End of training

README.md

@@ -1,9 +1,15 @@
 ---
 base_model: meta-llama/Llama-2-7b-hf
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: llama2-7b-dpo-full-wo-live_qa-ep3
   results: []
@@ -14,17 +20,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama2-7b-dpo-full-wo-live_qa-ep3
 
-This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on an unknown dataset.
+This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6297
-- Rewards/chosen: 0.0796
-- Rewards/rejected: -0.0955
-- Rewards/accuracies: 0.7679
-- Rewards/margins: 0.1751
-- Logps/rejected: -812.6376
-- Logps/chosen: -523.0222
-- Logits/rejected: -0.3805
-- Logits/chosen: -0.1364
+- Loss: 0.6073
+- Rewards/chosen: 0.0705
+- Rewards/rejected: -0.1795
+- Rewards/accuracies: 0.8214
+- Rewards/margins: 0.2499
+- Logps/rejected: -821.0361
+- Logps/chosen: -523.9339
+- Logits/rejected: -0.3815
+- Logits/chosen: -0.1586
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -0.15863379836082458,
+    "eval_logits/rejected": -0.3815014362335205,
+    "eval_logps/chosen": -523.9338989257812,
+    "eval_logps/rejected": -821.0360717773438,
+    "eval_loss": 0.6073186993598938,
+    "eval_rewards/accuracies": 0.8214285969734192,
+    "eval_rewards/chosen": 0.07047966867685318,
+    "eval_rewards/margins": 0.24993744492530823,
+    "eval_rewards/rejected": -0.17945778369903564,
+    "eval_runtime": 22.8385,
+    "eval_samples": 208,
+    "eval_samples_per_second": 9.107,
+    "eval_steps_per_second": 0.307,
     "train_loss": 0.6189163742643414,
     "train_runtime": 2381.9724,
     "train_samples": 10591,

config.json

@@ -23,6 +23,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.39.0.dev0",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.0,
+    "eval_logits/chosen": -0.15863379836082458,
+    "eval_logits/rejected": -0.3815014362335205,
+    "eval_logps/chosen": -523.9338989257812,
+    "eval_logps/rejected": -821.0360717773438,
+    "eval_loss": 0.6073186993598938,
+    "eval_rewards/accuracies": 0.8214285969734192,
+    "eval_rewards/chosen": 0.07047966867685318,
+    "eval_rewards/margins": 0.24993744492530823,
+    "eval_rewards/rejected": -0.17945778369903564,
+    "eval_runtime": 22.8385,
+    "eval_samples": 208,
+    "eval_samples_per_second": 9.107,
+    "eval_steps_per_second": 0.307
+}