End of training

README.md

@@ -3,6 +3,8 @@ library_name: transformers
 license: llama3.1
 base_model: mlfoundations-dev/stackexchange_christianity
 tags:
+- llama-factory
+- full
 - trl
 - dpo
 - llama-factory
@@ -17,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # simpo-stackexchange_christianity
 
-This model is a fine-tuned version of [mlfoundations-dev/stackexchange_christianity](https://huggingface.co/mlfoundations-dev/stackexchange_christianity) on an unknown dataset.
+This model is a fine-tuned version of [mlfoundations-dev/stackexchange_christianity](https://huggingface.co/mlfoundations-dev/stackexchange_christianity) on the mlfoundations-dev/gemma2-ultrafeedback-armorm dataset.
 It achieves the following results on the evaluation set:
 - Loss: 2.5820
 - Rewards/chosen: -43.5673

all_results.json

@@ -0,0 +1,20 @@
+{
+    "epoch": 0.9997172745264349,
+    "eval_logits/chosen": -0.925058126449585,
+    "eval_logits/rejected": -0.9405221939086914,
+    "eval_logps/chosen": -4.356727600097656,
+    "eval_logps/rejected": -5.03577995300293,
+    "eval_loss": 2.5820231437683105,
+    "eval_rewards/accuracies": 0.7914438247680664,
+    "eval_rewards/chosen": -43.56727981567383,
+    "eval_rewards/margins": 6.790517807006836,
+    "eval_rewards/rejected": -50.35779571533203,
+    "eval_runtime": 50.3639,
+    "eval_samples_per_second": 59.15,
+    "eval_steps_per_second": 3.713,
+    "total_flos": 134366991482880.0,
+    "train_loss": 3.371998559026157,
+    "train_runtime": 3776.6556,
+    "train_samples_per_second": 14.984,
+    "train_steps_per_second": 0.117
+}

eval_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 0.9997172745264349,
+    "eval_logits/chosen": -0.925058126449585,
+    "eval_logits/rejected": -0.9405221939086914,
+    "eval_logps/chosen": -4.356727600097656,
+    "eval_logps/rejected": -5.03577995300293,
+    "eval_loss": 2.5820231437683105,
+    "eval_rewards/accuracies": 0.7914438247680664,
+    "eval_rewards/chosen": -43.56727981567383,
+    "eval_rewards/margins": 6.790517807006836,
+    "eval_rewards/rejected": -50.35779571533203,
+    "eval_runtime": 50.3639,
+    "eval_samples_per_second": 59.15,
+    "eval_steps_per_second": 3.713
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.9997172745264349,
+    "total_flos": 134366991482880.0,
+    "train_loss": 3.371998559026157,
+    "train_runtime": 3776.6556,
+    "train_samples_per_second": 14.984,
+    "train_steps_per_second": 0.117
+}