Model save

README.md

@@ -3,16 +3,10 @@ library_name: transformers
 license: apache-2.0
 base_model: alignment-handbook/zephyr-7b-sft-full
 tags:
-- alignment-handbook
-- trl
-- dpo
-- generated_from_trainer
 - trl
 - dpo
 - alignment-handbook
 - generated_from_trainer
-datasets:
-- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-dpo-full
   results: []
@@ -23,17 +17,17 @@ 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: -0.0120
-- Rewards/chosen: -1.5839
-- Rewards/rejected: -2.6984
+- Logits/chosen: 0.5837
+- Logits/rejected: 1.9213
+- Logps/chosen: -439.8712
+- Logps/rejected: -529.8352
+- Loss: 0.7707
 - Rewards/accuracies: 0.7857
-- Rewards/margins: 1.1145
-- Logps/rejected: -530.0449
-- Logps/chosen: -440.3677
-- Logits/rejected: 1.9295
-- Logits/chosen: 0.5950
+- Rewards/chosen: -1.5789
+- Rewards/margins: 1.1174
+- Rewards/rejected: -2.6963
 
 ## Model description
 

all_results.json

@@ -15,8 +15,8 @@
     "eval_steps_per_second": 0.144,
     "total_flos": 0.0,
     "train_loss": 0.0,
-    "train_runtime": 2.6585,
+    "train_runtime": 2.3316,
     "train_samples": 61134,
-    "train_samples_per_second": 22996.052,
-    "train_steps_per_second": 359.231
+    "train_samples_per_second": 26219.362,
+    "train_steps_per_second": 409.584
 }

config.json

@@ -22,6 +22,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.2",
-  "use_cache": true,
+  "use_cache": false,
   "vocab_size": 32000
 }

runs/Apr19_07-04-57_aai-a1003/events.out.tfevents.1745014052.aai-a1003.181396.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4da894af2a05a653f506470b3d7c66bcea371f94fd8b2b4d13590d7619e8753
+size 6518

train_results.json

@@ -2,8 +2,8 @@
     "epoch": 0.9994767137624281,
     "total_flos": 0.0,
     "train_loss": 0.0,
-    "train_runtime": 2.6585,
+    "train_runtime": 2.3316,
     "train_samples": 61134,
-    "train_samples_per_second": 22996.052,
-    "train_steps_per_second": 359.231
+    "train_samples_per_second": 26219.362,
+    "train_steps_per_second": 409.584
 }

trainer_state.json

@@ -1597,9 +1597,9 @@
       "step": 955,
       "total_flos": 0.0,
       "train_loss": 0.0,
-      "train_runtime": 2.6585,
-      "train_samples_per_second": 22996.052,
-      "train_steps_per_second": 359.231
+      "train_runtime": 2.3316,
+      "train_samples_per_second": 26219.362,
+      "train_steps_per_second": 409.584
     }
   ],
   "logging_steps": 10,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3e6fe0039cff9858c99511576bf5982134bf7265a708760f709aebe2dd79c8cd
+oid sha256:91e9dad5d87e89f75b8c5a80d2e3aaf2884c3f9d7093b8bb5753e99d9d555184
 size 7480