- handle collection for multiple trains
- how many tensors to collect when sunk
- change loglevel for get_learning_rate
- update calculation of `max_file_size`
- fix how collect_tensor_freq counting

- fix writer pool not exit when max_file_size too small
- fix API docs to illustrate `collect_tensor_freq` and `max_file_size`

Add more detail note for SummaryCollector and SummaryRecord,
else if it is used not right, some proplem will be caused.

Merge pull request !2770 from LiHongzhang/capture_time

I added a SummaryCollector to help users automatically collect information
such as the network, loss, learning rate and so on, making it easier to collect this information.

It also can collect train lineage and eval lineage information which is
collected by TrainLineage Callback and EvalLineage Callback in
MindInsight.

I also add some UT for SummaryCollect to keep the code correct.

Merge pull request !2193 from wenkai/wk1_log_level_0617

1. When initing SummaryRecord, if check failed, self._close is not set,
   which leads to 'SummaryRecord' object has no attribute '_closed'
   AttributeError later on.

2. There may be too many processes for handling summary adapting.
   See issue #I1K6K7

Signed-off-by: candanzg <zhangshucheng@huawei.com>

Merge pull request !464 from candanzg/summary_event_notes

Signed-off-by: candanzg <zhangshucheng@huawei.com>

Signed-off-by: leonwanghui <leon.wanghui@huawei.com>