Statistical Inference — the formal definition of a consistent estimator

Suppose that $X_1, X_2, X_3, ..., X_n$ be a random sample of size n. An estimator $T_n$ calculated from this sample is said to be a consistent estimator of parameter $\theta$ for all $\varepsilon>0, \eta>0$ and $n\ge m$ if