scenario-kd-po-ner-full-xlmr_data-univner_full44
This model is a fine-tuned version of haryoaw/scenario-TCR-NER_data-univner_full on the None dataset. It achieves the following results on the evaluation set:
- Loss: 47.6447
- Precision: 0.8149
- Recall: 0.8310
- F1: 0.8229
- Accuracy: 0.9817
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: 44
- 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 |
---|---|---|---|---|---|---|---|
102.0525 | 0.2911 | 500 | 80.2861 | 0.7542 | 0.7806 | 0.7672 | 0.9771 |
74.8813 | 0.5822 | 1000 | 72.2140 | 0.7820 | 0.8016 | 0.7917 | 0.9792 |
68.8504 | 0.8732 | 1500 | 68.2456 | 0.7966 | 0.8015 | 0.7991 | 0.9798 |
64.7421 | 1.1643 | 2000 | 64.8443 | 0.7960 | 0.8084 | 0.8021 | 0.9801 |
61.5075 | 1.4554 | 2500 | 62.6862 | 0.8082 | 0.8088 | 0.8085 | 0.9803 |
59.5531 | 1.7465 | 3000 | 60.6463 | 0.8092 | 0.8171 | 0.8131 | 0.9807 |
57.7737 | 2.0375 | 3500 | 59.1526 | 0.8051 | 0.8212 | 0.8131 | 0.9807 |
55.8137 | 2.3286 | 4000 | 57.6582 | 0.8162 | 0.8156 | 0.8159 | 0.9811 |
54.535 | 2.6197 | 4500 | 56.4505 | 0.8116 | 0.8277 | 0.8196 | 0.9816 |
53.4845 | 2.9108 | 5000 | 55.3659 | 0.8145 | 0.8256 | 0.8200 | 0.9812 |
52.2754 | 3.2019 | 5500 | 54.6164 | 0.8185 | 0.8100 | 0.8142 | 0.9810 |
51.3189 | 3.4929 | 6000 | 53.6241 | 0.8019 | 0.8319 | 0.8167 | 0.9811 |
50.5871 | 3.7840 | 6500 | 52.9170 | 0.8191 | 0.8325 | 0.8258 | 0.9818 |
49.8884 | 4.0751 | 7000 | 52.2826 | 0.8184 | 0.8279 | 0.8231 | 0.9815 |
49.1939 | 4.3662 | 7500 | 51.7106 | 0.8129 | 0.8368 | 0.8247 | 0.9817 |
48.7488 | 4.6573 | 8000 | 51.2393 | 0.8162 | 0.8352 | 0.8256 | 0.9820 |
48.2589 | 4.9483 | 8500 | 50.8117 | 0.8196 | 0.8286 | 0.8241 | 0.9816 |
47.5894 | 5.2394 | 9000 | 50.3237 | 0.8130 | 0.8300 | 0.8214 | 0.9814 |
47.2432 | 5.5305 | 9500 | 50.0318 | 0.8211 | 0.8279 | 0.8245 | 0.9817 |
47.0358 | 5.8216 | 10000 | 49.7203 | 0.8077 | 0.8364 | 0.8218 | 0.9815 |
46.6045 | 6.1126 | 10500 | 49.4091 | 0.8207 | 0.8273 | 0.8240 | 0.9819 |
46.3028 | 6.4037 | 11000 | 49.1730 | 0.8191 | 0.8303 | 0.8247 | 0.9821 |
46.0718 | 6.6948 | 11500 | 48.9163 | 0.8261 | 0.8357 | 0.8309 | 0.9823 |
45.7463 | 6.9859 | 12000 | 48.7349 | 0.8247 | 0.8329 | 0.8288 | 0.9823 |
45.5286 | 7.2770 | 12500 | 48.4513 | 0.8184 | 0.8321 | 0.8252 | 0.9820 |
45.3206 | 7.5680 | 13000 | 48.2289 | 0.8219 | 0.8355 | 0.8286 | 0.9820 |
45.2792 | 7.8591 | 13500 | 48.2565 | 0.8258 | 0.8299 | 0.8279 | 0.9821 |
45.0271 | 8.1502 | 14000 | 47.9724 | 0.8234 | 0.8377 | 0.8305 | 0.9821 |
44.8801 | 8.4413 | 14500 | 47.8890 | 0.8183 | 0.8362 | 0.8272 | 0.9822 |
44.8699 | 8.7324 | 15000 | 47.8381 | 0.8210 | 0.8322 | 0.8265 | 0.9821 |
44.72 | 9.0234 | 15500 | 47.7280 | 0.8169 | 0.8316 | 0.8242 | 0.9818 |
44.5876 | 9.3145 | 16000 | 47.7288 | 0.8213 | 0.8305 | 0.8259 | 0.9820 |
44.538 | 9.6056 | 16500 | 47.6315 | 0.8200 | 0.8345 | 0.8272 | 0.9821 |
44.5589 | 9.8967 | 17000 | 47.6447 | 0.8149 | 0.8310 | 0.8229 | 0.9817 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.1.1+cu121
- Datasets 2.14.5
- Tokenizers 0.19.1
- Downloads last month
- 4
Model tree for haryoaw/scenario-kd-po-ner-full-xlmr_data-univner_full44
Base model
FacebookAI/xlm-roberta-base