scenario-kd-pre-ner-full-xlmr_data-univner_full66
This model is a fine-tuned version of FacebookAI/xlm-roberta-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.4655
- Precision: 0.8114
- Recall: 0.8137
- F1: 0.8126
- Accuracy: 0.9807
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: 3e-05
- train_batch_size: 8
- eval_batch_size: 32
- seed: 66
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
Training results
Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
---|---|---|---|---|---|---|---|
1.497 | 0.2911 | 500 | 0.9072 | 0.6658 | 0.6611 | 0.6634 | 0.9685 |
0.7828 | 0.5822 | 1000 | 0.7395 | 0.7001 | 0.7482 | 0.7233 | 0.9729 |
0.6597 | 0.8732 | 1500 | 0.6711 | 0.7843 | 0.7178 | 0.7496 | 0.9753 |
0.573 | 1.1643 | 2000 | 0.6213 | 0.7522 | 0.7820 | 0.7668 | 0.9773 |
0.5148 | 1.4554 | 2500 | 0.6290 | 0.7325 | 0.7919 | 0.7610 | 0.9759 |
0.497 | 1.7465 | 3000 | 0.5801 | 0.7759 | 0.7780 | 0.7769 | 0.9778 |
0.4659 | 2.0375 | 3500 | 0.5765 | 0.7926 | 0.7764 | 0.7844 | 0.9786 |
0.4098 | 2.3286 | 4000 | 0.5585 | 0.7868 | 0.7853 | 0.7860 | 0.9789 |
0.4085 | 2.6197 | 4500 | 0.5536 | 0.7862 | 0.8042 | 0.7951 | 0.9793 |
0.3979 | 2.9108 | 5000 | 0.5326 | 0.7902 | 0.8077 | 0.7989 | 0.9798 |
0.3568 | 3.2019 | 5500 | 0.5366 | 0.7925 | 0.7922 | 0.7924 | 0.9793 |
0.3523 | 3.4929 | 6000 | 0.5277 | 0.8058 | 0.7870 | 0.7963 | 0.9792 |
0.3361 | 3.7840 | 6500 | 0.5239 | 0.7851 | 0.8159 | 0.8002 | 0.9792 |
0.3298 | 4.0751 | 7000 | 0.5126 | 0.7993 | 0.8074 | 0.8033 | 0.9800 |
0.3053 | 4.3662 | 7500 | 0.5124 | 0.8074 | 0.7961 | 0.8017 | 0.9796 |
0.3099 | 4.6573 | 8000 | 0.5019 | 0.7953 | 0.8145 | 0.8048 | 0.9799 |
0.3031 | 4.9483 | 8500 | 0.4978 | 0.8133 | 0.8009 | 0.8071 | 0.9801 |
0.2834 | 5.2394 | 9000 | 0.5067 | 0.8160 | 0.8044 | 0.8101 | 0.9804 |
0.2767 | 5.5305 | 9500 | 0.4905 | 0.8104 | 0.8096 | 0.8100 | 0.9804 |
0.2799 | 5.8216 | 10000 | 0.4812 | 0.8092 | 0.8058 | 0.8075 | 0.9804 |
0.2735 | 6.1126 | 10500 | 0.4849 | 0.8110 | 0.8104 | 0.8107 | 0.9805 |
0.261 | 6.4037 | 11000 | 0.4817 | 0.8100 | 0.8114 | 0.8107 | 0.9803 |
0.2587 | 6.6948 | 11500 | 0.4814 | 0.8127 | 0.8152 | 0.8139 | 0.9810 |
0.2593 | 6.9859 | 12000 | 0.4812 | 0.8171 | 0.8090 | 0.8130 | 0.9806 |
0.247 | 7.2770 | 12500 | 0.4816 | 0.8037 | 0.8173 | 0.8104 | 0.9807 |
0.2452 | 7.5680 | 13000 | 0.4688 | 0.8130 | 0.8117 | 0.8124 | 0.9805 |
0.2426 | 7.8591 | 13500 | 0.4700 | 0.8130 | 0.8104 | 0.8117 | 0.9806 |
0.2404 | 8.1502 | 14000 | 0.4680 | 0.8127 | 0.8175 | 0.8151 | 0.9809 |
0.2347 | 8.4413 | 14500 | 0.4723 | 0.8156 | 0.8160 | 0.8158 | 0.9810 |
0.2356 | 8.7324 | 15000 | 0.4720 | 0.8115 | 0.8186 | 0.8151 | 0.9807 |
0.2347 | 9.0234 | 15500 | 0.4634 | 0.8199 | 0.8178 | 0.8188 | 0.9813 |
0.2301 | 9.3145 | 16000 | 0.4631 | 0.8172 | 0.8158 | 0.8165 | 0.9809 |
0.2287 | 9.6056 | 16500 | 0.4621 | 0.8125 | 0.8147 | 0.8136 | 0.9808 |
0.2253 | 9.8967 | 17000 | 0.4655 | 0.8114 | 0.8137 | 0.8126 | 0.9807 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.1.1+cu121
- Datasets 2.14.5
- Tokenizers 0.19.1
- Downloads last month
- 5
Model tree for haryoaw/scenario-kd-pre-ner-full-xlmr_data-univner_full66
Base model
FacebookAI/xlm-roberta-base