End of training

README.md

@@ -2,9 +2,15 @@
 license: other
 base_model: HuggingFaceH4/zephyr-7b-gemma-sft-v0.1
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- argilla/dpo-mix-7k
 model-index:
 - name: zephyr-7b-gemma-dpo
   results: []
@@ -15,17 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-gemma-dpo
 
-This model is a fine-tuned version of [HuggingFaceH4/zephyr-7b-gemma-sft-v0.1](https://huggingface.co/HuggingFaceH4/zephyr-7b-gemma-sft-v0.1) on an unknown dataset.
+This model is a fine-tuned version of [HuggingFaceH4/zephyr-7b-gemma-sft-v0.1](https://huggingface.co/HuggingFaceH4/zephyr-7b-gemma-sft-v0.1) on the argilla/dpo-mix-7k dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4692
-- Rewards/chosen: -2.9810
-- Rewards/rejected: -4.6451
-- Rewards/accuracies: 0.7604
-- Rewards/margins: 1.6641
-- Logps/rejected: -453.0530
-- Logps/chosen: -422.7995
-- Logits/rejected: 91.0574
-- Logits/chosen: 96.7661
+- Loss: 0.4683
+- Rewards/chosen: -3.0221
+- Rewards/rejected: -4.6813
+- Rewards/accuracies: 0.7708
+- Rewards/margins: 1.6592
+- Logps/rejected: -453.7782
+- Logps/chosen: -423.6228
+- Logits/rejected: 90.9822
+- Logits/chosen: 96.7158
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.971563981042654,
+    "eval_logits/chosen": 96.71578216552734,
+    "eval_logits/rejected": 90.98221588134766,
+    "eval_logps/chosen": -423.6227722167969,
+    "eval_logps/rejected": -453.7782287597656,
+    "eval_loss": 0.468290776014328,
+    "eval_rewards/accuracies": 0.7708333134651184,
+    "eval_rewards/chosen": -3.0221338272094727,
+    "eval_rewards/margins": 1.6591955423355103,
+    "eval_rewards/rejected": -4.681329727172852,
+    "eval_runtime": 58.6185,
+    "eval_samples": 750,
+    "eval_samples_per_second": 12.795,
+    "eval_steps_per_second": 0.409,
     "total_flos": 0.0,
     "train_loss": 0.3921648321243433,
     "train_runtime": 1190.3032,

config.json

@@ -24,6 +24,6 @@
   "rope_theta": 10000.0,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.40.1",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 256000
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.971563981042654,
-    "eval_logits/chosen": 96.0078125,
-    "eval_logits/rejected": 90.1099853515625,
-    "eval_logps/chosen": -434.6161193847656,
-    "eval_logps/rejected": -466.3072814941406,
-    "eval_loss": 0.4578173756599426,
-    "eval_rewards/accuracies": 0.7604166865348816,
-    "eval_rewards/chosen": -3.5473945140838623,
-    "eval_rewards/margins": 1.6644223928451538,
-    "eval_rewards/rejected": -5.211816787719727,
-    "eval_runtime": 124.9704,
+    "eval_logits/chosen": 96.71578216552734,
+    "eval_logits/rejected": 90.98221588134766,
+    "eval_logps/chosen": -423.6227722167969,
+    "eval_logps/rejected": -453.7782287597656,
+    "eval_loss": 0.468290776014328,
+    "eval_rewards/accuracies": 0.7708333134651184,
+    "eval_rewards/chosen": -3.0221338272094727,
+    "eval_rewards/margins": 1.6591955423355103,
+    "eval_rewards/rejected": -4.681329727172852,
+    "eval_runtime": 58.6185,
     "eval_samples": 750,
-    "eval_samples_per_second": 6.001,
-    "eval_steps_per_second": 0.192
+    "eval_samples_per_second": 12.795,
+    "eval_steps_per_second": 0.409
 }

runs/Apr27_13-04-12_gcp002/events.out.tfevents.1714226630.gcp002.2844214.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:66c87c7f080f1633e045b5954eafce01ef8cb719091ee8e0632700867593562a
+size 815