Model save

README.md

@@ -0,0 +1,77 @@
+---
+license: gemma
+library_name: peft
+tags:
+- trl
+- dpo
+- llama-factory
+- generated_from_trainer
+base_model: google/gemma-2b-it
+model-index:
+- name: Gemma-2B-It-ORPO
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# Gemma-2B-It-ORPO
+
+This model is a fine-tuned version of [google/gemma-2b-it](https://huggingface.co/google/gemma-2b-it) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.3996
+- Rewards/chosen: -0.1332
+- Rewards/rejected: -0.1501
+- Rewards/accuracies: 0.5460
+- Rewards/margins: 0.0169
+- Logps/rejected: -1.5010
+- Logps/chosen: -1.3319
+- Logits/rejected: -20.7443
+- Logits/chosen: -20.6555
+- Sft Loss: 1.3319
+- Odds Ratio Loss: 0.6769
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 5e-06
+- train_batch_size: 2
+- eval_batch_size: 2
+- seed: 42
+- gradient_accumulation_steps: 8
+- total_train_batch_size: 16
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 0.1
+- num_epochs: 3.0
+
+### Training results
+
+| Training Loss | Epoch  | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen | Sft Loss | Odds Ratio Loss |
+|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|:--------:|:---------------:|
+| 1.5296        | 0.8891 | 500  | 1.4462          | -0.1378        | -0.1548          | 0.5470             | 0.0170          | -1.5481        | -1.3781      | -21.5745        | -21.4875      | 1.3781   | 0.6806          |
+| 1.4761        | 1.7782 | 1000 | 1.4072          | -0.1340        | -0.1510          | 0.5350             | 0.0170          | -1.5095        | -1.3395      | -20.8914        | -20.8038      | 1.3395   | 0.6771          |
+| 1.4105        | 2.6673 | 1500 | 1.3996          | -0.1332        | -0.1501          | 0.5460             | 0.0169          | -1.5010        | -1.3319      | -20.7443        | -20.6555      | 1.3319   | 0.6769          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.40.1
+- Pytorch 2.3.0
+- Datasets 2.19.0
+- Tokenizers 0.19.1

trainer_log.jsonl

@@ -151,3 +151,22 @@
 {"current_steps": 1490, "total_steps": 1686, "loss": 1.3758, "accuracy": 0.5062500238418579, "learning_rate": 1.6490167940538343e-07, "epoch": 2.6494776617026004, "percentage": 88.37, "elapsed_time": "1:57:42", "remaining_time": "0:15:29"}
 {"current_steps": 1500, "total_steps": 1686, "loss": 1.4105, "accuracy": 0.5249999761581421, "learning_rate": 1.4866882516191339e-07, "epoch": 2.6672593909757722, "percentage": 88.97, "elapsed_time": "1:58:30", "remaining_time": "0:14:41"}
 {"current_steps": 1500, "total_steps": 1686, "eval_loss": 1.3996269702911377, "epoch": 2.6672593909757722, "percentage": 88.97, "elapsed_time": "1:59:50", "remaining_time": "0:14:51"}
+{"current_steps": 1510, "total_steps": 1686, "loss": 1.3972, "accuracy": 0.59375, "learning_rate": 1.3325243551706057e-07, "epoch": 2.685041120248944, "percentage": 89.56, "elapsed_time": "2:00:42", "remaining_time": "0:14:04"}
+{"current_steps": 1520, "total_steps": 1686, "loss": 1.4273, "accuracy": 0.518750011920929, "learning_rate": 1.1865786358165737e-07, "epoch": 2.702822849522116, "percentage": 90.15, "elapsed_time": "2:01:27", "remaining_time": "0:13:15"}
+{"current_steps": 1530, "total_steps": 1686, "loss": 1.4502, "accuracy": 0.574999988079071, "learning_rate": 1.0489017710262311e-07, "epoch": 2.720604578795288, "percentage": 90.75, "elapsed_time": "2:02:14", "remaining_time": "0:12:27"}
+{"current_steps": 1540, "total_steps": 1686, "loss": 1.4177, "accuracy": 0.512499988079071, "learning_rate": 9.195415670326446e-08, "epoch": 2.73838630806846, "percentage": 91.34, "elapsed_time": "2:03:01", "remaining_time": "0:11:39"}
+{"current_steps": 1550, "total_steps": 1686, "loss": 1.3834, "accuracy": 0.48750001192092896, "learning_rate": 7.985429422327384e-08, "epoch": 2.7561680373416317, "percentage": 91.93, "elapsed_time": "2:03:46", "remaining_time": "0:10:51"}
+{"current_steps": 1560, "total_steps": 1686, "loss": 1.5216, "accuracy": 0.518750011920929, "learning_rate": 6.859479115900818e-08, "epoch": 2.773949766614803, "percentage": 92.53, "elapsed_time": "2:04:32", "remaining_time": "0:10:03"}
+{"current_steps": 1570, "total_steps": 1686, "loss": 1.3449, "accuracy": 0.543749988079071, "learning_rate": 5.817955720457902e-08, "epoch": 2.791731495887975, "percentage": 93.12, "elapsed_time": "2:05:17", "remaining_time": "0:09:15"}
+{"current_steps": 1580, "total_steps": 1686, "loss": 1.423, "accuracy": 0.543749988079071, "learning_rate": 4.861220889427199e-08, "epoch": 2.809513225161147, "percentage": 93.71, "elapsed_time": "2:06:01", "remaining_time": "0:08:27"}
+{"current_steps": 1590, "total_steps": 1686, "loss": 1.4137, "accuracy": 0.518750011920929, "learning_rate": 3.9896068346758074e-08, "epoch": 2.827294954434319, "percentage": 94.31, "elapsed_time": "2:06:46", "remaining_time": "0:07:39"}
+{"current_steps": 1600, "total_steps": 1686, "loss": 1.4049, "accuracy": 0.4625000059604645, "learning_rate": 3.203416211153832e-08, "epoch": 2.8450766837074903, "percentage": 94.9, "elapsed_time": "2:07:32", "remaining_time": "0:06:51"}
+{"current_steps": 1610, "total_steps": 1686, "loss": 1.4637, "accuracy": 0.48750001192092896, "learning_rate": 2.5029220118019393e-08, "epoch": 2.8628584129806622, "percentage": 95.49, "elapsed_time": "2:08:19", "remaining_time": "0:06:03"}
+{"current_steps": 1620, "total_steps": 1686, "loss": 1.3231, "accuracy": 0.625, "learning_rate": 1.8883674727586122e-08, "epoch": 2.880640142253834, "percentage": 96.09, "elapsed_time": "2:09:03", "remaining_time": "0:05:15"}
+{"current_steps": 1630, "total_steps": 1686, "loss": 1.4301, "accuracy": 0.48124998807907104, "learning_rate": 1.3599659889000639e-08, "epoch": 2.898421871527006, "percentage": 96.68, "elapsed_time": "2:09:51", "remaining_time": "0:04:27"}
+{"current_steps": 1640, "total_steps": 1686, "loss": 1.3923, "accuracy": 0.4749999940395355, "learning_rate": 9.179010397421528e-09, "epoch": 2.916203600800178, "percentage": 97.27, "elapsed_time": "2:10:41", "remaining_time": "0:03:39"}
+{"current_steps": 1650, "total_steps": 1686, "loss": 1.2759, "accuracy": 0.512499988079071, "learning_rate": 5.623261257296509e-09, "epoch": 2.93398533007335, "percentage": 97.86, "elapsed_time": "2:11:27", "remaining_time": "0:02:52"}
+{"current_steps": 1660, "total_steps": 1686, "loss": 1.415, "accuracy": 0.5, "learning_rate": 2.933647149357122e-09, "epoch": 2.9517670593465217, "percentage": 98.46, "elapsed_time": "2:12:13", "remaining_time": "0:02:04"}
+{"current_steps": 1670, "total_steps": 1686, "loss": 1.3912, "accuracy": 0.48750001192092896, "learning_rate": 1.1111020018930717e-09, "epoch": 2.969548788619693, "percentage": 99.05, "elapsed_time": "2:12:59", "remaining_time": "0:01:16"}
+{"current_steps": 1680, "total_steps": 1686, "loss": 1.3388, "accuracy": 0.59375, "learning_rate": 1.5625866646051813e-10, "epoch": 2.987330517892865, "percentage": 99.64, "elapsed_time": "2:13:45", "remaining_time": "0:00:28"}
+{"current_steps": 1686, "total_steps": 1686, "epoch": 2.997999555456768, "percentage": 100.0, "elapsed_time": "2:14:15", "remaining_time": "0:00:00"}