End of training

README.md

@@ -2,9 +2,15 @@
 license: apache-2.0
 base_model: alignment-handbook/zephyr-7b-sft-full
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-dpo-full
   results: []
@@ -15,17 +21,17 @@ 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 an unknown 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 HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.1475
-- Rewards/chosen: -10.8669
-- Rewards/rejected: -16.0106
+- Loss: 1.1487
+- Rewards/chosen: -10.8742
+- Rewards/rejected: -16.0045
 - Rewards/accuracies: 0.7285
-- Rewards/margins: 5.1437
-- Logps/rejected: -424.5231
-- Logps/chosen: -383.5809
-- Logits/rejected: -0.5915
-- Logits/chosen: -1.0022
+- Rewards/margins: 5.1303
+- Logps/rejected: -424.4627
+- Logps/chosen: -383.6538
+- Logits/rejected: -0.5906
+- Logits/chosen: -1.0023
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 3.0,
+    "eval_logits/chosen": -1.0022624731063843,
+    "eval_logits/rejected": -0.5906034111976624,
+    "eval_logps/chosen": -383.65380859375,
+    "eval_logps/rejected": -424.4626770019531,
+    "eval_loss": 1.148720145225525,
+    "eval_rewards/accuracies": 0.728515625,
+    "eval_rewards/chosen": -10.874195098876953,
+    "eval_rewards/margins": 5.130314826965332,
+    "eval_rewards/rejected": -16.0045108795166,
+    "eval_runtime": 215.9812,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 9.26,
+    "eval_steps_per_second": 0.148,
     "total_flos": 0.0,
     "train_loss": 0.21746732159101023,
     "train_runtime": 62085.3048,

config.json

@@ -21,6 +21,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.40.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
-    "epoch": 0.9994767137624281,
-    "eval_logits/chosen": -2.5900135040283203,
-    "eval_logits/rejected": -2.5359914302825928,
-    "eval_logps/chosen": -282.1169128417969,
-    "eval_logps/rejected": -272.686279296875,
-    "eval_loss": 0.5283266305923462,
-    "eval_rewards/accuracies": 0.773809552192688,
-    "eval_rewards/chosen": -0.016292406246066093,
-    "eval_rewards/margins": 1.2304461002349854,
-    "eval_rewards/rejected": -1.2467385530471802,
-    "eval_runtime": 205.7802,
+    "epoch": 3.0,
+    "eval_logits/chosen": -1.0022624731063843,
+    "eval_logits/rejected": -0.5906034111976624,
+    "eval_logps/chosen": -383.65380859375,
+    "eval_logps/rejected": -424.4626770019531,
+    "eval_loss": 1.148720145225525,
+    "eval_rewards/accuracies": 0.728515625,
+    "eval_rewards/chosen": -10.874195098876953,
+    "eval_rewards/margins": 5.130314826965332,
+    "eval_rewards/rejected": -16.0045108795166,
+    "eval_runtime": 215.9812,
     "eval_samples": 2000,
-    "eval_samples_per_second": 9.719,
-    "eval_steps_per_second": 0.306
+    "eval_samples_per_second": 9.26,
+    "eval_steps_per_second": 0.148
 }

runs/May26_06-26-16_ubuntu/events.out.tfevents.1716767808.ubuntu.2970062.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f87e59bd24220b0926516f26b803c428192e39ec7d60b7d7128e70eebc710d8
+size 828