amazonPolarity_BERT_5E
This model is a fine-tuned version of bert-base-cased on the amazon_polarity dataset. It achieves the following results on the evaluation set:
- Loss: 0.4402
- Accuracy: 0.9067
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: 1e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
0.7011 | 0.03 | 50 | 0.6199 | 0.7 |
0.6238 | 0.05 | 100 | 0.4710 | 0.8133 |
0.4478 | 0.08 | 150 | 0.3249 | 0.8733 |
0.3646 | 0.11 | 200 | 0.3044 | 0.86 |
0.3244 | 0.13 | 250 | 0.2548 | 0.86 |
0.2734 | 0.16 | 300 | 0.2666 | 0.88 |
0.2784 | 0.19 | 350 | 0.2416 | 0.88 |
0.2706 | 0.21 | 400 | 0.2660 | 0.88 |
0.2368 | 0.24 | 450 | 0.2522 | 0.8867 |
0.2449 | 0.27 | 500 | 0.3135 | 0.88 |
0.262 | 0.29 | 550 | 0.2718 | 0.8733 |
0.2111 | 0.32 | 600 | 0.2494 | 0.8933 |
0.2459 | 0.35 | 650 | 0.2468 | 0.8867 |
0.2264 | 0.37 | 700 | 0.3049 | 0.8667 |
0.2572 | 0.4 | 750 | 0.2054 | 0.8933 |
0.1749 | 0.43 | 800 | 0.3489 | 0.86 |
0.2423 | 0.45 | 850 | 0.2142 | 0.8933 |
0.1931 | 0.48 | 900 | 0.2096 | 0.9067 |
0.2444 | 0.51 | 950 | 0.3404 | 0.8733 |
0.2666 | 0.53 | 1000 | 0.2378 | 0.9067 |
0.2311 | 0.56 | 1050 | 0.2416 | 0.9067 |
0.2269 | 0.59 | 1100 | 0.3188 | 0.8733 |
0.2143 | 0.61 | 1150 | 0.2343 | 0.9 |
0.2181 | 0.64 | 1200 | 0.2606 | 0.8667 |
0.2151 | 0.67 | 1250 | 0.1888 | 0.9133 |
0.2694 | 0.69 | 1300 | 0.3982 | 0.8467 |
0.2408 | 0.72 | 1350 | 0.1978 | 0.9067 |
0.2043 | 0.75 | 1400 | 0.2125 | 0.9 |
0.2081 | 0.77 | 1450 | 0.2680 | 0.8933 |
0.2361 | 0.8 | 1500 | 0.3723 | 0.8467 |
0.2503 | 0.83 | 1550 | 0.3427 | 0.8733 |
0.1983 | 0.85 | 1600 | 0.2525 | 0.9067 |
0.1947 | 0.88 | 1650 | 0.2427 | 0.9133 |
0.2411 | 0.91 | 1700 | 0.2448 | 0.9 |
0.2381 | 0.93 | 1750 | 0.3354 | 0.88 |
0.1852 | 0.96 | 1800 | 0.3078 | 0.8667 |
0.2427 | 0.99 | 1850 | 0.2408 | 0.9 |
0.1582 | 1.01 | 1900 | 0.2698 | 0.9133 |
0.159 | 1.04 | 1950 | 0.3383 | 0.9 |
0.1833 | 1.07 | 2000 | 0.2849 | 0.9 |
0.1257 | 1.09 | 2050 | 0.5376 | 0.8667 |
0.1513 | 1.12 | 2100 | 0.4469 | 0.88 |
0.1869 | 1.15 | 2150 | 0.3415 | 0.8933 |
0.1342 | 1.17 | 2200 | 0.3021 | 0.8867 |
0.1404 | 1.2 | 2250 | 0.3619 | 0.88 |
0.1576 | 1.23 | 2300 | 0.2815 | 0.9 |
0.1419 | 1.25 | 2350 | 0.4351 | 0.8867 |
0.1491 | 1.28 | 2400 | 0.3025 | 0.9133 |
0.1914 | 1.31 | 2450 | 0.3011 | 0.9067 |
0.1265 | 1.33 | 2500 | 0.3953 | 0.88 |
0.128 | 1.36 | 2550 | 0.2557 | 0.9333 |
0.1631 | 1.39 | 2600 | 0.2226 | 0.9333 |
0.1019 | 1.41 | 2650 | 0.3638 | 0.9133 |
0.1551 | 1.44 | 2700 | 0.3591 | 0.9 |
0.1853 | 1.47 | 2750 | 0.5005 | 0.8733 |
0.1578 | 1.49 | 2800 | 0.2662 | 0.92 |
0.1522 | 1.52 | 2850 | 0.2545 | 0.9267 |
0.1188 | 1.55 | 2900 | 0.3874 | 0.88 |
0.1638 | 1.57 | 2950 | 0.3003 | 0.92 |
0.1583 | 1.6 | 3000 | 0.2702 | 0.92 |
0.1844 | 1.63 | 3050 | 0.2183 | 0.9333 |
0.1365 | 1.65 | 3100 | 0.3322 | 0.8933 |
0.1683 | 1.68 | 3150 | 0.2069 | 0.9467 |
0.168 | 1.71 | 3200 | 0.4046 | 0.8667 |
0.1907 | 1.73 | 3250 | 0.3411 | 0.8933 |
0.1695 | 1.76 | 3300 | 0.1992 | 0.9333 |
0.1851 | 1.79 | 3350 | 0.2370 | 0.92 |
0.1302 | 1.81 | 3400 | 0.3058 | 0.9133 |
0.1353 | 1.84 | 3450 | 0.3134 | 0.9067 |
0.1428 | 1.87 | 3500 | 0.3767 | 0.8667 |
0.1642 | 1.89 | 3550 | 0.3239 | 0.8867 |
0.1319 | 1.92 | 3600 | 0.4725 | 0.86 |
0.1714 | 1.95 | 3650 | 0.3115 | 0.8867 |
0.1265 | 1.97 | 3700 | 0.3621 | 0.8867 |
0.1222 | 2.0 | 3750 | 0.3665 | 0.8933 |
0.0821 | 2.03 | 3800 | 0.2482 | 0.9133 |
0.1136 | 2.05 | 3850 | 0.3244 | 0.9 |
0.0915 | 2.08 | 3900 | 0.4745 | 0.8733 |
0.0967 | 2.11 | 3950 | 0.2346 | 0.94 |
0.0962 | 2.13 | 4000 | 0.3139 | 0.92 |
0.1001 | 2.16 | 4050 | 0.2944 | 0.9267 |
0.086 | 2.19 | 4100 | 0.5542 | 0.86 |
0.0588 | 2.21 | 4150 | 0.4377 | 0.9 |
0.1056 | 2.24 | 4200 | 0.3540 | 0.9133 |
0.0899 | 2.27 | 4250 | 0.5661 | 0.8733 |
0.0737 | 2.29 | 4300 | 0.5683 | 0.8733 |
0.1152 | 2.32 | 4350 | 0.2997 | 0.9333 |
0.0852 | 2.35 | 4400 | 0.5055 | 0.8933 |
0.1114 | 2.37 | 4450 | 0.3099 | 0.92 |
0.0821 | 2.4 | 4500 | 0.3026 | 0.9267 |
0.0698 | 2.43 | 4550 | 0.3250 | 0.92 |
0.1123 | 2.45 | 4600 | 0.3674 | 0.9 |
0.1196 | 2.48 | 4650 | 0.4539 | 0.8733 |
0.0617 | 2.51 | 4700 | 0.3446 | 0.92 |
0.0939 | 2.53 | 4750 | 0.3302 | 0.92 |
0.1114 | 2.56 | 4800 | 0.5149 | 0.8733 |
0.1154 | 2.59 | 4850 | 0.4935 | 0.8867 |
0.1495 | 2.61 | 4900 | 0.4706 | 0.8933 |
0.0858 | 2.64 | 4950 | 0.4048 | 0.9 |
0.0767 | 2.67 | 5000 | 0.3849 | 0.9133 |
0.0569 | 2.69 | 5050 | 0.5491 | 0.8867 |
0.1058 | 2.72 | 5100 | 0.5872 | 0.8733 |
0.0899 | 2.75 | 5150 | 0.3159 | 0.92 |
0.0757 | 2.77 | 5200 | 0.5861 | 0.8733 |
0.1305 | 2.8 | 5250 | 0.3633 | 0.9133 |
0.1027 | 2.83 | 5300 | 0.3972 | 0.9133 |
0.1259 | 2.85 | 5350 | 0.4197 | 0.8933 |
0.1255 | 2.88 | 5400 | 0.4583 | 0.8867 |
0.0981 | 2.91 | 5450 | 0.4657 | 0.8933 |
0.0736 | 2.93 | 5500 | 0.4036 | 0.9133 |
0.116 | 2.96 | 5550 | 0.3026 | 0.9067 |
0.0692 | 2.99 | 5600 | 0.3409 | 0.9133 |
0.0721 | 3.01 | 5650 | 0.5598 | 0.8733 |
0.052 | 3.04 | 5700 | 0.4130 | 0.9133 |
0.0661 | 3.07 | 5750 | 0.2589 | 0.9333 |
0.0667 | 3.09 | 5800 | 0.4484 | 0.9067 |
0.0599 | 3.12 | 5850 | 0.4883 | 0.9 |
0.0406 | 3.15 | 5900 | 0.4516 | 0.9067 |
0.0837 | 3.17 | 5950 | 0.3394 | 0.9267 |
0.0636 | 3.2 | 6000 | 0.4649 | 0.8867 |
0.0861 | 3.23 | 6050 | 0.5046 | 0.8933 |
0.0667 | 3.25 | 6100 | 0.3252 | 0.92 |
0.0401 | 3.28 | 6150 | 0.2771 | 0.94 |
0.0998 | 3.31 | 6200 | 0.4509 | 0.9 |
0.0209 | 3.33 | 6250 | 0.4666 | 0.8933 |
0.0747 | 3.36 | 6300 | 0.5430 | 0.8867 |
0.0678 | 3.39 | 6350 | 0.4050 | 0.9067 |
0.0685 | 3.41 | 6400 | 0.3738 | 0.92 |
0.0654 | 3.44 | 6450 | 0.4486 | 0.9 |
0.0496 | 3.47 | 6500 | 0.4386 | 0.9067 |
0.0379 | 3.49 | 6550 | 0.4547 | 0.9067 |
0.0897 | 3.52 | 6600 | 0.4197 | 0.9133 |
0.0729 | 3.55 | 6650 | 0.2855 | 0.9333 |
0.0515 | 3.57 | 6700 | 0.4459 | 0.9067 |
0.0588 | 3.6 | 6750 | 0.3627 | 0.92 |
0.0724 | 3.63 | 6800 | 0.4060 | 0.9267 |
0.0607 | 3.65 | 6850 | 0.4505 | 0.9133 |
0.0252 | 3.68 | 6900 | 0.5465 | 0.8933 |
0.0594 | 3.71 | 6950 | 0.4786 | 0.9067 |
0.0743 | 3.73 | 7000 | 0.4163 | 0.9267 |
0.0506 | 3.76 | 7050 | 0.3801 | 0.92 |
0.0548 | 3.79 | 7100 | 0.3557 | 0.9267 |
0.0932 | 3.81 | 7150 | 0.4278 | 0.9133 |
0.0643 | 3.84 | 7200 | 0.4673 | 0.9 |
0.0631 | 3.87 | 7250 | 0.3611 | 0.92 |
0.0793 | 3.89 | 7300 | 0.3956 | 0.9067 |
0.0729 | 3.92 | 7350 | 0.6630 | 0.8733 |
0.0552 | 3.95 | 7400 | 0.4259 | 0.8867 |
0.0432 | 3.97 | 7450 | 0.3615 | 0.92 |
0.0697 | 4.0 | 7500 | 0.5116 | 0.88 |
0.0463 | 4.03 | 7550 | 0.3334 | 0.94 |
0.046 | 4.05 | 7600 | 0.4704 | 0.8867 |
0.0371 | 4.08 | 7650 | 0.3323 | 0.94 |
0.0809 | 4.11 | 7700 | 0.3503 | 0.92 |
0.0285 | 4.13 | 7750 | 0.3360 | 0.92 |
0.0469 | 4.16 | 7800 | 0.3365 | 0.9333 |
0.041 | 4.19 | 7850 | 0.5726 | 0.88 |
0.0447 | 4.21 | 7900 | 0.4564 | 0.9067 |
0.0144 | 4.24 | 7950 | 0.5521 | 0.8867 |
0.0511 | 4.27 | 8000 | 0.5661 | 0.88 |
0.0481 | 4.29 | 8050 | 0.3445 | 0.94 |
0.036 | 4.32 | 8100 | 0.3247 | 0.94 |
0.0662 | 4.35 | 8150 | 0.3647 | 0.9333 |
0.051 | 4.37 | 8200 | 0.5024 | 0.9 |
0.0546 | 4.4 | 8250 | 0.4737 | 0.8933 |
0.0526 | 4.43 | 8300 | 0.4067 | 0.92 |
0.0291 | 4.45 | 8350 | 0.3862 | 0.9267 |
0.0292 | 4.48 | 8400 | 0.5101 | 0.9 |
0.0426 | 4.51 | 8450 | 0.4207 | 0.92 |
0.0771 | 4.53 | 8500 | 0.5525 | 0.8867 |
0.0668 | 4.56 | 8550 | 0.4487 | 0.9067 |
0.0585 | 4.59 | 8600 | 0.3574 | 0.9267 |
0.0375 | 4.61 | 8650 | 0.3980 | 0.92 |
0.0508 | 4.64 | 8700 | 0.4064 | 0.92 |
0.0334 | 4.67 | 8750 | 0.3031 | 0.94 |
0.0257 | 4.69 | 8800 | 0.3340 | 0.9333 |
0.0165 | 4.72 | 8850 | 0.4011 | 0.92 |
0.0553 | 4.75 | 8900 | 0.4243 | 0.9133 |
0.0597 | 4.77 | 8950 | 0.3685 | 0.9267 |
0.0407 | 4.8 | 9000 | 0.4262 | 0.9133 |
0.032 | 4.83 | 9050 | 0.4080 | 0.9133 |
0.0573 | 4.85 | 9100 | 0.4416 | 0.9133 |
0.0308 | 4.88 | 9150 | 0.4397 | 0.9133 |
0.0494 | 4.91 | 9200 | 0.4476 | 0.9067 |
0.015 | 4.93 | 9250 | 0.4419 | 0.9067 |
0.0443 | 4.96 | 9300 | 0.4347 | 0.9133 |
0.0479 | 4.99 | 9350 | 0.4402 | 0.9067 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.13.0
- Datasets 2.6.1
- Tokenizers 0.13.1
- Downloads last month
- 10