End of training

README.md

@@ -2,9 +2,15 @@
 license: gemma
 base_model: tanliboy/zephyr-7b-gemma-sft
 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 [tanliboy/zephyr-7b-gemma-sft](https://huggingface.co/tanliboy/zephyr-7b-gemma-sft) on an unknown dataset.
+This model is a fine-tuned version of [tanliboy/zephyr-7b-gemma-sft](https://huggingface.co/tanliboy/zephyr-7b-gemma-sft) on the argilla/dpo-mix-7k dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4722
-- Rewards/chosen: -0.0658
-- Rewards/rejected: -1.2673
+- Loss: 0.4833
+- Rewards/chosen: -0.1121
+- Rewards/rejected: -1.2301
 - Rewards/accuracies: 0.7396
-- Rewards/margins: 1.2015
-- Logps/rejected: -720.2745
-- Logps/chosen: -697.6023
-- Logits/rejected: 152.9660
-- Logits/chosen: 153.1356
+- Rewards/margins: 1.1180
+- Logps/rejected: -719.5310
+- Logps/chosen: -698.5295
+- Logits/rejected: 153.0066
+- Logits/chosen: 153.1078
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.971563981042654,
+    "eval_logits/chosen": 153.10781860351562,
+    "eval_logits/rejected": 153.006591796875,
+    "eval_logps/chosen": -698.5294799804688,
+    "eval_logps/rejected": -719.531005859375,
+    "eval_loss": 0.4833243787288666,
+    "eval_rewards/accuracies": 0.7395833134651184,
+    "eval_rewards/chosen": -0.11214617639780045,
+    "eval_rewards/margins": 1.1179896593093872,
+    "eval_rewards/rejected": -1.2301357984542847,
+    "eval_runtime": 116.3725,
+    "eval_samples": 750,
+    "eval_samples_per_second": 6.445,
+    "eval_steps_per_second": 0.206,
     "total_flos": 0.0,
     "train_loss": 0.3875045489806395,
     "train_runtime": 2331.3598,

config.json

@@ -24,6 +24,6 @@
   "rope_theta": 10000.0,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.40.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 256000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.971563981042654,
+    "eval_logits/chosen": 153.10781860351562,
+    "eval_logits/rejected": 153.006591796875,
+    "eval_logps/chosen": -698.5294799804688,
+    "eval_logps/rejected": -719.531005859375,
+    "eval_loss": 0.4833243787288666,
+    "eval_rewards/accuracies": 0.7395833134651184,
+    "eval_rewards/chosen": -0.11214617639780045,
+    "eval_rewards/margins": 1.1179896593093872,
+    "eval_rewards/rejected": -1.2301357984542847,
+    "eval_runtime": 116.3725,
+    "eval_samples": 750,
+    "eval_samples_per_second": 6.445,
+    "eval_steps_per_second": 0.206
+}