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-safeDPO
   results: []
@@ -15,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-sft-safeDPO
 
-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: 0.5426
-- Rewards/chosen: -2.2658
-- Rewards/rejected: -2.9947
-- Rewards/accuracies: 0.7215
-- Rewards/margins: 0.7289
-- Logps/rejected: -469.7990
-- Logps/chosen: -389.3405
-- Logits/rejected: -1.7796
-- Logits/chosen: -1.7904
+- Loss: 0.5427
+- Rewards/chosen: -2.2653
+- Rewards/rejected: -2.9944
+- Rewards/accuracies: 0.7197
+- Rewards/margins: 0.7291
+- Logps/rejected: -469.7640
+- Logps/chosen: -389.2892
+- Logits/rejected: -1.7787
+- Logits/chosen: -1.7892
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -1.7892440557479858,
+    "eval_logits/rejected": -1.7786749601364136,
+    "eval_logps/chosen": -389.2891845703125,
+    "eval_logps/rejected": -469.7640075683594,
+    "eval_loss": 0.5426768064498901,
+    "eval_rewards/accuracies": 0.719748854637146,
+    "eval_rewards/chosen": -2.265324115753174,
+    "eval_rewards/margins": 0.7290500402450562,
+    "eval_rewards/rejected": -2.9943742752075195,
+    "eval_runtime": 524.2847,
+    "eval_samples": 7000,
+    "eval_samples_per_second": 13.352,
+    "eval_steps_per_second": 0.418,
     "train_loss": 0.55459001399405,
     "train_runtime": 38266.551,
     "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": 1.0,
+    "eval_logits/chosen": -1.7892440557479858,
+    "eval_logits/rejected": -1.7786749601364136,
+    "eval_logps/chosen": -389.2891845703125,
+    "eval_logps/rejected": -469.7640075683594,
+    "eval_loss": 0.5426768064498901,
+    "eval_rewards/accuracies": 0.719748854637146,
+    "eval_rewards/chosen": -2.265324115753174,
+    "eval_rewards/margins": 0.7290500402450562,
+    "eval_rewards/rejected": -2.9943742752075195,
+    "eval_runtime": 524.2847,
+    "eval_samples": 7000,
+    "eval_samples_per_second": 13.352,
+    "eval_steps_per_second": 0.418
+}

runs/May05_19-13-15_gilbreth-j001.rcac.purdue.edu/events.out.tfevents.1714991721.gilbreth-j001.rcac.purdue.edu.6343.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e81638da659ed3ab68146aa7095c7dadb3f396d215f4799206fb6d41fdf89fdf
+size 828