End of training

README.md

@@ -3,9 +3,15 @@ library_name: transformers
 license: other
 base_model: trl-lib/qwen1.5-0.5b-sft
 tags:
+- alignment-handbook
 - trl
 - simpo
 - generated_from_trainer
+- trl
+- simpo
+- generated_from_trainer
+datasets:
+- yakazimir/ultrafeedback_binarized
 model-index:
 - name: qwen_cUNL_entropy_0_01
   results: []
@@ -16,18 +22,18 @@ should probably proofread and complete it, then remove this comment. -->
 
 # qwen_cUNL_entropy_0_01
 
-This model is a fine-tuned version of [trl-lib/qwen1.5-0.5b-sft](https://huggingface.co/trl-lib/qwen1.5-0.5b-sft) on an unknown dataset.
+This model is a fine-tuned version of [trl-lib/qwen1.5-0.5b-sft](https://huggingface.co/trl-lib/qwen1.5-0.5b-sft) on the yakazimir/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.5806
 - Sft Loss: 4.6392
-- Rewards/chosen: -4.6683
+- Rewards/chosen: -4.6682
 - Rewards/rejected: -5.7215
 - Rewards/accuracies: 0.7292
 - Rewards/margins: 1.0533
 - Logps/rejected: -5.7215
-- Logps/chosen: -4.6683
+- Logps/chosen: -4.6682
 - Logits/rejected: 0.0927
-- Logits/chosen: 0.0156
+- Logits/chosen: 0.0157
 
 ## Model description
 

all_results.json

@@ -1,5 +1,19 @@
 {
     "epoch": 2.999297541394882,
+    "eval_logits/chosen": 0.015653692185878754,
+    "eval_logits/rejected": 0.0926935076713562,
+    "eval_logps/chosen": -4.668216705322266,
+    "eval_logps/rejected": -5.721502304077148,
+    "eval_loss": 0.5805643200874329,
+    "eval_rewards/accuracies": 0.7292284965515137,
+    "eval_rewards/chosen": -4.668216705322266,
+    "eval_rewards/margins": 1.0532848834991455,
+    "eval_rewards/rejected": -5.721502304077148,
+    "eval_runtime": 42.6158,
+    "eval_samples": 1345,
+    "eval_samples_per_second": 31.561,
+    "eval_sft_loss": 4.639245510101318,
+    "eval_steps_per_second": 7.908,
     "total_flos": 0.0,
     "train_loss": 0.5697717436673725,
     "train_runtime": 31532.5259,

config.json

@@ -22,7 +22,7 @@
   "tie_word_embeddings": true,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.2",
-  "use_cache": false,
+  "use_cache": true,
   "use_sliding_window": false,
   "vocab_size": 151648
 }

eval_results.json

@@ -0,0 +1,17 @@
+{
+    "epoch": 2.999297541394882,
+    "eval_logits/chosen": 0.015653692185878754,
+    "eval_logits/rejected": 0.0926935076713562,
+    "eval_logps/chosen": -4.668216705322266,
+    "eval_logps/rejected": -5.721502304077148,
+    "eval_loss": 0.5805643200874329,
+    "eval_rewards/accuracies": 0.7292284965515137,
+    "eval_rewards/chosen": -4.668216705322266,
+    "eval_rewards/margins": 1.0532848834991455,
+    "eval_rewards/rejected": -5.721502304077148,
+    "eval_runtime": 42.6158,
+    "eval_samples": 1345,
+    "eval_samples_per_second": 31.561,
+    "eval_sft_loss": 4.639245510101318,
+    "eval_steps_per_second": 7.908
+}