Edit model card

salt_language_ID

This model is a fine-tuned version of google/t5-efficient-tiny on the generator dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0127
  • Accuracy: 0.9805

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: 64
  • eval_batch_size: 64
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 10
  • training_steps: 20000

Training results

Training Loss Epoch Step Validation Loss Accuracy
0.5069 0.025 500 0.1145 0.8337
0.0644 0.05 1000 0.0489 0.9170
0.0511 0.075 1500 0.0605 0.9056
0.0462 0.1 2000 0.0332 0.9432
0.0411 0.125 2500 0.0358 0.9385
0.0409 0.15 3000 0.0267 0.9509
0.0365 0.175 3500 0.0244 0.9563
0.0359 0.2 4000 0.0285 0.9536
0.035 0.225 4500 0.0355 0.9388
0.0321 0.25 5000 0.0264 0.9570
0.0327 0.275 5500 0.0278 0.9513
0.0313 0.3 6000 0.0217 0.9630
0.0305 0.325 6500 0.0255 0.9556
0.0285 0.35 7000 0.0187 0.9630
0.0293 0.375 7500 0.0225 0.9620
0.0264 0.4 8000 0.0228 0.9614
0.0272 0.425 8500 0.0195 0.9664
0.0268 0.45 9000 0.0178 0.9688
0.0259 0.475 9500 0.0164 0.9677
0.0256 0.5 10000 0.0167 0.9721
0.0241 0.525 10500 0.0182 0.9647
0.0235 0.55 11000 0.0212 0.9657
0.0239 0.575 11500 0.0145 0.9735
0.0239 0.6 12000 0.0173 0.9704
0.0234 0.625 12500 0.0152 0.9768
0.0229 0.65 13000 0.0181 0.9698
0.023 0.675 13500 0.0154 0.9735
0.0224 0.7 14000 0.0157 0.9708
0.0221 0.725 14500 0.0155 0.9714
0.0219 0.75 15000 0.0145 0.9755
0.0213 0.775 15500 0.0159 0.9735
0.0197 0.8 16000 0.0129 0.9751
0.0206 0.825 16500 0.0154 0.9724
0.02 0.85 17000 0.0140 0.9724
0.0209 0.875 17500 0.0115 0.9772
0.0191 0.9 18000 0.0129 0.9735
0.0194 0.925 18500 0.0120 0.9765
0.0191 0.95 19000 0.0133 0.9741
0.0183 0.975 19500 0.0166 0.9731
0.0207 1.0 20000 0.0127 0.9805

Framework versions

  • Transformers 4.40.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1
Downloads last month
2,331
Safetensors
Model size
15.6M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for yigagilbert/salt_language_ID

Finetuned
(7)
this model

Evaluation results