vibhorag101's picture
update model card README.md
43eb21b
|
raw
history blame
11.9 kB
metadata
license: mit
base_model: roberta-base
tags:
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: roberta-base-emotion-prediction-phr-2
    results: []

roberta-base-emotion-prediction-phr-2

This model is a fine-tuned version of roberta-base on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3524
  • Accuracy: 0.2522
  • Micro Precision: 0.7220
  • Micro Recall: 0.6522
  • Micro F1: 0.6853
  • Micro Roc Auc: 0.7908

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: 2e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • 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 Accuracy Micro Precision Micro Recall Micro F1 Micro Roc Auc
0.4952 0.12 100 0.4515 0.1574 0.5861 0.3505 0.4386 0.6404
0.4152 0.23 200 0.3839 0.2041 0.7102 0.4593 0.5578 0.7033
0.3878 0.35 300 0.3625 0.2341 0.7384 0.5198 0.6101 0.7340
0.3764 0.47 400 0.3506 0.2412 0.7666 0.5092 0.6119 0.7328
0.372 0.58 500 0.3450 0.2375 0.7686 0.5251 0.6239 0.7403
0.3588 0.7 600 0.3464 0.2249 0.7804 0.4964 0.6068 0.7286
0.3383 0.82 700 0.3471 0.2470 0.7503 0.5578 0.6398 0.7528
0.3489 0.94 800 0.3284 0.2620 0.7702 0.5682 0.6539 0.7603
0.3287 1.05 900 0.3214 0.2820 0.7707 0.5936 0.6706 0.7720
0.3158 1.17 1000 0.3352 0.2657 0.7580 0.5814 0.6580 0.7646
0.3247 1.29 1100 0.3219 0.2811 0.7696 0.6031 0.6763 0.7762
0.3159 1.4 1200 0.3237 0.2688 0.7479 0.6138 0.6743 0.7778
0.3207 1.52 1300 0.3217 0.2461 0.7676 0.5767 0.6586 0.7638
0.3087 1.64 1400 0.3253 0.2424 0.7484 0.5883 0.6587 0.7663
0.3057 1.75 1500 0.3174 0.2728 0.7587 0.6116 0.6773 0.7785
0.3099 1.87 1600 0.3150 0.2774 0.7683 0.6001 0.6738 0.7746
0.3006 1.99 1700 0.3176 0.2633 0.7636 0.5881 0.6645 0.7685
0.285 2.11 1800 0.3177 0.2722 0.7363 0.6484 0.6896 0.7915
0.2886 2.22 1900 0.3156 0.2768 0.7734 0.5935 0.6716 0.7723
0.2785 2.34 2000 0.3101 0.2808 0.7692 0.6151 0.6836 0.7816
0.2801 2.46 2100 0.3121 0.2728 0.7739 0.5956 0.6732 0.7734
0.2876 2.57 2200 0.3166 0.2777 0.7577 0.6157 0.6794 0.7802
0.2769 2.69 2300 0.3143 0.2881 0.7691 0.6124 0.6819 0.7803
0.2755 2.81 2400 0.3133 0.2792 0.7577 0.6263 0.6857 0.7850
0.2815 2.92 2500 0.3197 0.2716 0.7406 0.6466 0.6904 0.7914
0.2671 3.04 2600 0.3133 0.2857 0.7549 0.6438 0.6949 0.7925
0.2431 3.16 2700 0.3225 0.2722 0.7515 0.6320 0.6866 0.7866
0.2512 3.27 2800 0.3221 0.2743 0.7616 0.6106 0.6778 0.7784
0.2574 3.39 2900 0.3191 0.2737 0.7561 0.6214 0.6822 0.7825
0.2527 3.51 3000 0.3207 0.2666 0.7443 0.6315 0.6833 0.7852
0.2615 3.63 3100 0.3170 0.2670 0.7443 0.6471 0.6923 0.7923
0.2583 3.74 3200 0.3122 0.2685 0.7729 0.6068 0.6799 0.7783
0.2543 3.86 3300 0.3175 0.2709 0.7492 0.6432 0.6921 0.7913
0.2546 3.98 3400 0.3164 0.2752 0.7661 0.6186 0.6845 0.7828
0.2274 4.09 3500 0.3172 0.2759 0.7437 0.6426 0.6895 0.7902
0.2328 4.21 3600 0.3214 0.2737 0.7548 0.6297 0.6866 0.7861
0.2354 4.33 3700 0.3192 0.2792 0.7546 0.6310 0.6872 0.7866
0.2238 4.44 3800 0.3199 0.2709 0.7453 0.6444 0.6912 0.7912
0.2376 4.56 3900 0.3176 0.2734 0.7599 0.6247 0.6857 0.7846
0.2344 4.68 4000 0.3189 0.2639 0.7437 0.6390 0.6874 0.7885
0.2222 4.8 4100 0.3222 0.2636 0.7436 0.6409 0.6884 0.7894
0.232 4.91 4200 0.3227 0.2725 0.7472 0.6426 0.6910 0.7907
0.2367 5.03 4300 0.3243 0.2670 0.7463 0.6339 0.6855 0.7866
0.2154 5.15 4400 0.3257 0.2593 0.7366 0.6513 0.6913 0.7929
0.2089 5.26 4500 0.3261 0.2700 0.7416 0.6453 0.6901 0.7910
0.2081 5.38 4600 0.3269 0.2731 0.7602 0.6133 0.6789 0.7794
0.2116 5.5 4700 0.3308 0.2593 0.7229 0.6687 0.6947 0.7983
0.2128 5.61 4800 0.3263 0.2660 0.7422 0.6432 0.6891 0.7902
0.2059 5.73 4900 0.3295 0.2728 0.7356 0.6550 0.6929 0.7944
0.2103 5.85 5000 0.3301 0.2814 0.7442 0.6510 0.6945 0.7940
0.2151 5.96 5100 0.3300 0.2541 0.7221 0.6598 0.6896 0.7942
0.1954 6.08 5200 0.3325 0.2765 0.7476 0.6381 0.6885 0.7887
0.2028 6.2 5300 0.3316 0.2559 0.7364 0.6400 0.6848 0.7878
0.1911 6.32 5400 0.3332 0.2553 0.7370 0.6386 0.6843 0.7873
0.2015 6.43 5500 0.3349 0.2645 0.7308 0.6538 0.6902 0.7931
0.1901 6.55 5600 0.3389 0.2587 0.7197 0.6682 0.6930 0.7975
0.197 6.67 5700 0.3349 0.2728 0.7400 0.6424 0.6878 0.7895
0.1907 6.78 5800 0.3354 0.2627 0.7454 0.6349 0.6857 0.7870
0.1853 6.9 5900 0.3420 0.2657 0.7356 0.6513 0.6909 0.7927
0.1841 7.02 6000 0.3399 0.2584 0.7308 0.6554 0.6910 0.7937
0.1739 7.13 6100 0.3409 0.2620 0.7364 0.6446 0.6874 0.7898
0.1768 7.25 6200 0.3417 0.2593 0.7314 0.6474 0.6868 0.7902
0.1762 7.37 6300 0.3384 0.2654 0.7398 0.6373 0.6847 0.7871
0.177 7.49 6400 0.3448 0.2541 0.7237 0.6547 0.6875 0.7922
0.1787 7.6 6500 0.3422 0.2513 0.7317 0.6425 0.6842 0.7881
0.1793 7.72 6600 0.3452 0.2611 0.7231 0.6582 0.6891 0.7936
0.1772 7.84 6700 0.3470 0.2587 0.7193 0.6618 0.6894 0.7946
0.1799 7.95 6800 0.3459 0.2547 0.7238 0.6494 0.6846 0.7898
0.1726 8.07 6900 0.3477 0.2507 0.7259 0.6419 0.6813 0.7869
0.1672 8.19 7000 0.3489 0.2492 0.7215 0.6499 0.6838 0.7897
0.1664 8.3 7100 0.3474 0.2498 0.7197 0.6491 0.6826 0.7890
0.1712 8.42 7200 0.3477 0.2516 0.7309 0.6404 0.6827 0.7870
0.166 8.54 7300 0.3487 0.2553 0.7209 0.6547 0.6862 0.7917
0.1706 8.65 7400 0.3487 0.2538 0.7239 0.6518 0.6860 0.7909
0.1674 8.77 7500 0.3506 0.2538 0.7216 0.6541 0.6862 0.7916
0.1655 8.89 7600 0.3476 0.2553 0.7283 0.6465 0.6849 0.7893
0.1609 9.01 7700 0.3498 0.2495 0.7273 0.6443 0.6833 0.7882
0.1647 9.12 7800 0.3507 0.2522 0.7255 0.6423 0.6814 0.7870
0.1531 9.24 7900 0.3503 0.2522 0.7292 0.6426 0.6832 0.7878
0.1577 9.36 8000 0.3524 0.2528 0.7212 0.6569 0.6875 0.7927
0.1592 9.47 8100 0.3517 0.2519 0.7186 0.6536 0.6845 0.7908
0.1615 9.59 8200 0.3514 0.2510 0.7183 0.6529 0.6841 0.7905
0.1529 9.71 8300 0.3515 0.2516 0.7221 0.6489 0.6835 0.7893
0.1607 9.82 8400 0.3520 0.2528 0.7212 0.6499 0.6837 0.7896
0.1506 9.94 8500 0.3524 0.2522 0.7220 0.6522 0.6853 0.7908

Framework versions

  • Transformers 4.31.0
  • Pytorch 2.1.0+cu121
  • Datasets 2.14.5
  • Tokenizers 0.13.3