End of training

README.md

@@ -3,9 +3,15 @@ library_name: transformers
 license: llama3
 base_model: meta-llama/Meta-Llama-3-8B-Instruct
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- trl-lib/ultrafeedback_binarized
 model-index:
 - name: llama-3-8b-dpo-full
   results: []
@@ -16,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama-3-8b-dpo-full
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on an unknown dataset.
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the trl-lib/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6494
-- Rewards/chosen: -0.1752
-- Rewards/rejected: -0.2195
+- Loss: 0.6491
+- Rewards/chosen: -0.1814
+- Rewards/rejected: -0.2255
 - Rewards/accuracies: 0.5625
-- Rewards/margins: 0.0443
-- Logps/rejected: -418.5782
-- Logps/chosen: -335.3811
-- Logits/rejected: -1.1582
-- Logits/chosen: -1.0463
+- Rewards/margins: 0.0441
+- Logps/rejected: -419.1795
+- Logps/chosen: -335.9990
+- Logits/rejected: -1.1373
+- Logits/chosen: -1.0280
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.9969104016477858,
+    "eval_logits/chosen": -1.0279620885849,
+    "eval_logits/rejected": -1.1372919082641602,
+    "eval_logps/chosen": -335.9989929199219,
+    "eval_logps/rejected": -419.17950439453125,
+    "eval_loss": 0.6490800976753235,
+    "eval_rewards/accuracies": 0.5625,
+    "eval_rewards/chosen": -0.18141408264636993,
+    "eval_rewards/margins": 0.044088974595069885,
+    "eval_rewards/rejected": -0.22550305724143982,
+    "eval_runtime": 7.6515,
+    "eval_samples": 1000,
+    "eval_samples_per_second": 130.694,
+    "eval_steps_per_second": 1.046,
     "total_flos": 0.0,
     "train_loss": 0.6619468080110786,
     "train_runtime": 1288.9507,

config.json

@@ -25,6 +25,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.45.1",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 128256
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
-    "epoch": 0.998691442030882,
-    "eval_logits/chosen": -2.6246745586395264,
-    "eval_logits/rejected": -2.7917981147766113,
-    "eval_logps/chosen": -539.5550537109375,
-    "eval_logps/rejected": -599.9959106445312,
-    "eval_loss": 0.5575631260871887,
-    "eval_rewards/accuracies": 0.7341269850730896,
-    "eval_rewards/chosen": -1.3506927490234375,
-    "eval_rewards/margins": 0.6923374533653259,
-    "eval_rewards/rejected": -2.043030261993408,
-    "eval_runtime": 39.0144,
-    "eval_samples": 2000,
-    "eval_samples_per_second": 51.263,
-    "eval_steps_per_second": 1.615
+    "epoch": 0.9969104016477858,
+    "eval_logits/chosen": -1.0279620885849,
+    "eval_logits/rejected": -1.1372919082641602,
+    "eval_logps/chosen": -335.9989929199219,
+    "eval_logps/rejected": -419.17950439453125,
+    "eval_loss": 0.6490800976753235,
+    "eval_rewards/accuracies": 0.5625,
+    "eval_rewards/chosen": -0.18141408264636993,
+    "eval_rewards/margins": 0.044088974595069885,
+    "eval_rewards/rejected": -0.22550305724143982,
+    "eval_runtime": 7.6515,
+    "eval_samples": 1000,
+    "eval_samples_per_second": 130.694,
+    "eval_steps_per_second": 1.046
 }

runs/Oct08_11-25-31_a2ap-dgx018/events.out.tfevents.1728359699.a2ap-dgx018.1893204.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1bd0bbdd78e6c5baed8f17b81d09429e0744d1e21bf6d282a78d8edaa66391b7
+size 815