End of training

README.md

@@ -2,9 +2,16 @@
 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:
+- HuggingFaceH4/ultrafeedback_binarized
+- HuggingFaceH4/orca_dpo_pairs
 model-index:
 - name: gemma-7b-dpo-full-mix2-beta-0.1
   results: []
@@ -15,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # gemma-7b-dpo-full-mix2-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 HuggingFaceH4/ultrafeedback_binarized and the HuggingFaceH4/orca_dpo_pairs datasets.
 It achieves the following results on the evaluation set:
-- Loss: 0.4134
-- Rewards/chosen: -0.3763
-- Rewards/rejected: -3.5060
-- Rewards/accuracies: 0.8032
-- Rewards/margins: 3.1296
-- Logps/rejected: -413.8586
-- Logps/chosen: -392.1099
-- Logits/rejected: 83.6363
-- Logits/chosen: 82.8991
+- Loss: 0.4056
+- Rewards/chosen: -0.3995
+- Rewards/rejected: -3.5721
+- Rewards/accuracies: 0.7926
+- Rewards/margins: 3.1726
+- Logps/rejected: -414.5198
+- Logps/chosen: -392.3416
+- Logits/rejected: 83.8425
+- Logits/chosen: 83.1641
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": 83.16413879394531,
+    "eval_logits/rejected": 83.84245300292969,
+    "eval_logps/chosen": -392.341552734375,
+    "eval_logps/rejected": -414.519775390625,
+    "eval_loss": 0.40556877851486206,
+    "eval_rewards/accuracies": 0.792553186416626,
+    "eval_rewards/chosen": -0.3995126485824585,
+    "eval_rewards/margins": 3.1725716590881348,
+    "eval_rewards/rejected": -3.5720841884613037,
+    "eval_runtime": 140.3273,
+    "eval_samples": 3000,
+    "eval_samples_per_second": 21.379,
+    "eval_steps_per_second": 0.67,
     "train_loss": 0.5623811487565961,
     "train_runtime": 6612.2068,
     "train_samples": 72994,

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": 1.0,
+    "eval_logits/chosen": 83.16413879394531,
+    "eval_logits/rejected": 83.84245300292969,
+    "eval_logps/chosen": -392.341552734375,
+    "eval_logps/rejected": -414.519775390625,
+    "eval_loss": 0.40556877851486206,
+    "eval_rewards/accuracies": 0.792553186416626,
+    "eval_rewards/chosen": -0.3995126485824585,
+    "eval_rewards/margins": 3.1725716590881348,
+    "eval_rewards/rejected": -3.5720841884613037,
+    "eval_runtime": 140.3273,
+    "eval_samples": 3000,
+    "eval_samples_per_second": 21.379,
+    "eval_steps_per_second": 0.67
+}

runs/Feb29_21-52-12_ip-26-0-165-38/events.out.tfevents.1709250706.ip-26-0-165-38.2014140.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7f21e438fe89b1ea7864c63c293b8715bdc9fd552fc9478caba50f9eaec9751
+size 828