End of training

README.md

@@ -2,9 +2,15 @@
 license: apache-2.0
 base_model: tanliboy/zephyr-qwen2-7b-sft
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-qwen2-7b-dpo
   results: []
@@ -15,17 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-qwen2-7b-dpo
 
-This model is a fine-tuned version of [tanliboy/zephyr-qwen2-7b-sft](https://huggingface.co/tanliboy/zephyr-qwen2-7b-sft) on an unknown dataset.
+This model is a fine-tuned version of [tanliboy/zephyr-qwen2-7b-sft](https://huggingface.co/tanliboy/zephyr-qwen2-7b-sft) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5497
-- Rewards/chosen: -0.3456
-- Rewards/rejected: -0.7435
-- Rewards/accuracies: 0.7040
-- Rewards/margins: 0.3979
-- Logps/rejected: -340.0029
-- Logps/chosen: -335.0901
-- Logits/rejected: 0.0097
-- Logits/chosen: 0.1434
+- Loss: 0.5492
+- Rewards/chosen: -0.3377
+- Rewards/rejected: -0.7347
+- Rewards/accuracies: 0.7120
+- Rewards/margins: 0.3970
+- Logps/rejected: -339.1164
+- Logps/chosen: -334.2959
+- Logits/rejected: 0.0096
+- Logits/chosen: 0.1430
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.998691442030882,
+    "eval_logits/chosen": 0.14297574758529663,
+    "eval_logits/rejected": 0.009554736316204071,
+    "eval_logps/chosen": -334.2958984375,
+    "eval_logps/rejected": -339.11639404296875,
+    "eval_loss": 0.5491579174995422,
+    "eval_rewards/accuracies": 0.7120000123977661,
+    "eval_rewards/chosen": -0.3376711905002594,
+    "eval_rewards/margins": 0.3970130383968353,
+    "eval_rewards/rejected": -0.7346842288970947,
+    "eval_runtime": 86.7709,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 23.049,
+    "eval_steps_per_second": 1.441,
     "total_flos": 0.0,
     "train_loss": 0.5905132868504874,
     "train_runtime": 7548.8448,

config.json

@@ -22,7 +22,7 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.40.2",
-  "use_cache": false,
+  "use_cache": true,
   "use_sliding_window": false,
   "vocab_size": 152064
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 0.998691442030882,
+    "eval_logits/chosen": 0.14297574758529663,
+    "eval_logits/rejected": 0.009554736316204071,
+    "eval_logps/chosen": -334.2958984375,
+    "eval_logps/rejected": -339.11639404296875,
+    "eval_loss": 0.5491579174995422,
+    "eval_rewards/accuracies": 0.7120000123977661,
+    "eval_rewards/chosen": -0.3376711905002594,
+    "eval_rewards/margins": 0.3970130383968353,
+    "eval_rewards/rejected": -0.7346842288970947,
+    "eval_runtime": 86.7709,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 23.049,
+    "eval_steps_per_second": 1.441
+}

runs/Jun20_07-06-58_action-graph-trainer/events.out.tfevents.1718875722.action-graph-trainer.12409.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea145d3671e1273973879c185c80ffc6eccb1c368a31173a6fbafd2d58df94ca
+size 828