End of training

README.md

@@ -3,9 +3,15 @@ library_name: transformers
 license: llama3.1
 base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: lambda-llama-3-8b-dpo-test
   results: []
@@ -16,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # lambda-llama-3-8b-dpo-test
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) on an unknown dataset.
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5776
-- Rewards/chosen: -1.0426
-- Rewards/rejected: -1.6196
+- Loss: 0.5772
+- Rewards/chosen: -1.0430
+- Rewards/rejected: -1.6270
 - Rewards/accuracies: 0.7063
-- Rewards/margins: 0.5769
-- Logps/rejected: -557.9310
-- Logps/chosen: -509.0269
-- Logits/rejected: -2.6060
-- Logits/chosen: -2.4454
+- Rewards/margins: 0.5841
+- Logps/rejected: -558.6804
+- Logps/chosen: -509.0577
+- Logits/rejected: -2.5324
+- Logits/chosen: -2.3779
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.9984301412872841,
+    "eval_logits/chosen": -2.3779289722442627,
+    "eval_logits/rejected": -2.5323691368103027,
+    "eval_logps/chosen": -509.05767822265625,
+    "eval_logps/rejected": -558.6803588867188,
+    "eval_loss": 0.577161967754364,
+    "eval_rewards/accuracies": 0.7063491940498352,
+    "eval_rewards/chosen": -1.0429552793502808,
+    "eval_rewards/margins": 0.5840887427330017,
+    "eval_rewards/rejected": -1.6270442008972168,
+    "eval_runtime": 88.0288,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 22.72,
+    "eval_steps_per_second": 0.716,
     "total_flos": 0.0,
     "train_loss": 0.6095632167232361,
     "train_runtime": 6900.3625,

config.json

@@ -34,6 +34,6 @@
   "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 +1,16 @@
+{
+    "epoch": 0.9984301412872841,
+    "eval_logits/chosen": -2.3779289722442627,
+    "eval_logits/rejected": -2.5323691368103027,
+    "eval_logps/chosen": -509.05767822265625,
+    "eval_logps/rejected": -558.6803588867188,
+    "eval_loss": 0.577161967754364,
+    "eval_rewards/accuracies": 0.7063491940498352,
+    "eval_rewards/chosen": -1.0429552793502808,
+    "eval_rewards/margins": 0.5840887427330017,
+    "eval_rewards/rejected": -1.6270442008972168,
+    "eval_runtime": 88.0288,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 22.72,
+    "eval_steps_per_second": 0.716
+}

runs/Sep18_05-34-48_action-graph-trainer/events.out.tfevents.1726645681.action-graph-trainer.2590949.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cee8d99cc06b68f719923a25121d0600d7c9ba66c6c08e67a83e34513defc0ac
+size 828