End of training

README.md

@@ -2,9 +2,15 @@
 license: apache-2.0
 base_model: martimfasantos/tinyllama-1.1b-sum-sft-full_old
 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-simpo_beta2.0_gamma1.6_LR5e-8_3epochs
   results: []
@@ -15,7 +21,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # tinyllama-1.1b-sum-simpo_beta2.0_gamma1.6_LR5e-8_3epochs
 
-This model is a fine-tuned version of [martimfasantos/tinyllama-1.1b-sum-sft-full_old](https://huggingface.co/martimfasantos/tinyllama-1.1b-sum-sft-full_old) on an unknown dataset.
+This model is a fine-tuned version of [martimfasantos/tinyllama-1.1b-sum-sft-full_old](https://huggingface.co/martimfasantos/tinyllama-1.1b-sum-sft-full_old) on the openai/summarize_from_feedback dataset.
 It achieves the following results on the evaluation set:
 - Loss: 1.5991
 - Rewards/chosen: -4.1926

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 3.0,
+    "eval_logits/chosen": -3.337226629257202,
+    "eval_logits/rejected": -3.3338418006896973,
+    "eval_logps/chosen": -2.096304178237915,
+    "eval_logps/rejected": -2.3068747520446777,
+    "eval_loss": 1.5990804433822632,
+    "eval_rewards/accuracies": 0.6231412887573242,
+    "eval_rewards/chosen": -4.19260835647583,
+    "eval_rewards/margins": 0.4211415648460388,
+    "eval_rewards/rejected": -4.6137495040893555,
+    "eval_runtime": 154.7299,
+    "eval_samples": 4304,
+    "eval_samples_per_second": 27.816,
+    "eval_steps_per_second": 3.477,
     "total_flos": 0.0,
     "train_loss": 1.5625922015596196,
     "train_runtime": 80753.7869,

config.json

@@ -24,6 +24,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "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": -3.337226629257202,
+    "eval_logits/rejected": -3.3338418006896973,
+    "eval_logps/chosen": -2.096304178237915,
+    "eval_logps/rejected": -2.3068747520446777,
+    "eval_loss": 1.5990804433822632,
+    "eval_rewards/accuracies": 0.6231412887573242,
+    "eval_rewards/chosen": -4.19260835647583,
+    "eval_rewards/margins": 0.4211415648460388,
+    "eval_rewards/rejected": -4.6137495040893555,
+    "eval_runtime": 154.7299,
+    "eval_samples": 4304,
+    "eval_samples_per_second": 27.816,
+    "eval_steps_per_second": 3.477
+}

runs/Jul05_13-38-00_poseidon/events.out.tfevents.1720268336.poseidon.1633264.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2bd2af206877c355140d2368766ae9ae31c8d50ff7516286d0f1abf13830c01b
+size 841