VuongQuoc's picture
Model save
8789767
metadata
license: mit
base_model: microsoft/deberta-v3-large
tags:
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: checkpoints_28_9_microsoft_deberta_V5
    results: []

checkpoints_28_9_microsoft_deberta_V5

This model is a fine-tuned version of microsoft/deberta-v3-large on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.6408
  • Map@3: 0.8542
  • Accuracy: 0.76

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: 4
  • eval_batch_size: 4
  • seed: 42
  • gradient_accumulation_steps: 32
  • total_train_batch_size: 128
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.2
  • num_epochs: 1

Training results

Training Loss Epoch Step Validation Loss Map@3 Accuracy
1.6111 0.05 25 1.6092 0.5092 0.325
1.6139 0.11 50 1.6085 0.7 0.575
1.6096 0.16 75 1.5867 0.7583 0.645
1.2905 0.21 100 1.1496 0.7767 0.66
1.0263 0.27 125 0.8628 0.8067 0.705
0.9475 0.32 150 0.7252 0.8458 0.75
0.841 0.37 175 0.7018 0.8492 0.76
0.8301 0.43 200 0.7137 0.8492 0.755
0.823 0.48 225 0.6633 0.8525 0.755
0.8263 0.53 250 0.6751 0.8608 0.765
0.7962 0.59 275 0.6704 0.8542 0.755
0.8013 0.64 300 0.6583 0.8525 0.755
0.789 0.69 325 0.6497 0.8533 0.76
0.7979 0.75 350 0.6512 0.8525 0.755
0.7751 0.8 375 0.6445 0.8583 0.765
0.7993 0.85 400 0.6424 0.8558 0.765
0.7685 0.91 425 0.6408 0.8542 0.76
0.7807 0.96 450 0.6408 0.8542 0.76

Framework versions

  • Transformers 4.32.1
  • Pytorch 2.0.0
  • Datasets 2.9.0
  • Tokenizers 0.13.3