Edit model card

t5-base-pointer-adv-top_v2

This model is a fine-tuned version of google/mt5-base on the top_v2 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0255
  • Exact Match: 0.8472

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: 0.001
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 64
  • total_train_batch_size: 512
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • training_steps: 3000

Training results

Training Loss Epoch Step Validation Loss Exact Match
2.2938 0.41 200 0.5532 0.0012
0.671 0.82 400 0.1624 0.1610
0.5276 1.23 600 0.0692 0.2157
0.4196 1.64 800 0.0491 0.2259
0.3593 2.05 1000 0.0400 0.2291
0.3471 2.46 1200 0.0335 0.2297
0.3416 2.87 1400 0.0307 0.2318
0.3351 3.29 1600 0.0307 0.2334
0.3316 3.7 1800 0.0297 0.2343
0.3312 4.11 2000 0.0282 0.2344
0.3271 4.52 2200 0.0262 0.2365
0.3241 4.93 2400 0.0263 0.2365
0.3227 5.34 2600 0.0259 0.2368
0.3201 5.75 2800 0.0257 0.2365
0.3227 6.16 3000 0.0255 0.2365

Framework versions

  • Transformers 4.24.0
  • Pytorch 1.13.0+cu117
  • Datasets 2.7.0
  • Tokenizers 0.13.2
Downloads last month
2