bge-small-en-v1.5-quality
This model is a fine-tuned version of BAAI/bge-small-en-v1.5 on the agentlans/prompt-quality dataset. It achieves the following results on the evaluation set:
- Loss: 1.3925
- Mse: 1.3925
Examples
| Prompt | Quality |
|---|---|
| How can make the branches of a Cactus grow leaves? | -2.891 |
| Think of something creative to do with a banana. | -3.078 |
| List top 5 differences between men and women | -1.242 |
| Name the popular tourist attraction in a given city. Copenhagen | -0.9883 |
| how can an MBA applied to the field of biotech? | -1.625 |
| How can I configure i8N for reports developed in TIBCO Jaspersoft server | -0.832 |
| what are the criteria to identify who to interview when doing market research | -1.047 |
| Do earthquake warning systems take into account potential volcanic triggers for tsunamis? | -0.4102 |
| Make a list of items required to create a Harry Potter themed party. | -0.3574 |
| Write a tongue twister about a quirky quail in the style of Shel Silverstein. | 2.516 |
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 20.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Mse |
|---|---|---|---|---|
| 1.7064 | 1.0 | 9976 | 1.5169 | 1.5169 |
| 1.3945 | 2.0 | 19952 | 1.5205 | 1.5205 |
| 1.0004 | 3.0 | 29928 | 1.3925 | 1.3925 |
| 0.8248 | 4.0 | 39904 | 1.5392 | 1.5392 |
| 0.6471 | 5.0 | 49880 | 1.5778 | 1.5778 |
| 0.5265 | 6.0 | 59856 | 1.6023 | 1.6023 |
| 0.4439 | 7.0 | 69832 | 1.6765 | 1.6765 |
| 0.36 | 8.0 | 79808 | 1.6086 | 1.6086 |
| 0.3072 | 9.0 | 89784 | 1.5445 | 1.5445 |
| 0.2535 | 10.0 | 99760 | 1.6000 | 1.6000 |
| 0.209 | 11.0 | 109736 | 1.5538 | 1.5538 |
| 0.1827 | 12.0 | 119712 | 1.5323 | 1.5323 |
| 0.1614 | 13.0 | 129688 | 1.5196 | 1.5196 |
| 0.133 | 14.0 | 139664 | 1.4935 | 1.4935 |
| 0.127 | 15.0 | 149640 | 1.5164 | 1.5164 |
| 0.1076 | 16.0 | 159616 | 1.5022 | 1.5022 |
| 0.0926 | 17.0 | 169592 | 1.4562 | 1.4562 |
| 0.0835 | 18.0 | 179568 | 1.4688 | 1.4688 |
| 0.0775 | 19.0 | 189544 | 1.4841 | 1.4841 |
| 0.0769 | 20.0 | 199520 | 1.4606 | 1.4606 |
Framework versions
- Transformers 5.0.0.dev0
- Pytorch 2.9.1+cu128
- Datasets 4.4.1
- Tokenizers 0.22.1
License
This fine-tuned model is released under the MIT License.
This model is based on BAAI/bge-small-en-v1.5, which is also licensed under the MIT License. The copyright notice and license for the base model are preserved in LICENSE.
Copyright (c) 2026 @agentlans
Commercial use, modification, distribution, and private use are permitted under the terms of the MIT License.