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-ipo
   results: []
@@ -15,17 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-gemma-ipo
 
-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: 60.8626
-- Rewards/chosen: -0.5007
-- Rewards/rejected: -0.6906
+- Loss: 61.0152
+- Rewards/chosen: -0.4988
+- Rewards/rejected: -0.6909
 - Rewards/accuracies: 0.8021
-- Rewards/margins: 0.1899
-- Logps/rejected: -15.3697
-- Logps/chosen: -11.4648
-- Logits/rejected: 99.7591
-- Logits/chosen: 99.5497
+- Rewards/margins: 0.1921
+- Logps/rejected: -15.3755
+- Logps/chosen: -11.4268
+- Logits/rejected: 99.7522
+- Logits/chosen: 99.5411
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.971563981042654,
+    "eval_logits/chosen": 99.54110717773438,
+    "eval_logits/rejected": 99.75216674804688,
+    "eval_logps/chosen": -11.42675495147705,
+    "eval_logps/rejected": -15.375484466552734,
+    "eval_loss": 61.015228271484375,
+    "eval_rewards/accuracies": 0.8020833134651184,
+    "eval_rewards/chosen": -0.49881264567375183,
+    "eval_rewards/margins": 0.1921073943376541,
+    "eval_rewards/rejected": -0.6909201145172119,
+    "eval_runtime": 120.2204,
+    "eval_samples": 750,
+    "eval_samples_per_second": 6.239,
+    "eval_steps_per_second": 0.2,
     "total_flos": 0.0,
     "train_loss": 77.37614455589882,
     "train_runtime": 2312.9834,

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

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.971563981042654,
+    "eval_logits/chosen": 99.54110717773438,
+    "eval_logits/rejected": 99.75216674804688,
+    "eval_logps/chosen": -11.42675495147705,
+    "eval_logps/rejected": -15.375484466552734,
+    "eval_loss": 61.015228271484375,
+    "eval_rewards/accuracies": 0.8020833134651184,
+    "eval_rewards/chosen": -0.49881264567375183,
+    "eval_rewards/margins": 0.1921073943376541,
+    "eval_rewards/rejected": -0.6909201145172119,
+    "eval_runtime": 120.2204,
+    "eval_samples": 750,
+    "eval_samples_per_second": 6.239,
+    "eval_steps_per_second": 0.2
+}

runs/Apr27_17-32-57_660111d13776/events.out.tfevents.1714238537.660111d13776.65508.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd81d46a89a7691c8aac101fc3fad898d4bc71b21a988a9dac0fcb457eff3d0a
+size 815