End of training

README.md

@@ -2,9 +2,15 @@
 license: gemma
 base_model: google/gemma-2b
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- argilla/dpo-mix-7k
 model-index:
 - name: zephyr-2b-gemma-dpo
   results: []
@@ -16,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://zebra.wandb.io/cto/distillm/runs/vnc8ka21)
 # zephyr-2b-gemma-dpo
 
-This model is a fine-tuned version of [google/gemma-2b](https://huggingface.co/google/gemma-2b) on an unknown dataset.
+This model is a fine-tuned version of [google/gemma-2b](https://huggingface.co/google/gemma-2b) on the argilla/dpo-mix-7k dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6497
-- Rewards/chosen: -0.0395
-- Rewards/rejected: -0.1328
-- Rewards/accuracies: 0.6354
-- Rewards/margins: 0.0933
-- Logps/rejected: -378.4776
-- Logps/chosen: -386.1444
-- Logits/rejected: -25.9802
-- Logits/chosen: -26.9529
+- Loss: 0.6493
+- Rewards/chosen: -0.0415
+- Rewards/rejected: -0.1402
+- Rewards/accuracies: 0.6875
+- Rewards/margins: 0.0986
+- Logps/rejected: -378.6258
+- Logps/chosen: -386.1853
+- Logits/rejected: -25.9826
+- Logits/chosen: -26.9604
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.971563981042654,
+    "eval_logits/chosen": -26.960426330566406,
+    "eval_logits/rejected": -25.98261260986328,
+    "eval_logps/chosen": -386.185302734375,
+    "eval_logps/rejected": -378.6257629394531,
+    "eval_loss": 0.6492888331413269,
+    "eval_rewards/accuracies": 0.6875,
+    "eval_rewards/chosen": -0.04152223840355873,
+    "eval_rewards/margins": 0.09864477068185806,
+    "eval_rewards/rejected": -0.1401670128107071,
+    "eval_runtime": 6.8392,
+    "eval_samples": 750,
+    "eval_samples_per_second": 109.662,
+    "eval_steps_per_second": 3.509,
     "total_flos": 0.0,
     "train_loss": 0.661209049133154,
     "train_runtime": 432.2704,

config.json

@@ -24,6 +24,6 @@
   "rope_theta": 10000.0,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.42.4",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 256000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.971563981042654,
+    "eval_logits/chosen": -26.960426330566406,
+    "eval_logits/rejected": -25.98261260986328,
+    "eval_logps/chosen": -386.185302734375,
+    "eval_logps/rejected": -378.6257629394531,
+    "eval_loss": 0.6492888331413269,
+    "eval_rewards/accuracies": 0.6875,
+    "eval_rewards/chosen": -0.04152223840355873,
+    "eval_rewards/margins": 0.09864477068185806,
+    "eval_rewards/rejected": -0.1401670128107071,
+    "eval_runtime": 6.8392,
+    "eval_samples": 750,
+    "eval_samples_per_second": 109.662,
+    "eval_steps_per_second": 3.509
+}

runs/Jul12_09-48-54_ale-distillm-8-0-0/events.out.tfevents.1720774969.ale-distillm-8-0-0.1202.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9d915ada9c410742f74fb7bf0195b865f90ed07b43ea56d354150997db6a4f3
+size 815