Update README.md

README.md

@@ -176,7 +176,33 @@ detect_lang(
 
 ## Training procedure
 
-* WIP
+
+Fine-tuning was done via the `Trainer` API. Here is the [Jupyter notebook](https://github.com/DominguesM/language-detection/blob/main/Language_Detector_Lora.ipynb) with the training code.
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+
+- learning_rate: 2e-05
+- train_batch_size: 64
+- eval_batch_size: 128
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- num_epochs: 2
+
+### Training results
+
+The validation results on the `valid` split of the Language Identification dataset are summarised here below.
+
+| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1     |
+|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
+| 1.4403        | 1.0   | 1094 | 0.0591          | 0.9952   | 0.9952 |
+| 0.0256        | 2.0   | 2188 | 0.0272          | 0.9955   | 0.9955 |
+
+In short, it achieves the following results on the validation set:
+
+- Loss: 0.0298
+- Accuracy: 0.9946
+- F1: 0.9946
 
 ### Framework versions