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-0913
   results: []
@@ -16,18 +22,18 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama3.1-cpo_j-full-0913
 
-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: 1.4447
-- Rewards/chosen: -16.0215
-- Rewards/rejected: -16.8041
-- Rewards/accuracies: 0.6261
-- Rewards/margins: 0.7826
-- Logps/rejected: -168.0413
-- Logps/chosen: -160.2150
-- Logits/rejected: -0.3389
-- Logits/chosen: -0.3606
-- Nll Loss: 0.2798
+- Loss: 1.4404
+- Rewards/chosen: -16.0845
+- Rewards/rejected: -16.8741
+- Rewards/accuracies: 0.6326
+- Rewards/margins: 0.7896
+- Logps/rejected: -168.7413
+- Logps/chosen: -160.8449
+- Logits/rejected: -0.3487
+- Logits/chosen: -0.3704
+- Nll Loss: 0.2790
 
 ## Model description
 

all_results.json

@@ -1,5 +1,19 @@
 {
     "epoch": 0.9985553308292401,
+    "eval_logits/chosen": -0.37041226029396057,
+    "eval_logits/rejected": -0.3487028181552887,
+    "eval_logps/chosen": -160.84490966796875,
+    "eval_logps/rejected": -168.74127197265625,
+    "eval_loss": 1.4404404163360596,
+    "eval_nll_loss": 0.278992623090744,
+    "eval_rewards/accuracies": 0.6326087117195129,
+    "eval_rewards/chosen": -16.084491729736328,
+    "eval_rewards/margins": 0.7896348834037781,
+    "eval_rewards/rejected": -16.874126434326172,
+    "eval_runtime": 76.7035,
+    "eval_samples": 1826,
+    "eval_samples_per_second": 23.806,
+    "eval_steps_per_second": 1.499,
     "total_flos": 0.0,
     "train_loss": 1.6635627029118714,
     "train_runtime": 10783.4889,

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": -0.37041226029396057,
+    "eval_logits/rejected": -0.3487028181552887,
+    "eval_logps/chosen": -160.84490966796875,
+    "eval_logps/rejected": -168.74127197265625,
+    "eval_loss": 1.4404404163360596,
+    "eval_nll_loss": 0.278992623090744,
+    "eval_rewards/accuracies": 0.6326087117195129,
+    "eval_rewards/chosen": -16.084491729736328,
+    "eval_rewards/margins": 0.7896348834037781,
+    "eval_rewards/rejected": -16.874126434326172,
+    "eval_runtime": 76.7035,
+    "eval_samples": 1826,
+    "eval_samples_per_second": 23.806,
+    "eval_steps_per_second": 1.499
+}

runs/Sep13_20-11-24_jjb_prism_dev2/events.out.tfevents.1726237696.jjb_prism_dev2.77810.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd4f66ac5041f8db204150e122444f7bf052baffe5fa7ea72e0288a96332cce4
+size 880