eval_steps_per_second": 0.362,
+      "eval_runtime": 86.2141,
+      "eval_samples_per_second": 23.198,
+      "eval_steps_per_second": 0.371,
       "step": 400
     },
     {
@@ -749,9 +749,9 @@
       "step": 478,
       "total_flos": 0.0,
       "train_loss": 0.5366686437918052,
-      "train_runtime": 5259.7251,
-      "train_samples_per_second": 11.623,
-      "train_steps_per_second": 0.091
+      "train_runtime": 5196.7487,
+      "train_samples_per_second": 11.764,
+      "train_steps_per_second": 0.092
     }
   ],
   "logging_steps": 10,