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
 - cpo
 - generated_from_trainer
+- trl
+- cpo
+- generated_from_trainer
+datasets:
+- princeton-nlp/llama3-ultrafeedback
 model-index:
 - name: llama3.1-cpo-full-0911
   results: []
@@ -16,7 +22,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama3.1-cpo-full-0911
 
-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 princeton-nlp/llama3-ultrafeedback dataset.
 It achieves the following results on the evaluation set:
 - Loss: 1.5984
 - Rewards/chosen: -14.3945

all_results.json

@@ -1,5 +1,19 @@
 {
     "epoch": 2.9956659924877203,
+    "eval_logits/chosen": -0.34079235792160034,
+    "eval_logits/rejected": -0.31415677070617676,
+    "eval_logps/chosen": -143.9447784423828,
+    "eval_logps/rejected": -155.83648681640625,
+    "eval_loss": 1.5984355211257935,
+    "eval_nll_loss": 0.3937048017978668,
+    "eval_rewards/accuracies": 0.6304348111152649,
+    "eval_rewards/chosen": -14.394478797912598,
+    "eval_rewards/margins": 1.1891697645187378,
+    "eval_rewards/rejected": -15.583648681640625,
+    "eval_runtime": 73.3871,
+    "eval_samples": 1826,
+    "eval_samples_per_second": 24.882,
+    "eval_steps_per_second": 1.567,
     "total_flos": 0.0,
     "train_loss": 1.025652496903031,
     "train_runtime": 27913.7357,

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 +1,17 @@
+{
+    "epoch": 2.9956659924877203,
+    "eval_logits/chosen": -0.34079235792160034,
+    "eval_logits/rejected": -0.31415677070617676,
+    "eval_logps/chosen": -143.9447784423828,
+    "eval_logps/rejected": -155.83648681640625,
+    "eval_loss": 1.5984355211257935,
+    "eval_nll_loss": 0.3937048017978668,
+    "eval_rewards/accuracies": 0.6304348111152649,
+    "eval_rewards/chosen": -14.394478797912598,
+    "eval_rewards/margins": 1.1891697645187378,
+    "eval_rewards/rejected": -15.583648681640625,
+    "eval_runtime": 73.3871,
+    "eval_samples": 1826,
+    "eval_samples_per_second": 24.882,
+    "eval_steps_per_second": 1.567
+}

runs/Sep12_07-12-04_jjb_prism_dev2/events.out.tfevents.1726121763.jjb_prism_dev2.69652.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eca4c585d7c62219dd15b0ef2eb0bffd8a137a6058b86bbb1e6d4beac92200d2
+size 880