End of training

README.md

@@ -3,6 +3,8 @@ library_name: transformers
 license: llama3.1
 base_model: mlfoundations-dev/oh_teknium_scaling_down_random_0.4
 tags:
+- llama-factory
+- full
 - trl
 - dpo
 - llama-factory
@@ -17,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # simpo-oh_teknium_scaling_down_random_0.4
 
-This model is a fine-tuned version of [mlfoundations-dev/oh_teknium_scaling_down_random_0.4](https://huggingface.co/mlfoundations-dev/oh_teknium_scaling_down_random_0.4) on an unknown dataset.
+This model is a fine-tuned version of [mlfoundations-dev/oh_teknium_scaling_down_random_0.4](https://huggingface.co/mlfoundations-dev/oh_teknium_scaling_down_random_0.4) on the mlfoundations-dev/gemma2-ultrafeedback-armorm dataset.
 It achieves the following results on the evaluation set:
 - Loss: 2.9306
 - Rewards/chosen: -28.9839

all_results.json

@@ -0,0 +1,20 @@
+{
+    "epoch": 0.9997172745264349,
+    "eval_logits/chosen": -1.0814050436019897,
+    "eval_logits/rejected": -1.0757393836975098,
+    "eval_logps/chosen": -2.8983895778656006,
+    "eval_logps/rejected": -3.4255356788635254,
+    "eval_loss": 2.930593490600586,
+    "eval_rewards/accuracies": 0.7573529481887817,
+    "eval_rewards/chosen": -28.983896255493164,
+    "eval_rewards/margins": 5.271461486816406,
+    "eval_rewards/rejected": -34.25535583496094,
+    "eval_runtime": 100.6112,
+    "eval_samples_per_second": 29.609,
+    "eval_steps_per_second": 1.859,
+    "total_flos": 134366991482880.0,
+    "train_loss": 3.5634871910060695,
+    "train_runtime": 7608.7168,
+    "train_samples_per_second": 7.438,
+    "train_steps_per_second": 0.058
+}

eval_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 0.9997172745264349,
+    "eval_logits/chosen": -1.0814050436019897,
+    "eval_logits/rejected": -1.0757393836975098,
+    "eval_logps/chosen": -2.8983895778656006,
+    "eval_logps/rejected": -3.4255356788635254,
+    "eval_loss": 2.930593490600586,
+    "eval_rewards/accuracies": 0.7573529481887817,
+    "eval_rewards/chosen": -28.983896255493164,
+    "eval_rewards/margins": 5.271461486816406,
+    "eval_rewards/rejected": -34.25535583496094,
+    "eval_runtime": 100.6112,
+    "eval_samples_per_second": 29.609,
+    "eval_steps_per_second": 1.859
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.9997172745264349,
+    "total_flos": 134366991482880.0,
+    "train_loss": 3.5634871910060695,
+    "train_runtime": 7608.7168,
+    "train_samples_per_second": 7.438,
+    "train_steps_per_second": 0.058
+}