mindspore2022

Commit Graph

Author	SHA1	Message	Date
Li Hongzhang	299469babb	address the importance of closing the SummaryRecord and illustrate how	5 years ago
Li Hongzhang	97d8673018	warn when values duplicate and set mode to 'eval' to avoid extra recording	5 years ago
chenzomi	a834a6308e	change some comment name in the whole project	5 years ago
ougongchang	939cd29d7e	Add a callback named SummaryCollector and delete SummaryStep callback 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.	5 years ago
Li Hongzhang	0921c1e538	enhance the SummaryRecord with set_mode and add_value	5 years ago
mindspore-ci-bot	373832d030	!2193 fix log level too high: step has no summary record is normal Merge pull request !2193 from wenkai/wk1_log_level_0617	5 years ago
wenkai	a2bad5c72d	fix log level too high: step has no record is normal.	5 years ago
Li Hongzhang	d31e14f593	fix having too many processes and no attribute of '_closed' 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	5 years ago
Li Hongzhang	ccf49b7c0e	shape is tuple, not a integer	5 years ago
李鸿章	e798fe2de7	mimic numpy behavior when min == max and give warnings when too large	5 years ago
李鸿章	324195280c	summary handle not a image tensor	5 years ago
shijianning	f2230fb070	fix pylint warnings	5 years ago
李鸿章	32c1d558f4	SummaryRecord as context manager	5 years ago
guozhijian	1b3a0277b7	回退 'Pull Request !713 : Use a resident process to write summary files'	5 years ago
李鸿章	69d3abfdd3	reduce dead step(step % flush_step > 0) summary	5 years ago
李鸿章	e0890e6acc	use a resident process for ordered and fast writing summary	5 years ago
李鸿章	abbdf78897	reduce expensive calculation and memory usage	5 years ago
李鸿章	2ac60a1ad4	style changes and nonfunctional modifies	5 years ago
wenkai	ab04b3dc4b	fix np.histograms(bins='auto') sometimes calc very small width and very large bucket number, which lead to error/long compute time.	5 years ago
candanzg	7e8e44ef56	fixed summary graph bug Signed-off-by: candanzg <zhangshucheng@huawei.com>	5 years ago
mindspore-ci-bot	f9849bf5aa	!464 Supplement summary note Merge pull request !464 from candanzg/summary_event_notes	6 years ago
candanzg	a91f82d79f	Supplement summary log Signed-off-by: candanzg <zhangshucheng@huawei.com>	6 years ago
wenkai	cce61d462c	histogram python dev	6 years ago
zhunaipan	930a1fb0a8	initial version Signed-off-by: leonwanghui <leon.wanghui@huawei.com>	6 years ago

24 Commits (c94dea6a512eddb6cbe8b591268d82d7b9aa3209)