End of training

README.md

@@ -3,9 +3,15 @@ library_name: transformers
 license: apache-2.0
 base_model: alignment-handbook/zephyr-7b-sft-full
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-align-scan-0.0-1.0-linear-2
   results: []
@@ -16,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-align-scan-0.0-1.0-linear-2
 
-This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on an unknown dataset.
+This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.8359
-- Rewards/chosen: 1.9508
-- Rewards/rejected: 1.2167
+- Loss: 0.8962
+- Rewards/chosen: 1.9939
+- Rewards/rejected: 0.7547
 - Rewards/accuracies: 0.3333
-- Rewards/margins: 0.7341
-- Logps/rejected: -79.9043
-- Logps/chosen: -72.5286
-- Logits/rejected: -2.5408
-- Logits/chosen: -2.5561
+- Rewards/margins: 1.2392
+- Logps/rejected: -80.3691
+- Logps/chosen: -72.4852
+- Logits/rejected: -2.5165
+- Logits/chosen: -2.5326
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 2.0,
+    "eval_logits/chosen": -2.5326499938964844,
+    "eval_logits/rejected": -2.516491174697876,
+    "eval_logps/chosen": -72.48519134521484,
+    "eval_logps/rejected": -80.36906433105469,
+    "eval_loss": 0.8962341547012329,
+    "eval_rewards/accuracies": 0.3333333432674408,
+    "eval_rewards/chosen": 1.9939475059509277,
+    "eval_rewards/margins": 1.2392425537109375,
+    "eval_rewards/rejected": 0.7547050714492798,
+    "eval_runtime": 113.8391,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 17.569,
+    "eval_steps_per_second": 0.553,
     "total_flos": 0.0,
     "train_loss": 0.572633953144153,
     "train_runtime": 2040.9175,

config.json

@@ -22,6 +22,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 2.0,
+    "eval_logits/chosen": -2.5326499938964844,
+    "eval_logits/rejected": -2.516491174697876,
+    "eval_logps/chosen": -72.48519134521484,
+    "eval_logps/rejected": -80.36906433105469,
+    "eval_loss": 0.8962341547012329,
+    "eval_rewards/accuracies": 0.3333333432674408,
+    "eval_rewards/chosen": 1.9939475059509277,
+    "eval_rewards/margins": 1.2392425537109375,
+    "eval_rewards/rejected": 0.7547050714492798,
+    "eval_runtime": 113.8391,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 17.569,
+    "eval_steps_per_second": 0.553
+}

runs/Sep24_09-18-41_qa-a100-004.crc.nd.edu/events.out.tfevents.1727186425.qa-a100-004.crc.nd.edu.1804213.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:98d32696c213e02386c171feb1ae812cc7efa93a36e6b75787263adafe1feb0f
+size 828