End of training

README.md

@@ -2,9 +2,15 @@
 license: apache-2.0
 base_model: martimfasantos/tinyllama-1.1b-chat-sft-full
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: tinyllama-1.1b-chat-dpo-full
   results: []
@@ -15,17 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # tinyllama-1.1b-chat-dpo-full
 
-This model is a fine-tuned version of [martimfasantos/tinyllama-1.1b-chat-sft-full](https://huggingface.co/martimfasantos/tinyllama-1.1b-chat-sft-full) on an unknown dataset.
+This model is a fine-tuned version of [martimfasantos/tinyllama-1.1b-chat-sft-full](https://huggingface.co/martimfasantos/tinyllama-1.1b-chat-sft-full) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5862
-- Rewards/chosen: -1.1608
-- Rewards/rejected: -1.6138
-- Rewards/accuracies: 0.6885
-- Rewards/margins: 0.4530
-- Logps/rejected: -458.4823
-- Logps/chosen: -452.2973
-- Logits/rejected: -2.3882
-- Logits/chosen: -2.4306
+- Loss: 0.5860
+- Rewards/chosen: -1.1602
+- Rewards/rejected: -1.6135
+- Rewards/accuracies: 0.6890
+- Rewards/margins: 0.4533
+- Logps/rejected: -458.4552
+- Logps/chosen: -452.2377
+- Logits/rejected: -2.3877
+- Logits/chosen: -2.4300
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.4300496578216553,
+    "eval_logits/rejected": -2.387667179107666,
+    "eval_logps/chosen": -452.2376708984375,
+    "eval_logps/rejected": -458.4551696777344,
+    "eval_loss": 0.5859559178352356,
+    "eval_rewards/accuracies": 0.6890000104904175,
+    "eval_rewards/chosen": -1.1601862907409668,
+    "eval_rewards/margins": 0.45331665873527527,
+    "eval_rewards/rejected": -1.613502860069275,
+    "eval_runtime": 304.9269,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 6.559,
+    "eval_steps_per_second": 0.82,
     "total_flos": 0.0,
     "train_loss": 0.6060711596530634,
     "train_runtime": 35916.4658,

config.json

@@ -24,6 +24,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "float32",
   "transformers_version": "4.41.1",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.0,
+    "eval_logits/chosen": -2.4300496578216553,
+    "eval_logits/rejected": -2.387667179107666,
+    "eval_logps/chosen": -452.2376708984375,
+    "eval_logps/rejected": -458.4551696777344,
+    "eval_loss": 0.5859559178352356,
+    "eval_rewards/accuracies": 0.6890000104904175,
+    "eval_rewards/chosen": -1.1601862907409668,
+    "eval_rewards/margins": 0.45331665873527527,
+    "eval_rewards/rejected": -1.613502860069275,
+    "eval_runtime": 304.9269,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 6.559,
+    "eval_steps_per_second": 0.82
+}

runs/Jun02_10-42-23_poseidon/events.out.tfevents.1717361786.poseidon.2371184.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e2c8613aa979165aef43571d8aba80de6a982e218c9a6a86fb2e87310f85a9a
+size 828