End of training

README.md

@@ -3,9 +3,15 @@ library_name: transformers
 license: apache-2.0
 base_model: alignment-handbook/zephyr-7b-sft-full
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-align-scan-0.0-0.2-polynomial-3
   results: []
@@ -16,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-align-scan-0.0-0.2-polynomial-3
 
-This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on an unknown dataset.
+This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6467
-- Rewards/chosen: -0.1914
-- Rewards/rejected: -0.6247
+- Loss: 0.7011
+- Rewards/chosen: -1.4864
+- Rewards/rejected: -2.0925
 - Rewards/accuracies: 0.3313
-- Rewards/margins: 0.4333
-- Logps/rejected: -83.7654
-- Logps/chosen: -75.2992
-- Logits/rejected: -2.5078
-- Logits/chosen: -2.5251
+- Rewards/margins: 0.6061
+- Logps/rejected: -89.9617
+- Logps/chosen: -80.7659
+- Logits/rejected: -2.4878
+- Logits/chosen: -2.5051
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 3.0,
+    "eval_logits/chosen": -2.5050723552703857,
+    "eval_logits/rejected": -2.487805128097534,
+    "eval_logps/chosen": -80.76593017578125,
+    "eval_logps/rejected": -89.9616928100586,
+    "eval_loss": 0.701084554195404,
+    "eval_rewards/accuracies": 0.3313491940498352,
+    "eval_rewards/chosen": -1.4863682985305786,
+    "eval_rewards/margins": 0.6061023473739624,
+    "eval_rewards/rejected": -2.092470407485962,
+    "eval_runtime": 114.0061,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 17.543,
+    "eval_steps_per_second": 0.553,
     "total_flos": 0.0,
     "train_loss": 0.4920133708251847,
     "train_runtime": 3526.2726,

config.json

@@ -22,6 +22,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 3.0,
+    "eval_logits/chosen": -2.5050723552703857,
+    "eval_logits/rejected": -2.487805128097534,
+    "eval_logps/chosen": -80.76593017578125,
+    "eval_logps/rejected": -89.9616928100586,
+    "eval_loss": 0.701084554195404,
+    "eval_rewards/accuracies": 0.3313491940498352,
+    "eval_rewards/chosen": -1.4863682985305786,
+    "eval_rewards/margins": 0.6061023473739624,
+    "eval_rewards/rejected": -2.092470407485962,
+    "eval_runtime": 114.0061,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 17.543,
+    "eval_steps_per_second": 0.553
+}

runs/Sep24_12-46-46_qa-a100-004.crc.nd.edu/events.out.tfevents.1727200406.qa-a100-004.crc.nd.edu.1837605.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14486b1fb737145d7ec8efc24f3730c70aeb98d4be728d94f7a6209a2bdc956f
+size 828