[SPARK-16653][ML][Optimizer] update ANN convergence tolerance param default to 1e-6

[WeichenXu123]

What changes were proposed in this pull request?

replace ANN convergence tolerance param default
from 1e-4 to 1e-6

so that it will be the same with other algorithms in MLLib which use LBFGS as optimizer.

How was this patch tested?

Existing Test.

[srowen]

Although it's not necessarily true that two different algos using L-BFGS need about the same convergence tolerance, I do agree in this case that they probably should match up. It's a small change to the defaults, so LGTM

[SparkQA]

Test build #62610 has finished for PR 14286 at commit cd86df5.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[srowen]

Merged to master