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_j-full-0912
   results: []
@@ -16,18 +22,18 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama3.1-cpo_j-full-0912
 
-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: 1.4429
-- Rewards/chosen: -16.1577
-- Rewards/rejected: -16.9120
-- Rewards/accuracies: 0.6304
-- Rewards/margins: 0.7544
-- Logps/rejected: -169.1204
-- Logps/chosen: -161.5765
-- Logits/rejected: -0.3509
-- Logits/chosen: -0.3812
-- Nll Loss: 0.2845
+- Loss: 1.4395
+- Rewards/chosen: -16.1609
+- Rewards/rejected: -16.9344
+- Rewards/accuracies: 0.6326
+- Rewards/margins: 0.7735
+- Logps/rejected: -169.3439
+- Logps/chosen: -161.6093
+- Logits/rejected: -0.3578
+- Logits/chosen: -0.3883
+- Nll Loss: 0.2841
 
 ## Model description
 

all_results.json

@@ -1,5 +1,19 @@
 {
     "epoch": 0.9985553308292401,
+    "eval_logits/chosen": -0.3882608115673065,
+    "eval_logits/rejected": -0.3577967882156372,
+    "eval_logps/chosen": -161.60928344726562,
+    "eval_logps/rejected": -169.34393310546875,
+    "eval_loss": 1.4395084381103516,
+    "eval_nll_loss": 0.28411635756492615,
+    "eval_rewards/accuracies": 0.6326087117195129,
+    "eval_rewards/chosen": -16.160930633544922,
+    "eval_rewards/margins": 0.773465096950531,
+    "eval_rewards/rejected": -16.93439483642578,
+    "eval_runtime": 76.7453,
+    "eval_samples": 1826,
+    "eval_samples_per_second": 23.793,
+    "eval_steps_per_second": 1.498,
     "total_flos": 0.0,
     "train_loss": 1.6653077276768509,
     "train_runtime": 9934.0195,

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": 0.9985553308292401,
+    "eval_logits/chosen": -0.3882608115673065,
+    "eval_logits/rejected": -0.3577967882156372,
+    "eval_logps/chosen": -161.60928344726562,
+    "eval_logps/rejected": -169.34393310546875,
+    "eval_loss": 1.4395084381103516,
+    "eval_nll_loss": 0.28411635756492615,
+    "eval_rewards/accuracies": 0.6326087117195129,
+    "eval_rewards/chosen": -16.160930633544922,
+    "eval_rewards/margins": 0.773465096950531,
+    "eval_rewards/rejected": -16.93439483642578,
+    "eval_runtime": 76.7453,
+    "eval_samples": 1826,
+    "eval_samples_per_second": 23.793,
+    "eval_steps_per_second": 1.498
+}

runs/Sep12_20-48-52_jjb_prism_dev2/events.out.tfevents.1726152785.jjb_prism_dev2.76248.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:345d4080688067c750928ef1cd4c34643384ae6017626d4ae5791eba670373ae
+size 880