End of training

README.md

@@ -2,9 +2,15 @@
 license: other
 base_model: lewtun/gemma-7b-sft-full-deita-10k-v0
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- argilla/dpo-mix-7k
 model-index:
 - name: gemma-7b-dpo-full-mix1-beta-0.1
   results: []
@@ -15,7 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # gemma-7b-dpo-full-mix1-beta-0.1
 
-This model is a fine-tuned version of [lewtun/gemma-7b-sft-full-deita-10k-v0](https://huggingface.co/lewtun/gemma-7b-sft-full-deita-10k-v0) on the None dataset.
+This model is a fine-tuned version of [lewtun/gemma-7b-sft-full-deita-10k-v0](https://huggingface.co/lewtun/gemma-7b-sft-full-deita-10k-v0) on the argilla/dpo-mix-7k dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.6160
+- Rewards/chosen: -1.9864
+- Rewards/rejected: -3.4684
+- Rewards/accuracies: 0.7083
+- Rewards/margins: 1.4821
+- Logps/rejected: -486.2339
+- Logps/chosen: -473.2680
+- Logits/rejected: 99.1940
+- Logits/chosen: 105.4826
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.99,
+    "eval_logits/chosen": 105.48260498046875,
+    "eval_logits/rejected": 99.1939697265625,
+    "eval_logps/chosen": -473.2679748535156,
+    "eval_logps/rejected": -486.23388671875,
+    "eval_loss": 0.6159864068031311,
+    "eval_rewards/accuracies": 0.7083333134651184,
+    "eval_rewards/chosen": -1.986350178718567,
+    "eval_rewards/margins": 1.4820828437805176,
+    "eval_rewards/rejected": -3.468433141708374,
+    "eval_runtime": 50.9713,
+    "eval_samples": 750,
+    "eval_samples_per_second": 14.714,
+    "eval_steps_per_second": 0.471,
     "train_loss": 0.6733019076860868,
     "train_runtime": 575.7905,
     "train_samples": 6750,

config.json

@@ -23,6 +23,6 @@
   "rope_theta": 10000.0,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.39.0.dev0",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 256000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 0.99,
+    "eval_logits/chosen": 105.48260498046875,
+    "eval_logits/rejected": 99.1939697265625,
+    "eval_logps/chosen": -473.2679748535156,
+    "eval_logps/rejected": -486.23388671875,
+    "eval_loss": 0.6159864068031311,
+    "eval_rewards/accuracies": 0.7083333134651184,
+    "eval_rewards/chosen": -1.986350178718567,
+    "eval_rewards/margins": 1.4820828437805176,
+    "eval_rewards/rejected": -3.468433141708374,
+    "eval_runtime": 50.9713,
+    "eval_samples": 750,
+    "eval_samples_per_second": 14.714,
+    "eval_steps_per_second": 0.471
+}

runs/Feb29_21-31-57_ip-26-0-171-230/events.out.tfevents.1709243154.ip-26-0-171-230.519606.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9510f0d2fec96ecb232a5e57c26cd29f853db81491f8d5e96cdac60e5958e9b
+size 815

runs/Feb29_21-37-35_ip-26-0-164-207/events.out.tfevents.1709242747.ip-26-0-164-207.2831688.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ec986d96e6531af2b46453256c5562f52a04c0dc10ed95586e91d902317ad966
-size 6898
+oid sha256:3f2bcf06a185979504d7c65efd1a004c040f2912d413b8849f83f48b467263c5
+size 8250