End of training

README.md

@@ -3,9 +3,15 @@ library_name: transformers
 license: apache-2.0
 base_model: pkarypis/gpt2-sft-dutch
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- BramVanroy/ultra_feedback_dutch
 model-index:
 - name: gpt2-dpo-dutch
   results: []
@@ -16,7 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # gpt2-dpo-dutch
 
-This model is a fine-tuned version of [pkarypis/gpt2-sft-dutch](https://huggingface.co/pkarypis/gpt2-sft-dutch) on an unknown dataset.
+This model is a fine-tuned version of [pkarypis/gpt2-sft-dutch](https://huggingface.co/pkarypis/gpt2-sft-dutch) on the BramVanroy/ultra_feedback_dutch dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.6936
+- Rewards/chosen: 0.0022
+- Rewards/rejected: -0.0016
+- Rewards/accuracies: 0.5172
+- Rewards/margins: 0.0038
+- Logps/rejected: -1243.3655
+- Logps/chosen: -1947.5537
+- Logits/rejected: -3.7675
+- Logits/chosen: -3.4617
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -3.4616706371307373,
+    "eval_logits/rejected": -3.767509937286377,
+    "eval_logps/chosen": -1947.5537109375,
+    "eval_logps/rejected": -1243.365478515625,
+    "eval_loss": 0.69362872838974,
+    "eval_rewards/accuracies": 0.5171641707420349,
+    "eval_rewards/chosen": 0.002152332803234458,
+    "eval_rewards/margins": 0.003774710465222597,
+    "eval_rewards/rejected": -0.0016223782440647483,
+    "eval_runtime": 207.5382,
+    "eval_samples": 5359,
+    "eval_samples_per_second": 25.822,
+    "eval_steps_per_second": 3.228,
     "total_flos": 0.0,
     "train_loss": 0.6930974917333634,
     "train_runtime": 41.8675,

config.json

@@ -25,6 +25,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.0,
+    "eval_logits/chosen": -3.4616706371307373,
+    "eval_logits/rejected": -3.767509937286377,
+    "eval_logps/chosen": -1947.5537109375,
+    "eval_logps/rejected": -1243.365478515625,
+    "eval_loss": 0.69362872838974,
+    "eval_rewards/accuracies": 0.5171641707420349,
+    "eval_rewards/chosen": 0.002152332803234458,
+    "eval_rewards/margins": 0.003774710465222597,
+    "eval_rewards/rejected": -0.0016223782440647483,
+    "eval_runtime": 207.5382,
+    "eval_samples": 5359,
+    "eval_samples_per_second": 25.822,
+    "eval_steps_per_second": 3.228
+}

runs/Sep28_15-49-51_ip-10-192-36-160/events.out.tfevents.1727538895.ip-10-192-36-160.3398194.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01af8173e9f46a663f0fe5a2d417ef359fe4cf88f6c65602178e68d922677033
+size 815