End of training

README.md

@@ -2,9 +2,15 @@
 license: apache-2.0
 base_model: martimfasantos/tinyllama-1.1b-sum-sft-full
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- openai/summarize_from_feedback
 model-index:
 - name: tinyllama-1.1b-sum-dpo-full_LR2e-7_3epochs
   results: []
@@ -15,17 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # tinyllama-1.1b-sum-dpo-full_LR2e-7_3epochs
 
-This model is a fine-tuned version of [martimfasantos/tinyllama-1.1b-sum-sft-full](https://huggingface.co/martimfasantos/tinyllama-1.1b-sum-sft-full) on an unknown dataset.
+This model is a fine-tuned version of [martimfasantos/tinyllama-1.1b-sum-sft-full](https://huggingface.co/martimfasantos/tinyllama-1.1b-sum-sft-full) on the openai/summarize_from_feedback dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6412
-- Rewards/chosen: -1.5933
-- Rewards/rejected: -1.9040
-- Rewards/accuracies: 0.6296
-- Rewards/margins: 0.3107
-- Logps/rejected: -253.1478
-- Logps/chosen: -218.3473
-- Logits/rejected: -2.1506
-- Logits/chosen: -2.1702
+- Loss: 0.6411
+- Rewards/chosen: -1.5955
+- Rewards/rejected: -1.9066
+- Rewards/accuracies: 0.6273
+- Rewards/margins: 0.3112
+- Logps/rejected: -253.4108
+- Logps/chosen: -218.5612
+- Logits/rejected: -2.1502
+- Logits/chosen: -2.1697
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 3.0,
+    "eval_logits/chosen": -2.1697230339050293,
+    "eval_logits/rejected": -2.15024995803833,
+    "eval_logps/chosen": -218.56118774414062,
+    "eval_logps/rejected": -253.41075134277344,
+    "eval_loss": 0.641146719455719,
+    "eval_rewards/accuracies": 0.6273234486579895,
+    "eval_rewards/chosen": -1.5954570770263672,
+    "eval_rewards/margins": 0.3111543357372284,
+    "eval_rewards/rejected": -1.906611442565918,
+    "eval_runtime": 381.9987,
+    "eval_samples": 4304,
+    "eval_samples_per_second": 11.267,
+    "eval_steps_per_second": 1.408,
     "total_flos": 0.0,
     "train_loss": 0.5850592939272546,
     "train_runtime": 88547.231,

config.json

@@ -24,6 +24,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "float32",
   "transformers_version": "4.41.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 3.0,
+    "eval_logits/chosen": -2.1697230339050293,
+    "eval_logits/rejected": -2.15024995803833,
+    "eval_logps/chosen": -218.56118774414062,
+    "eval_logps/rejected": -253.41075134277344,
+    "eval_loss": 0.641146719455719,
+    "eval_rewards/accuracies": 0.6273234486579895,
+    "eval_rewards/chosen": -1.5954570770263672,
+    "eval_rewards/margins": 0.3111543357372284,
+    "eval_rewards/rejected": -1.906611442565918,
+    "eval_runtime": 381.9987,
+    "eval_samples": 4304,
+    "eval_samples_per_second": 11.267,
+    "eval_steps_per_second": 1.408
+}

runs/Jun08_12-33-54_poseidon/events.out.tfevents.1717939465.poseidon.3992514.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df2116f9b9a151011ad656fefa92e8cf1ced24e6a11a602a9ab987d2a82fe203
+size 841