metadata
{}
Test-Model
This model is a fine-tuned version of llama, and is specialized for document classification. The overall accuracy of the model is 0.978852 and the hit rate is 0.543211.
Training Parameters
Epochs: 5
Batch Size: 20
Threshold: 0.975
Token Length: 512
Data Set Size: 10000
Evenly Distributed: True
Municipalities: Heroey, Kongsberg, Boemlo, Luster, Maalselv
Learning Rate: 0.0001
Weight Decay: 0.01
Eval Accumulation Steps: 4
| Label | Accuracy | Accuracy Accumulation | Hit Rate | Hit Rate Accumulation |
|---|---|---|---|---|
| 0 | 0.9333333333333333 | 0.9333333333333333 | 0.9767441860465116 | 0.9767441860465116 |
| 1 | 0.95 | 0.95 | 0.9743589743589743 | 0.9743589743589743 |
| 2 | 0.9148936170212766 | 0.9148936170212766 | 0.9772727272727273 | 0.9772727272727273 |
| 3 | 0.4 | 0.4 | 0.5714285714285714 | 0.5714285714285714 |
| 4 | 0.25 | 0.25 | 0.6329113924050633 | 0.6329113924050633 |
| 5 | 0.8571428571428571 | 0.8571428571428571 | 0.9375 | 0.9375 |
| 6 | 0.9888888888888889 | 0.9888888888888889 | 1.04093567251462 | 1.04093567251462 |
| 7 | 0.94 | 0.94 | 0.9791666666666666 | 0.9791666666666666 |
| 8 | 0.6666666666666666 | 0.6666666666666666 | 0.8571428571428571 | 0.8571428571428571 |
| 9 | 0.9888888888888889 | 0.9888888888888889 | 0.994413407821229 | 0.994413407821229 |