Timeline for Expected performance of training tf.keras.Sequential model with model.fit, model.fit_generator and model.train_on_batch
Current License: CC BY-SA 4.0
21 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 5 at 7:19 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Jun 6 at 15:05 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Feb 7 at 4:04 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Oct 4, 2023 at 17:00 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Jun 6, 2023 at 8:04 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Feb 4, 2023 at 7:06 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Oct 6, 2022 at 6:04 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| May 26, 2022 at 10:01 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Jan 25, 2022 at 8:02 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Sep 19, 2021 at 8:07 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| May 19, 2021 at 12:06 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Jan 16, 2021 at 3:06 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Sep 16, 2020 at 18:03 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| May 18, 2020 at 2:03 | history | bumped | CommunityBot | This question has answers that may be good or bad; the system has marked it active so that they can be reviewed. | |
| Apr 13, 2020 at 16:29 | history | edited | Tuukka Nieminen | CC BY-SA 4.0 |
Fixed the code as per the comments below.
|
| Apr 13, 2020 at 12:36 | answer | added | Tuukka Nieminen | timeline score: 0 | |
| Mar 3, 2020 at 11:19 | comment | added | Tuukka Nieminen | If you are experiencing memory problems with the generator, the parameter max_queue_size has a direct impact on it. The somewhat large value 1024 I used might cause problems if a single batch does reserve a considerable amount of memory. | |
| Mar 3, 2020 at 11:19 | comment | added | Tuukka Nieminen | Thank you for your comment. No, I haven't been able to resolve the issue. I suspected that the large arrays passed as an argument to the generator might cause the issue, but another, definitely a bit dirty approach with global variables did not help at all. Also, the number of workers associated to the generator does not have an impact to computation times, With more workers, the queue fills up more rapidly, but after it has been filled, no impact whatsoever. | |
| Feb 26, 2020 at 17:30 | comment | added | CAta.RAy | Did you figure this out? I am having similar problems and I have even experienced problems with the size of the Batches. fit_generator has to be considerable smaller to fit on memory. | |
| Dec 18, 2019 at 15:10 | review | First posts | |||
| Dec 18, 2019 at 15:45 | |||||
| Dec 18, 2019 at 15:08 | history | asked | Tuukka Nieminen | CC BY-SA 4.0 |