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-SALT
+  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-SALT
+
+This model is a fine-tuned version of [google/gemma-2b-it](https://huggingface.co/google/gemma-2b-it) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.3294
+- Rewards/chosen: -0.1257
+- Rewards/rejected: -0.1400
+- Rewards/accuracies: 0.5345
+- Rewards/margins: 0.0142
+- Logps/rejected: -1.3996
+- Logps/chosen: -1.2573
+- Logits/rejected: -20.6959
+- Logits/chosen: -20.6644
+- Sft Loss: 1.2573
+- Odds Ratio Loss: 0.7207
+
+## 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 |
+|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|:--------:|:---------------:|
+| 1.4057        | 0.8082 | 500  | 1.3871          | -0.1315        | -0.1452          | 0.5318             | 0.0137          | -1.4522        | -1.3148      | -21.9402        | -21.9144      | 1.3148   | 0.7223          |
+| 1.2196        | 1.6165 | 1000 | 1.3397          | -0.1268        | -0.1408          | 0.5355             | 0.0140          | -1.4079        | -1.2678      | -20.9111        | -20.8823      | 1.2678   | 0.7187          |
+| 1.2883        | 2.4247 | 1500 | 1.3294          | -0.1257        | -0.1400          | 0.5345             | 0.0142          | -1.3996        | -1.2573      | -20.6959        | -20.6644      | 1.2573   | 0.7207          |
+
+
+### 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 +151,39 @@
 {"current_steps": 1490, "total_steps": 1854, "loss": 1.3968, "accuracy": 0.543749988079071, "learning_rate": 4.607082849092523e-07, "epoch": 2.40856738735098, "percentage": 80.37, "elapsed_time": "1:57:08", "remaining_time": "0:28:36"}
 {"current_steps": 1500, "total_steps": 1854, "loss": 1.2883, "accuracy": 0.5249999761581421, "learning_rate": 4.3649635614901405e-07, "epoch": 2.4247322691452817, "percentage": 80.91, "elapsed_time": "1:57:52", "remaining_time": "0:27:49"}
 {"current_steps": 1500, "total_steps": 1854, "eval_loss": 1.3293771743774414, "epoch": 2.4247322691452817, "percentage": 80.91, "elapsed_time": "1:59:15", "remaining_time": "0:28:08"}
+{"current_steps": 1510, "total_steps": 1854, "loss": 1.2928, "accuracy": 0.5375000238418579, "learning_rate": 4.128769732701973e-07, "epoch": 2.4408971509395836, "percentage": 81.45, "elapsed_time": "2:00:03", "remaining_time": "0:27:21"}
+{"current_steps": 1520, "total_steps": 1854, "loss": 1.3339, "accuracy": 0.5062500238418579, "learning_rate": 3.8985691870233046e-07, "epoch": 2.4570620327338855, "percentage": 81.98, "elapsed_time": "2:00:47", "remaining_time": "0:26:32"}
+{"current_steps": 1530, "total_steps": 1854, "loss": 1.3289, "accuracy": 0.518750011920929, "learning_rate": 3.6744280277467904e-07, "epoch": 2.4732269145281873, "percentage": 82.52, "elapsed_time": "2:01:30", "remaining_time": "0:25:43"}
+{"current_steps": 1540, "total_steps": 1854, "loss": 1.2195, "accuracy": 0.5874999761581421, "learning_rate": 3.456410618180503e-07, "epoch": 2.489391796322489, "percentage": 83.06, "elapsed_time": "2:02:14", "remaining_time": "0:24:55"}
+{"current_steps": 1550, "total_steps": 1854, "loss": 1.2861, "accuracy": 0.574999988079071, "learning_rate": 3.244579563165753e-07, "epoch": 2.5055566781167915, "percentage": 83.6, "elapsed_time": "2:02:57", "remaining_time": "0:24:07"}
+{"current_steps": 1560, "total_steps": 1854, "loss": 1.36, "accuracy": 0.5249999761581421, "learning_rate": 3.038995691099697e-07, "epoch": 2.521721559911093, "percentage": 84.14, "elapsed_time": "2:03:39", "remaining_time": "0:23:18"}
+{"current_steps": 1570, "total_steps": 1854, "loss": 1.3879, "accuracy": 0.5562499761581421, "learning_rate": 2.839718036468192e-07, "epoch": 2.5378864417053952, "percentage": 84.68, "elapsed_time": "2:04:24", "remaining_time": "0:22:30"}
+{"current_steps": 1580, "total_steps": 1854, "loss": 1.3346, "accuracy": 0.5, "learning_rate": 2.646803822893723e-07, "epoch": 2.5540513234996967, "percentage": 85.22, "elapsed_time": "2:05:09", "remaining_time": "0:21:42"}
+{"current_steps": 1590, "total_steps": 1854, "loss": 1.293, "accuracy": 0.574999988079071, "learning_rate": 2.460308446703341e-07, "epoch": 2.570216205293999, "percentage": 85.76, "elapsed_time": "2:05:53", "remaining_time": "0:20:54"}
+{"current_steps": 1600, "total_steps": 1854, "loss": 1.2481, "accuracy": 0.612500011920929, "learning_rate": 2.2802854610213143e-07, "epoch": 2.5863810870883004, "percentage": 86.3, "elapsed_time": "2:06:36", "remaining_time": "0:20:05"}
+{"current_steps": 1610, "total_steps": 1854, "loss": 1.3503, "accuracy": 0.581250011920929, "learning_rate": 2.106786560391072e-07, "epoch": 2.6025459688826027, "percentage": 86.84, "elapsed_time": "2:07:19", "remaining_time": "0:19:17"}
+{"current_steps": 1620, "total_steps": 1854, "loss": 1.2827, "accuracy": 0.53125, "learning_rate": 1.9398615659308255e-07, "epoch": 2.6187108506769046, "percentage": 87.38, "elapsed_time": "2:08:03", "remaining_time": "0:18:29"}
+{"current_steps": 1630, "total_steps": 1854, "loss": 1.3318, "accuracy": 0.543749988079071, "learning_rate": 1.7795584110272184e-07, "epoch": 2.6348757324712064, "percentage": 87.92, "elapsed_time": "2:08:50", "remaining_time": "0:17:42"}
+{"current_steps": 1640, "total_steps": 1854, "loss": 1.3008, "accuracy": 0.5249999761581421, "learning_rate": 1.6259231275709636e-07, "epoch": 2.6510406142655083, "percentage": 88.46, "elapsed_time": "2:09:36", "remaining_time": "0:16:54"}
+{"current_steps": 1650, "total_steps": 1854, "loss": 1.276, "accuracy": 0.5687500238418579, "learning_rate": 1.478999832738548e-07, "epoch": 2.66720549605981, "percentage": 89.0, "elapsed_time": "2:10:18", "remaining_time": "0:16:06"}
+{"current_steps": 1660, "total_steps": 1854, "loss": 1.2979, "accuracy": 0.5062500238418579, "learning_rate": 1.338830716323769e-07, "epoch": 2.683370377854112, "percentage": 89.54, "elapsed_time": "2:11:04", "remaining_time": "0:15:19"}
+{"current_steps": 1670, "total_steps": 1854, "loss": 1.2305, "accuracy": 0.543749988079071, "learning_rate": 1.205456028622723e-07, "epoch": 2.699535259648414, "percentage": 90.08, "elapsed_time": "2:11:46", "remaining_time": "0:14:31"}
+{"current_steps": 1680, "total_steps": 1854, "loss": 1.2805, "accuracy": 0.625, "learning_rate": 1.0789140688756805e-07, "epoch": 2.7157001414427158, "percentage": 90.61, "elapsed_time": "2:12:30", "remaining_time": "0:13:43"}
+{"current_steps": 1690, "total_steps": 1854, "loss": 1.3583, "accuracy": 0.4937500059604645, "learning_rate": 9.592411742693098e-08, "epoch": 2.7318650232370176, "percentage": 91.15, "elapsed_time": "2:13:12", "remaining_time": "0:12:55"}
+{"current_steps": 1700, "total_steps": 1854, "loss": 1.2587, "accuracy": 0.606249988079071, "learning_rate": 8.464717095022168e-08, "epoch": 2.7480299050313195, "percentage": 91.69, "elapsed_time": "2:13:54", "remaining_time": "0:12:07"}
+{"current_steps": 1710, "total_steps": 1854, "loss": 1.3366, "accuracy": 0.5, "learning_rate": 7.406380569169841e-08, "epoch": 2.7641947868256214, "percentage": 92.23, "elapsed_time": "2:14:38", "remaining_time": "0:11:20"}
+{"current_steps": 1720, "total_steps": 1854, "loss": 1.2649, "accuracy": 0.550000011920929, "learning_rate": 6.417706072013808e-08, "epoch": 2.7803596686199232, "percentage": 92.77, "elapsed_time": "2:15:23", "remaining_time": "0:10:32"}
+{"current_steps": 1730, "total_steps": 1854, "loss": 1.3517, "accuracy": 0.5687500238418579, "learning_rate": 5.498977506615294e-08, "epoch": 2.796524550414225, "percentage": 93.31, "elapsed_time": "2:16:09", "remaining_time": "0:09:45"}
+{"current_steps": 1740, "total_steps": 1854, "loss": 1.3367, "accuracy": 0.5625, "learning_rate": 4.6504586906947756e-08, "epoch": 2.812689432208527, "percentage": 93.85, "elapsed_time": "2:16:55", "remaining_time": "0:08:58"}
+{"current_steps": 1750, "total_steps": 1854, "loss": 1.4126, "accuracy": 0.5, "learning_rate": 3.8723932808754914e-08, "epoch": 2.828854314002829, "percentage": 94.39, "elapsed_time": "2:17:41", "remaining_time": "0:08:10"}
+{"current_steps": 1760, "total_steps": 1854, "loss": 1.2905, "accuracy": 0.5375000238418579, "learning_rate": 3.1650047027158014e-08, "epoch": 2.8450191957971307, "percentage": 94.93, "elapsed_time": "2:18:27", "remaining_time": "0:07:23"}
+{"current_steps": 1770, "total_steps": 1854, "loss": 1.1906, "accuracy": 0.574999988079071, "learning_rate": 2.5284960865517848e-08, "epoch": 2.8611840775914326, "percentage": 95.47, "elapsed_time": "2:19:13", "remaining_time": "0:06:36"}
+{"current_steps": 1780, "total_steps": 1854, "loss": 1.3062, "accuracy": 0.6000000238418579, "learning_rate": 1.9630502091670388e-08, "epoch": 2.8773489593857344, "percentage": 96.01, "elapsed_time": "2:19:56", "remaining_time": "0:05:49"}
+{"current_steps": 1790, "total_steps": 1854, "loss": 1.193, "accuracy": 0.5249999761581421, "learning_rate": 1.4688294413074677e-08, "epoch": 2.8935138411800363, "percentage": 96.55, "elapsed_time": "2:20:39", "remaining_time": "0:05:01"}
+{"current_steps": 1800, "total_steps": 1854, "loss": 1.2909, "accuracy": 0.4749999940395355, "learning_rate": 1.0459757010556626e-08, "epoch": 2.909678722974338, "percentage": 97.09, "elapsed_time": "2:21:22", "remaining_time": "0:04:14"}
+{"current_steps": 1810, "total_steps": 1854, "loss": 1.3451, "accuracy": 0.512499988079071, "learning_rate": 6.94610413078306e-09, "epoch": 2.92584360476864, "percentage": 97.63, "elapsed_time": "2:22:05", "remaining_time": "0:03:27"}
+{"current_steps": 1820, "total_steps": 1854, "loss": 1.2352, "accuracy": 0.59375, "learning_rate": 4.14834473758563e-09, "epoch": 2.942008486562942, "percentage": 98.17, "elapsed_time": "2:22:49", "remaining_time": "0:02:40"}
+{"current_steps": 1830, "total_steps": 1854, "loss": 1.2618, "accuracy": 0.606249988079071, "learning_rate": 2.067282222230349e-09, "epoch": 2.9581733683572438, "percentage": 98.71, "elapsed_time": "2:23:37", "remaining_time": "0:01:53"}
+{"current_steps": 1840, "total_steps": 1854, "loss": 1.1819, "accuracy": 0.550000011920929, "learning_rate": 7.035141727212979e-10, "epoch": 2.9743382501515456, "percentage": 99.24, "elapsed_time": "2:24:18", "remaining_time": "0:01:05"}
+{"current_steps": 1850, "total_steps": 1854, "loss": 1.3642, "accuracy": 0.543749988079071, "learning_rate": 5.743220219761592e-11, "epoch": 2.9905031319458475, "percentage": 99.78, "elapsed_time": "2:25:03", "remaining_time": "0:00:18"}
+{"current_steps": 1854, "total_steps": 1854, "epoch": 2.9969690846635686, "percentage": 100.0, "elapsed_time": "2:25:22", "remaining_time": "0:00:00"}