End of training

README.md

@@ -3,9 +3,15 @@ library_name: transformers
 license: llama3
 base_model: meta-llama/Meta-Llama-3-8B-Instruct
 tags:
+- alignment-handbook
 - trl
 - simpo
 - generated_from_trainer
+- trl
+- simpo
+- generated_from_trainer
+datasets:
+- yakazimir/llama3-ultrafeedback-armorm
 model-index:
 - name: llama3_qfUNL_best_entropy
   results: []
@@ -16,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama3_qfUNL_best_entropy
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on an unknown dataset.
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the yakazimir/llama3-ultrafeedback-armorm dataset.
 It achieves the following results on the evaluation set:
-- Loss: 2.0936
-- Rewards/chosen: -5.3144
-- Rewards/rejected: -7.1419
-- Rewards/accuracies: 0.7771
-- Rewards/margins: 1.8275
-- Logps/rejected: -0.7142
+- Loss: 2.0906
+- Rewards/chosen: -5.3136
+- Rewards/rejected: -7.1413
+- Rewards/accuracies: 0.7741
+- Rewards/margins: 1.8277
+- Logps/rejected: -0.7141
 - Logps/chosen: -0.5314
-- Logits/rejected: -1.3326
-- Logits/chosen: -1.3731
+- Logits/rejected: -1.3346
+- Logits/chosen: -1.3749
 - Semantic Entropy: 0.9976
 
 ## Model description

all_results.json

@@ -1,5 +1,19 @@
 {
     "epoch": 0.9989071038251366,
+    "eval_logits/chosen": -1.3748970031738281,
+    "eval_logits/rejected": -1.3346441984176636,
+    "eval_logps/chosen": -0.5313604474067688,
+    "eval_logps/rejected": -0.7141343355178833,
+    "eval_loss": 2.09059739112854,
+    "eval_rewards/accuracies": 0.7740963697433472,
+    "eval_rewards/chosen": -5.313604354858398,
+    "eval_rewards/margins": 1.8277385234832764,
+    "eval_rewards/rejected": -7.141342639923096,
+    "eval_runtime": 33.4268,
+    "eval_samples": 1318,
+    "eval_samples_per_second": 39.429,
+    "eval_semantic_entropy": 0.997556209564209,
+    "eval_steps_per_second": 2.483,
     "total_flos": 0.0,
     "train_loss": 2.4655840506438875,
     "train_runtime": 5955.1851,

config.json

@@ -24,6 +24,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 128256
 }

eval_results.json

@@ -0,0 +1,17 @@
+{
+    "epoch": 0.9989071038251366,
+    "eval_logits/chosen": -1.3748970031738281,
+    "eval_logits/rejected": -1.3346441984176636,
+    "eval_logps/chosen": -0.5313604474067688,
+    "eval_logps/rejected": -0.7141343355178833,
+    "eval_loss": 2.09059739112854,
+    "eval_rewards/accuracies": 0.7740963697433472,
+    "eval_rewards/chosen": -5.313604354858398,
+    "eval_rewards/margins": 1.8277385234832764,
+    "eval_rewards/rejected": -7.141342639923096,
+    "eval_runtime": 33.4268,
+    "eval_samples": 1318,
+    "eval_samples_per_second": 39.429,
+    "eval_semantic_entropy": 0.997556209564209,
+    "eval_steps_per_second": 2.483
+}