End of training

README.md

@@ -1,8 +1,11 @@
 ---
 base_model: deepseek-ai/deepseek-math-7b-base
+datasets:
+- zfz1/my_preference_gsm8k_deepseek
 library_name: peft
 license: other
 tags:
+- alignment-handbook
 - trl
 - orpo
 - generated_from_trainer
@@ -17,7 +20,20 @@ should probably proofread and complete it, then remove this comment. -->
 [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/thuzfz1/huggingface/runs/z4q90sz1)
 # deepseek-8b-orpo-lora
 
-This model is a fine-tuned version of [deepseek-ai/deepseek-math-7b-base](https://huggingface.co/deepseek-ai/deepseek-math-7b-base) on an unknown dataset.
+This model is a fine-tuned version of [deepseek-ai/deepseek-math-7b-base](https://huggingface.co/deepseek-ai/deepseek-math-7b-base) on the zfz1/my_preference_gsm8k_deepseek dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.6818
+- Rewards/chosen: -0.0338
+- Rewards/rejected: -0.0840
+- Rewards/accuracies: 0.8088
+- Rewards/margins: 0.0502
+- Logps/rejected: -0.8398
+- Logps/chosen: -0.3377
+- Logits/rejected: 34.4233
+- Logits/chosen: 35.5254
+- Nll Loss: 0.6414
+- Log Odds Ratio: -0.4212
+- Log Odds Chosen: 1.0634
 
 ## Model description
 

all_results.json

@@ -1,5 +1,21 @@
 {
     "epoch": 1.9936102236421727,
+    "eval_log_odds_chosen": 1.0634132623672485,
+    "eval_log_odds_ratio": -0.421150267124176,
+    "eval_logits/chosen": 35.52544021606445,
+    "eval_logits/rejected": 34.42332077026367,
+    "eval_logps/chosen": -0.3376733958721161,
+    "eval_logps/rejected": -0.8398498296737671,
+    "eval_loss": 0.6817505359649658,
+    "eval_nll_loss": 0.6413611173629761,
+    "eval_rewards/accuracies": 0.8088235259056091,
+    "eval_rewards/chosen": -0.033767346292734146,
+    "eval_rewards/margins": 0.0502176471054554,
+    "eval_rewards/rejected": -0.08398497849702835,
+    "eval_runtime": 252.7054,
+    "eval_samples": 5398,
+    "eval_samples_per_second": 21.361,
+    "eval_steps_per_second": 0.336,
     "total_flos": 0.0,
     "train_loss": 0.7846963420892373,
     "train_runtime": 5577.3844,

eval_results.json

@@ -1,19 +1,19 @@
 {
     "epoch": 1.9936102236421727,
-    "eval_log_odds_chosen": 0.05526316165924072,
-    "eval_log_odds_ratio": -0.6908599138259888,
-    "eval_logits/chosen": 30.44268798828125,
-    "eval_logits/rejected": 30.32880210876465,
-    "eval_logps/chosen": -0.7262152433395386,
-    "eval_logps/rejected": -0.7568790316581726,
-    "eval_loss": 1.1456295251846313,
-    "eval_nll_loss": 1.071852684020996,
-    "eval_rewards/accuracies": 0.5647059082984924,
-    "eval_rewards/chosen": -0.07262152433395386,
-    "eval_rewards/margins": 0.0030663711950182915,
-    "eval_rewards/rejected": -0.07568789273500443,
-    "eval_runtime": 247.9026,
+    "eval_log_odds_chosen": 1.0634132623672485,
+    "eval_log_odds_ratio": -0.421150267124176,
+    "eval_logits/chosen": 35.52544021606445,
+    "eval_logits/rejected": 34.42332077026367,
+    "eval_logps/chosen": -0.3376733958721161,
+    "eval_logps/rejected": -0.8398498296737671,
+    "eval_loss": 0.6817505359649658,
+    "eval_nll_loss": 0.6413611173629761,
+    "eval_rewards/accuracies": 0.8088235259056091,
+    "eval_rewards/chosen": -0.033767346292734146,
+    "eval_rewards/margins": 0.0502176471054554,
+    "eval_rewards/rejected": -0.08398497849702835,
+    "eval_runtime": 252.7054,
     "eval_samples": 5398,
-    "eval_samples_per_second": 21.775,
-    "eval_steps_per_second": 0.343
+    "eval_samples_per_second": 21.361,
+    "eval_steps_per_second": 0.336
 }

runs/Jul18_00-59-35_phe108-jieyuzhao-01/events.out.tfevents.1721295716.phe108-jieyuzhao-01.335789.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d0eb6ffd43afd6681159b970e21b0b92660842bb289b7862d8b896f1abca581
+size 997