End of training

README.md

@@ -1,9 +1,16 @@
 ---
 library_name: transformers
+base_model: data/OpenELM-1_1B-SFT-2
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: OpenELM-1_1B-DPO-full-2
   results: []
@@ -14,13 +21,13 @@ should probably proofread and complete it, then remove this comment. -->
 
 # OpenELM-1_1B-DPO-full-2
 
-This model was trained from scratch on an unknown dataset.
+This model is a fine-tuned version of [data/OpenELM-1_1B-SFT-2](https://huggingface.co/data/OpenELM-1_1B-SFT-2) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.7947
+- Loss: 0.7945
 - Rewards/chosen: -8.3125
-- Rewards/rejected: -10.5
+- Rewards/rejected: -10.4375
 - Rewards/accuracies: 0.7324
-- Rewards/margins: 2.1562
+- Rewards/margins: 2.1406
 - Logps/rejected: -1336.0
 - Logps/chosen: -1144.0
 - Logits/rejected: 5.5

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 2.998430141287284,
+    "eval_logits/chosen": 3.59375,
+    "eval_logits/rejected": 5.5,
+    "eval_logps/chosen": -1144.0,
+    "eval_logps/rejected": -1336.0,
+    "eval_loss": 0.7945234179496765,
+    "eval_rewards/accuracies": 0.732421875,
+    "eval_rewards/chosen": -8.3125,
+    "eval_rewards/margins": 2.140625,
+    "eval_rewards/rejected": -10.4375,
+    "eval_runtime": 46.9097,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 42.635,
+    "eval_steps_per_second": 0.682,
     "total_flos": 0.0,
     "train_loss": 0.26257153015070145,
     "train_runtime": 12376.5138,

config.json

@@ -120,6 +120,6 @@
   "share_input_output_layers": true,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 2.998430141287284,
+    "eval_logits/chosen": 3.59375,
+    "eval_logits/rejected": 5.5,
+    "eval_logps/chosen": -1144.0,
+    "eval_logps/rejected": -1336.0,
+    "eval_loss": 0.7945234179496765,
+    "eval_rewards/accuracies": 0.732421875,
+    "eval_rewards/chosen": -8.3125,
+    "eval_rewards/margins": 2.140625,
+    "eval_rewards/rejected": -10.4375,
+    "eval_runtime": 46.9097,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 42.635,
+    "eval_steps_per_second": 0.682
+}

runs/Aug29_19-27-59_xe8545-a100-17/events.out.tfevents.1724965454.xe8545-a100-17.2472796.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0790dc32d1801ab6e4f1b7077af7df0dceabfc95f28dc96ad7537a3b31a25294
+size 828