File size: 764 Bytes
308df6a 0bd1497 308df6a c312761 308df6a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
In this project, we aim to develop a text classification system to classify Turkish texts into specific categories. We are currently in the first phase of our project and in this context, we are researching and collecting Turkish text classification datasets from internet sources (HuggingFace, Kaggle, etc.). Dataset Statistics The combined dataset consists of a total of 1,227,879 instructions. The average length for each component is as follows: Instruction Length (in characters): 83.87 Input Length (in characters): 262.03 Output Length (in characters): 5.58 In terms of variability, the standard deviations for each component are: Instruction Length (in characters): 51.41 Input Length (in characters): 479.51 Output Length (in characters): 14.75 |