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-0.05-epoch-2
   results: []
@@ -15,17 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # gemma-7b-dpo-full-mix1-beta-0.05-epoch-2
 
-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: 0.4773
-- Rewards/chosen: -2.0019
-- Rewards/rejected: -3.4640
-- Rewards/accuracies: 0.7292
-- Rewards/margins: 1.4620
-- Logps/rejected: -520.8292
-- Logps/chosen: -493.4434
-- Logits/rejected: 93.2355
-- Logits/chosen: 99.2919
+- Loss: 0.4897
+- Rewards/chosen: -2.0476
+- Rewards/rejected: -3.4508
+- Rewards/accuracies: 0.7083
+- Rewards/margins: 1.4032
+- Logps/rejected: -520.5655
+- Logps/chosen: -494.3556
+- Logits/rejected: 93.1721
+- Logits/chosen: 99.2168
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.97,
+    "eval_logits/chosen": 99.2168197631836,
+    "eval_logits/rejected": 93.17211151123047,
+    "eval_logps/chosen": -494.3555603027344,
+    "eval_logps/rejected": -520.5654907226562,
+    "eval_loss": 0.4896722733974457,
+    "eval_rewards/accuracies": 0.7083333134651184,
+    "eval_rewards/chosen": -2.047553062438965,
+    "eval_rewards/margins": 1.4032453298568726,
+    "eval_rewards/rejected": -3.450798749923706,
+    "eval_runtime": 51.7881,
+    "eval_samples": 750,
+    "eval_samples_per_second": 14.482,
+    "eval_steps_per_second": 0.463,
     "train_loss": 0.3880494168171516,
     "train_runtime": 1239.5887,
     "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": 1.97,
+    "eval_logits/chosen": 99.2168197631836,
+    "eval_logits/rejected": 93.17211151123047,
+    "eval_logps/chosen": -494.3555603027344,
+    "eval_logps/rejected": -520.5654907226562,
+    "eval_loss": 0.4896722733974457,
+    "eval_rewards/accuracies": 0.7083333134651184,
+    "eval_rewards/chosen": -2.047553062438965,
+    "eval_rewards/margins": 1.4032453298568726,
+    "eval_rewards/rejected": -3.450798749923706,
+    "eval_runtime": 51.7881,
+    "eval_samples": 750,
+    "eval_samples_per_second": 14.482,
+    "eval_steps_per_second": 0.463
+}