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-7b-it
+model-index:
+- name: Gemma-7B-It-ORPO-SALT-HALF
+  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-7B-It-ORPO-SALT-HALF
+
+This model is a fine-tuned version of [google/gemma-7b-it](https://huggingface.co/google/gemma-7b-it) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.3159
+- Rewards/chosen: -0.1249
+- Rewards/rejected: -0.1471
+- Rewards/accuracies: 0.5619
+- Rewards/margins: 0.0222
+- Logps/rejected: -1.4709
+- Logps/chosen: -1.2488
+- Logits/rejected: 253.8645
+- Logits/chosen: 253.5439
+- Sft Loss: 1.2488
+- Odds Ratio Loss: 0.6713
+
+## 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 |
+|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|:--------:|:---------------:|
+| 1.422         | 0.8467 | 500  | 1.3896          | -0.1322        | -0.1546          | 0.5752             | 0.0224          | -1.5459        | -1.3222      | 250.5634        | 250.2739      | 1.3222   | 0.6733          |
+| 1.3103        | 1.6935 | 1000 | 1.3313          | -0.1264        | -0.1489          | 0.5695             | 0.0224          | -1.4886        | -1.2642      | 253.1350        | 252.8147      | 1.2642   | 0.6718          |
+| 1.2057        | 2.5402 | 1500 | 1.3159          | -0.1249        | -0.1471          | 0.5619             | 0.0222          | -1.4709        | -1.2488      | 253.8645        | 253.5439      | 1.2488   | 0.6713          |
+
+
+### 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 +151,31 @@
 {"current_steps": 1490, "total_steps": 1770, "loss": 1.2608, "accuracy": 0.574999988079071, "learning_rate": 3.024615823368371e-07, "epoch": 2.523285351397121, "percentage": 84.18, "elapsed_time": "5:27:34", "remaining_time": "1:01:33"}
 {"current_steps": 1500, "total_steps": 1770, "loss": 1.2057, "accuracy": 0.5625, "learning_rate": 2.8165102503600716e-07, "epoch": 2.5402201524132093, "percentage": 84.75, "elapsed_time": "5:29:42", "remaining_time": "0:59:20"}
 {"current_steps": 1500, "total_steps": 1770, "eval_loss": 1.3159173727035522, "epoch": 2.5402201524132093, "percentage": 84.75, "elapsed_time": "5:33:24", "remaining_time": "1:00:00"}
+{"current_steps": 1510, "total_steps": 1770, "loss": 1.3937, "accuracy": 0.543749988079071, "learning_rate": 2.615393769259039e-07, "epoch": 2.557154953429297, "percentage": 85.31, "elapsed_time": "5:35:40", "remaining_time": "0:57:47"}
+{"current_steps": 1520, "total_steps": 1770, "loss": 1.2334, "accuracy": 0.543749988079071, "learning_rate": 2.421329743475917e-07, "epoch": 2.574089754445385, "percentage": 85.88, "elapsed_time": "5:37:46", "remaining_time": "0:55:33"}
+{"current_steps": 1530, "total_steps": 1770, "loss": 1.3281, "accuracy": 0.5375000238418579, "learning_rate": 2.234379314486973e-07, "epoch": 2.5910245554614733, "percentage": 86.44, "elapsed_time": "5:39:57", "remaining_time": "0:53:19"}
+{"current_steps": 1540, "total_steps": 1770, "loss": 1.2628, "accuracy": 0.6000000238418579, "learning_rate": 2.0546013825709783e-07, "epoch": 2.6079593564775614, "percentage": 87.01, "elapsed_time": "5:42:10", "remaining_time": "0:51:06"}
+{"current_steps": 1550, "total_steps": 1770, "loss": 1.1414, "accuracy": 0.6187499761581421, "learning_rate": 1.88205258825217e-07, "epoch": 2.6248941574936495, "percentage": 87.57, "elapsed_time": "5:44:24", "remaining_time": "0:48:53"}
+{"current_steps": 1560, "total_steps": 1770, "loss": 1.2452, "accuracy": 0.550000011920929, "learning_rate": 1.7167872944552245e-07, "epoch": 2.6418289585097376, "percentage": 88.14, "elapsed_time": "5:46:29", "remaining_time": "0:46:38"}
+{"current_steps": 1570, "total_steps": 1770, "loss": 1.2281, "accuracy": 0.5, "learning_rate": 1.5588575693777142e-07, "epoch": 2.6587637595258258, "percentage": 88.7, "elapsed_time": "5:48:43", "remaining_time": "0:44:25"}
+{"current_steps": 1580, "total_steps": 1770, "loss": 1.3435, "accuracy": 0.5062500238418579, "learning_rate": 1.4083131700856428e-07, "epoch": 2.675698560541914, "percentage": 89.27, "elapsed_time": "5:50:58", "remaining_time": "0:42:12"}
+{"current_steps": 1590, "total_steps": 1770, "loss": 1.2371, "accuracy": 0.606249988079071, "learning_rate": 1.2652015268370315e-07, "epoch": 2.6926333615580016, "percentage": 89.83, "elapsed_time": "5:53:06", "remaining_time": "0:39:58"}
+{"current_steps": 1600, "total_steps": 1770, "loss": 1.3777, "accuracy": 0.581250011920929, "learning_rate": 1.1295677281386502e-07, "epoch": 2.7095681625740897, "percentage": 90.4, "elapsed_time": "5:55:21", "remaining_time": "0:37:45"}
+{"current_steps": 1610, "total_steps": 1770, "loss": 1.323, "accuracy": 0.574999988079071, "learning_rate": 1.0014545065404973e-07, "epoch": 2.726502963590178, "percentage": 90.96, "elapsed_time": "5:57:31", "remaining_time": "0:35:31"}
+{"current_steps": 1620, "total_steps": 1770, "loss": 1.285, "accuracy": 0.574999988079071, "learning_rate": 8.809022251725502e-08, "epoch": 2.743437764606266, "percentage": 91.53, "elapsed_time": "5:59:42", "remaining_time": "0:33:18"}
+{"current_steps": 1630, "total_steps": 1770, "loss": 1.3117, "accuracy": 0.5375000238418579, "learning_rate": 7.679488650280509e-08, "epoch": 2.7603725656223537, "percentage": 92.09, "elapsed_time": "6:01:57", "remaining_time": "0:31:05"}
+{"current_steps": 1640, "total_steps": 1770, "loss": 1.2449, "accuracy": 0.5375000238418579, "learning_rate": 6.626300129972563e-08, "epoch": 2.777307366638442, "percentage": 92.66, "elapsed_time": "6:03:59", "remaining_time": "0:28:51"}
+{"current_steps": 1650, "total_steps": 1770, "loss": 1.2878, "accuracy": 0.6000000238418579, "learning_rate": 5.649788506555065e-08, "epoch": 2.79424216765453, "percentage": 93.22, "elapsed_time": "6:06:05", "remaining_time": "0:26:37"}
+{"current_steps": 1660, "total_steps": 1770, "loss": 1.2631, "accuracy": 0.581250011920929, "learning_rate": 4.7502614380908474e-08, "epoch": 2.811176968670618, "percentage": 93.79, "elapsed_time": "6:08:16", "remaining_time": "0:24:24"}
+{"current_steps": 1670, "total_steps": 1770, "loss": 1.2531, "accuracy": 0.550000011920929, "learning_rate": 3.9280023280222066e-08, "epoch": 2.828111769686706, "percentage": 94.35, "elapsed_time": "6:10:25", "remaining_time": "0:22:10"}
+{"current_steps": 1680, "total_steps": 1770, "loss": 1.288, "accuracy": 0.5562499761581421, "learning_rate": 3.1832702358818855e-08, "epoch": 2.8450465707027943, "percentage": 94.92, "elapsed_time": "6:12:45", "remaining_time": "0:19:58"}
+{"current_steps": 1690, "total_steps": 1770, "loss": 1.2743, "accuracy": 0.5687500238418579, "learning_rate": 2.5162997956746647e-08, "epoch": 2.8619813717188824, "percentage": 95.48, "elapsed_time": "6:15:05", "remaining_time": "0:17:45"}
+{"current_steps": 1700, "total_steps": 1770, "loss": 1.3412, "accuracy": 0.5874999761581421, "learning_rate": 1.9273011419536914e-08, "epoch": 2.8789161727349706, "percentage": 96.05, "elapsed_time": "6:17:00", "remaining_time": "0:15:31"}
+{"current_steps": 1710, "total_steps": 1770, "loss": 1.2679, "accuracy": 0.5562499761581421, "learning_rate": 1.4164598436159083e-08, "epoch": 2.8958509737510583, "percentage": 96.61, "elapsed_time": "6:19:03", "remaining_time": "0:13:18"}
+{"current_steps": 1720, "total_steps": 1770, "loss": 1.213, "accuracy": 0.606249988079071, "learning_rate": 9.839368454371556e-09, "epoch": 2.9127857747671464, "percentage": 97.18, "elapsed_time": "6:21:14", "remaining_time": "0:11:04"}
+{"current_steps": 1730, "total_steps": 1770, "loss": 1.2129, "accuracy": 0.5874999761581421, "learning_rate": 6.298684173650649e-09, "epoch": 2.9297205757832345, "percentage": 97.74, "elapsed_time": "6:23:20", "remaining_time": "0:08:51"}
+{"current_steps": 1740, "total_steps": 1770, "loss": 1.2883, "accuracy": 0.5625, "learning_rate": 3.543661115860686e-09, "epoch": 2.9466553767993227, "percentage": 98.31, "elapsed_time": "6:25:17", "remaining_time": "0:06:38"}
+{"current_steps": 1750, "total_steps": 1770, "loss": 1.249, "accuracy": 0.574999988079071, "learning_rate": 1.575167273800693e-09, "epoch": 2.963590177815411, "percentage": 98.87, "elapsed_time": "6:27:15", "remaining_time": "0:04:25"}
+{"current_steps": 1760, "total_steps": 1770, "loss": 1.3394, "accuracy": 0.574999988079071, "learning_rate": 3.9382283773564676e-10, "epoch": 2.9805249788314985, "percentage": 99.44, "elapsed_time": "6:29:28", "remaining_time": "0:02:12"}
+{"current_steps": 1770, "total_steps": 1770, "loss": 1.3966, "accuracy": 0.606249988079071, "learning_rate": 0.0, "epoch": 2.9974597798475866, "percentage": 100.0, "elapsed_time": "6:31:33", "remaining_time": "0:00:00"}
+{"current_steps": 1770, "total_steps": 1770, "epoch": 2.9974597798475866, "percentage": 100.0, "elapsed_time": "6:31:33", "remaining_time": "0:00:00"}