End of training

README.md

@@ -2,9 +2,15 @@
 library_name: transformers
 base_model: allenai/tulu-2-7b
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- data/tulu-2-7b-uf-rlced-conifer-ref
 model-index:
 - name: uf-rlced-conifer_tulu-2-7b-group-dpo-no-clip
   results: []
@@ -15,7 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # uf-rlced-conifer_tulu-2-7b-group-dpo-no-clip
 
-This model is a fine-tuned version of [allenai/tulu-2-7b](https://huggingface.co/allenai/tulu-2-7b) on an unknown dataset.
+This model is a fine-tuned version of [allenai/tulu-2-7b](https://huggingface.co/allenai/tulu-2-7b) on the data/tulu-2-7b-uf-rlced-conifer-ref dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.6073
+- Rewards/chosen: -0.2204
+- Rewards/rejected: -0.4189
+- Rewards/accuracies: 0.7876
+- Rewards/margins: 0.1984
+- Logps/rejected: -527.1544
+- Logps/chosen: -483.4286
+- Logits/rejected: -0.9235
+- Logits/chosen: -0.9285
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.9993060374739764,
+    "eval_logits/chosen": -0.9284578561782837,
+    "eval_logits/rejected": -0.923488974571228,
+    "eval_logps/chosen": -483.4286193847656,
+    "eval_logps/rejected": -527.1544189453125,
+    "eval_loss": 0.6072572469711304,
+    "eval_rewards/accuracies": 0.7875939607620239,
+    "eval_rewards/chosen": -0.22044964134693146,
+    "eval_rewards/margins": 0.19842886924743652,
+    "eval_rewards/rejected": -0.4188785254955292,
+    "eval_runtime": 379.0532,
+    "eval_samples": 8491,
+    "eval_samples_per_second": 22.401,
+    "eval_steps_per_second": 0.351,
     "total_flos": 0.0,
     "train_loss": 0.008160642593995564,
     "train_runtime": 20291.8014,

config.json

@@ -24,6 +24,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.1",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 0.9993060374739764,
+    "eval_logits/chosen": -0.9284578561782837,
+    "eval_logits/rejected": -0.923488974571228,
+    "eval_logps/chosen": -483.4286193847656,
+    "eval_logps/rejected": -527.1544189453125,
+    "eval_loss": 0.6072572469711304,
+    "eval_rewards/accuracies": 0.7875939607620239,
+    "eval_rewards/chosen": -0.22044964134693146,
+    "eval_rewards/margins": 0.19842886924743652,
+    "eval_rewards/rejected": -0.4188785254955292,
+    "eval_runtime": 379.0532,
+    "eval_samples": 8491,
+    "eval_samples_per_second": 22.401,
+    "eval_steps_per_second": 0.351
+}