NT multi-species with gut_1024
Collection
Here are collected alle the Nucleotide Transformers multi-species models, developed by InstaDeep, and finetuned with my dataset gut_1024
•
5 items
•
Updated
gut_1024b-finetuned-lora-v2-100m-multi-species
This model is a fine-tuned version of InstaDeepAI/nucleotide-transformer-v2-100m-multi-species on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5193
- F1: 0.8251
- Matthews Correlation: 0.5308
- Accuracy: 0.7779
- F1 Score: 0.8251
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: 0.0005
- train_batch_size: 8
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 1000
Training results
| Training Loss | Epoch | Step | Validation Loss | F1 | Matthews Correlation | Accuracy | F1 Score |
|---|---|---|---|---|---|---|---|
| 0.7015 | 0.02 | 100 | 0.6761 | 0.7478 | 0.0 | 0.5971 | 0.7478 |
| 0.7134 | 0.04 | 200 | 0.6272 | 0.7788 | 0.3129 | 0.6744 | 0.7788 |
| 0.6992 | 0.05 | 300 | 0.5981 | 0.7736 | 0.3943 | 0.7145 | 0.7736 |
| 0.6285 | 0.07 | 400 | 0.6951 | 0.7336 | 0.3658 | 0.6905 | 0.7336 |
| 0.6476 | 0.09 | 500 | 0.6052 | 0.7631 | 0.3673 | 0.7019 | 0.7631 |
| 0.5939 | 0.11 | 600 | 0.5809 | 0.7720 | 0.4050 | 0.7179 | 0.7720 |
| 0.5497 | 0.12 | 700 | 0.5403 | 0.8133 | 0.4867 | 0.7576 | 0.8133 |
| 0.5188 | 0.14 | 800 | 0.5580 | 0.7846 | 0.4694 | 0.7441 | 0.7846 |
| 0.546 | 0.16 | 900 | 0.5767 | 0.8041 | 0.4243 | 0.7154 | 0.8041 |
| 0.5705 | 0.18 | 1000 | 0.5193 | 0.8251 | 0.5308 | 0.7779 | 0.8251 |
Framework versions
- Transformers 4.38.1
- Pytorch 2.1.0+cu121
- Datasets 2.17.1
- Tokenizers 0.15.2
- Downloads last month
- 4
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.