Edit model card

scenario-NON-KD-PO-COPY-CDF-CL-D2_data-cl-cardiff_cl_only_gamma

This model is a fine-tuned version of haryoaw/scenario-TCR_data-cl-cardiff_cl_only2 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 5.6776
  • Accuracy: 0.4560
  • F1: 0.4557

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-05
  • train_batch_size: 32
  • eval_batch_size: 32
  • seed: 11423
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Accuracy F1
No log 1.09 250 1.7114 0.4745 0.4746
0.5344 2.17 500 2.4454 0.4522 0.4441
0.5344 3.26 750 2.7492 0.4699 0.4693
0.2322 4.35 1000 3.0156 0.4336 0.4123
0.2322 5.43 1250 2.9947 0.4799 0.4819
0.1279 6.52 1500 3.7528 0.4506 0.4474
0.1279 7.61 1750 3.5806 0.4560 0.4560
0.0973 8.7 2000 3.8704 0.4545 0.4510
0.0973 9.78 2250 4.0113 0.4498 0.4450
0.0683 10.87 2500 3.5910 0.4676 0.4690
0.0683 11.96 2750 4.3665 0.4506 0.4480
0.0498 13.04 3000 4.2188 0.4799 0.4809
0.0498 14.13 3250 4.8550 0.4398 0.4315
0.0324 15.22 3500 4.7206 0.4630 0.4586
0.0324 16.3 3750 5.0101 0.4444 0.4392
0.025 17.39 4000 5.0926 0.4660 0.4667
0.025 18.48 4250 4.8167 0.4869 0.4886
0.0248 19.57 4500 5.1087 0.4606 0.4610
0.0248 20.65 4750 5.4661 0.4429 0.4365
0.0154 21.74 5000 5.2199 0.4730 0.4742
0.0154 22.83 5250 5.1442 0.4807 0.4825
0.0112 23.91 5500 5.4588 0.4691 0.4696
0.0112 25.0 5750 5.4026 0.4653 0.4654
0.0112 26.09 6000 5.6362 0.4506 0.4466
0.0112 27.17 6250 5.5405 0.4614 0.4620
0.0046 28.26 6500 5.6038 0.4591 0.4590
0.0046 29.35 6750 5.6776 0.4560 0.4557

Framework versions

  • Transformers 4.33.3
  • Pytorch 2.1.1+cu121
  • Datasets 2.14.5
  • Tokenizers 0.13.3
Downloads last month
4

Finetuned from