End of training

README.md

@@ -2,10 +2,16 @@
 library_name: transformers
 base_model: allenai/tulu-2-7b
 tags:
+- alignment-handbook
+- trl
+- dpo
+- generated_from_trainer
 - trl
 - dpo
 - alignment-handbook
 - generated_from_trainer
+datasets:
+- data/uf_rlced_conifer
 model-index:
 - name: uf-rlced-conifer_tulu-2-7b-dpo-full
   results: []
@@ -16,7 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # uf-rlced-conifer_tulu-2-7b-dpo-full
 
-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/uf_rlced_conifer dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.3316
+- Rewards/chosen: -2.6774
+- Rewards/rejected: -5.0456
+- Rewards/accuracies: 0.8383
+- Rewards/margins: 2.3682
+- Logps/rejected: -989.8275
+- Logps/chosen: -729.1251
+- Logits/rejected: -0.3176
+- Logits/chosen: -0.4437
 
 ## Model description
 

all_results.json

@@ -1,18 +1,18 @@
 {
     "epoch": 0.9993060374739764,
-    "eval_logits/chosen": -1.3140687942504883,
-    "eval_logits/rejected": -1.1924079656600952,
-    "eval_logps/chosen": -316.60687255859375,
-    "eval_logps/rejected": -434.5618896484375,
-    "eval_loss": 0.6931471228599548,
-    "eval_rewards/accuracies": 0.0,
-    "eval_rewards/chosen": 0.0,
-    "eval_rewards/margins": 0.0,
-    "eval_rewards/rejected": 0.0,
-    "eval_runtime": 2.7893,
-    "eval_samples": 50,
-    "eval_samples_per_second": 17.926,
-    "eval_steps_per_second": 0.359,
+    "eval_logits/chosen": -0.44369885325431824,
+    "eval_logits/rejected": -0.31763577461242676,
+    "eval_logps/chosen": -729.1251220703125,
+    "eval_logps/rejected": -989.8275146484375,
+    "eval_loss": 0.3316161632537842,
+    "eval_rewards/accuracies": 0.8383458852767944,
+    "eval_rewards/chosen": -2.6774144172668457,
+    "eval_rewards/margins": 2.368194341659546,
+    "eval_rewards/rejected": -5.0456085205078125,
+    "eval_runtime": 378.3021,
+    "eval_samples": 8491,
+    "eval_samples_per_second": 22.445,
+    "eval_steps_per_second": 0.352,
     "total_flos": 0.0,
     "train_loss": 0.4493948830498589,
     "train_runtime": 20175.981,

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

@@ -1,16 +1,16 @@
 {
-    "epoch": 1.0,
-    "eval_logits/chosen": -1.3140687942504883,
-    "eval_logits/rejected": -1.1924079656600952,
-    "eval_logps/chosen": -316.60687255859375,
-    "eval_logps/rejected": -434.5618896484375,
-    "eval_loss": 0.6931471228599548,
-    "eval_rewards/accuracies": 0.0,
-    "eval_rewards/chosen": 0.0,
-    "eval_rewards/margins": 0.0,
-    "eval_rewards/rejected": 0.0,
-    "eval_runtime": 2.7893,
-    "eval_samples": 50,
-    "eval_samples_per_second": 17.926,
-    "eval_steps_per_second": 0.359
+    "epoch": 0.9993060374739764,
+    "eval_logits/chosen": -0.44369885325431824,
+    "eval_logits/rejected": -0.31763577461242676,
+    "eval_logps/chosen": -729.1251220703125,
+    "eval_logps/rejected": -989.8275146484375,
+    "eval_loss": 0.3316161632537842,
+    "eval_rewards/accuracies": 0.8383458852767944,
+    "eval_rewards/chosen": -2.6774144172668457,
+    "eval_rewards/margins": 2.368194341659546,
+    "eval_rewards/rejected": -5.0456085205078125,
+    "eval_runtime": 378.3021,
+    "eval_samples": 8491,
+    "eval_samples_per_second": 22.445,
+    "eval_steps_per_second": 0.352
 }