End of training

README.md

@@ -2,9 +2,16 @@
 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
+- AmberYifan/safetyQA_DPO
 model-index:
 - name: zephyr-7b-sft-safeDPO3
   results: []
@@ -15,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-sft-safeDPO3
 
-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.
+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 and the AmberYifan/safetyQA_DPO datasets.
 It achieves the following results on the evaluation set:
 - Loss: 0.6446
-- Rewards/chosen: -8.0277
-- Rewards/rejected: -9.5354
-- Rewards/accuracies: 0.7163
-- Rewards/margins: 1.5078
-- Logps/rejected: -1123.8704
-- Logps/chosen: -965.5221
-- Logits/rejected: 3.5627
-- Logits/chosen: 4.0395
+- Rewards/chosen: -8.0278
+- Rewards/rejected: -9.5352
+- Rewards/accuracies: 0.7152
+- Rewards/margins: 1.5074
+- Logps/rejected: -1123.8456
+- Logps/chosen: -965.5345
+- Logits/rejected: 3.5622
+- Logits/chosen: 4.0391
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 3.0,
+    "eval_logits/chosen": 4.039111614227295,
+    "eval_logits/rejected": 3.562220811843872,
+    "eval_logps/chosen": -965.5345458984375,
+    "eval_logps/rejected": -1123.8455810546875,
+    "eval_loss": 0.6446220278739929,
+    "eval_rewards/accuracies": 0.7151826620101929,
+    "eval_rewards/chosen": -8.027778625488281,
+    "eval_rewards/margins": 1.5074113607406616,
+    "eval_rewards/rejected": -9.535189628601074,
+    "eval_runtime": 534.3288,
+    "eval_samples": 7000,
+    "eval_samples_per_second": 13.101,
+    "eval_steps_per_second": 0.41,
     "train_loss": 0.3444365044389086,
     "train_runtime": 113135.1128,
     "train_samples": 111134,

config.json

@@ -21,6 +21,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.39.0.dev0",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 3.0,
+    "eval_logits/chosen": 4.039111614227295,
+    "eval_logits/rejected": 3.562220811843872,
+    "eval_logps/chosen": -965.5345458984375,
+    "eval_logps/rejected": -1123.8455810546875,
+    "eval_loss": 0.6446220278739929,
+    "eval_rewards/accuracies": 0.7151826620101929,
+    "eval_rewards/chosen": -8.027778625488281,
+    "eval_rewards/margins": 1.5074113607406616,
+    "eval_rewards/rejected": -9.535189628601074,
+    "eval_runtime": 534.3288,
+    "eval_samples": 7000,
+    "eval_samples_per_second": 13.101,
+    "eval_steps_per_second": 0.41
+}

runs/May07_21-56-54_gilbreth-j001.rcac.purdue.edu/events.out.tfevents.1715247503.gilbreth-j001.rcac.purdue.edu.28663.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6dc71cf51e1ed2fbf491fe8d96b99676ca63e5d98cf165c169641db9a5451032
+size 828