allegroai
f02b1fc190
Add configuration files logging example
2023-10-02 20:43:44 +03:00
allegroai
ea9b613d4b
Update README
2023-09-30 00:02:40 +03:00
allegroai
2a5fd56fdb
Version bump to v1.13.1
2023-09-26 00:33:44 +03:00
allegroai
b83baa6f2e
Fix Hydra overrides integration ( #1120 )
2023-09-25 22:16:54 +03:00
allegroai
96646dc46a
Fix GPU memory used reports 0 when memory can not be queried per process (edge case)
2023-09-25 22:16:02 +03:00
allegroai
e71c257290
Version bump to v1.13.0
2023-09-22 16:26:06 +03:00
allegroai
c5675733a0
Fix GPU memory used reports 0 when memory can not be queried per process
2023-09-22 16:25:50 +03:00
allegroai
0442579e23
Fix scroll_id argument missing in events.get_scalar_metric_data endpoint (mainly for use in the APIClient)
2023-09-20 15:27:27 +03:00
allegroai
9e69b09c9f
Fix PEP8 issues
2023-09-20 15:26:50 +03:00
allegroai
701582d5ff
Fix incorrect module used
2023-09-16 20:53:47 +03:00
allegroai
1820423b02
Fix keep backwards compatibility for pipeline.add_step() positional arguments
2023-09-16 20:53:30 +03:00
allegroai
1b27ed07f7
Add force_single_script_file to Task.create() to avoid git repo auto detection
2023-09-16 20:52:42 +03:00
allegroai
76076b09dd
Fix compatibility issue with requests < 2.28
2023-09-12 00:51:47 +03:00
allegroai
d131ff2a45
Fix import sys should not appear in pipeline component auto detected packages
2023-09-12 00:51:25 +03:00
allegroai
96b89d76b8
Fix dataset zip extraction might fail when creating folders concurrently
2023-09-12 00:50:23 +03:00
natephysics
d458924160
Add support for recursive list, dict, and tuple ref parsing for pipeline controller.add step() parameter override ( #1099 )
...
* feat: ✨ Added optional support for list, dicts, tuples in Pipeline parameter_overrides
* style: 🎨 Updated to pass the flake8 formatting guidelines
* docs: 📝 There was a small typo I noticed in the documentation. Two extra '
2023-09-09 22:17:49 +03:00
allegroai
c922c40d13
Fix model reporting in tensorflow 2.13 does not work properly ( #1112 )
2023-09-08 22:14:53 +03:00
pollfly
cd61efe6df
Edit dataset_path docstring ( #1105 )
2023-09-06 17:07:59 +03:00
allegroai
04fe7dffe3
Fix OutputModel should prefer connected task name over filename stem
2023-09-02 17:49:02 +03:00
allegroai
71c74f977b
Add Model archive/unarchive methods
2023-08-31 00:10:38 +03:00
allegroai
713501c611
Fix hydra overrides integration (issue #552 )
2023-08-31 00:10:21 +03:00
allegroai
6756dec06a
Json serialization error should raise exception
2023-08-31 00:09:36 +03:00
Alex Burlacu
0959974920
Fix fastAI binding not reporting scalars if tensorboard is installed
2023-08-24 13:55:05 +03:00
Alex Burlacu
0b521b00a6
Use os.register_at_fork instead of monkey patching fork for python > 3.6
2023-08-21 14:06:59 +03:00
allegroai
f6ad5e6c06
Fix PyPI Downloads badge
2023-08-17 23:28:03 +03:00
Alex Burlacu
88d036b367
Update PyNVML to the latest NVIDIA version
2023-08-15 13:54:04 +03:00
Alex Burlacu
afe178b002
Bump version to 1.12.2
2023-08-11 19:57:07 +03:00
Alex Burlacu
83a04d438c
Adjust for pandas < 2.0
2023-08-11 19:56:08 +03:00
Alex Burlacu
d4b11dfa22
Create OutputModel base model lazily
2023-08-11 13:11:58 +03:00
Alex Burlacu
2c44bff461
Fix Hydra support both Hydra section overrides and
2023-08-11 13:09:19 +03:00
allegroai
c15f012e1b
Add PipelineController always_create_from_code=True (False is previous behavior where we deserialize always from backend when running remotely, new flow means the pipeline is always created from code)
...
Fix pipeline decorator does not read the pipeline arguments back from the backend when running remotely
2023-08-04 19:54:55 +03:00
allegroai
d26ce48dbe
Improve artifacts serialization: should now be consistent when serializing pandas object into gzip stream (csv.gz)
2023-08-04 19:08:51 +03:00
allegroai
197894735b
Make sure we work on a copy of task objects
2023-08-04 19:08:09 +03:00
allegroai
696c349488
Improve warning message on task execute_remotely
2023-08-04 19:07:35 +03:00
allegroai
b846772138
Fix pipeline monitor must be called after pipeline is completed (just in case we missed something)
2023-08-04 19:07:05 +03:00
allegroai
99c7eecbee
Fix pipeline should always use artifacts hash for checking cache hits
2023-08-04 19:06:15 +03:00
allegroai
46c6d2bf0f
Fix proxy object support for lists
2023-08-04 19:05:24 +03:00
Alex Burlacu
a8b71de3e4
Adjust a few things in the README
2023-08-04 14:30:10 +03:00
Alex Burlacu
92511efbde
Allow setting multiprocessing start method
2023-07-31 21:34:13 +03:00
Alex Burlacu
80ef359f45
Bump version and clarify docs
2023-07-31 21:33:28 +03:00
allegroai
e467c05fb4
Fix S3 uploads fail with LazyEvalWrapper type error ( #1081 )
2023-07-30 13:04:42 +03:00
Michael Pilosov
b1f99abc96
Raise error if dataset name empty ( #1083 )
...
* raise error if dataset name empty.
hidden dataset project when name is an empty string.
if the dataset project happens to match an existing project, it can be quite jarring to have one's history of results suddenly become hidden.
if there's an intent behind this behavior, please let me know and I'd gladly change it to a warning instead.
* move check to top of init, change message
* whitespace
2023-07-27 14:54:19 +03:00
pollfly
c451589298
Edit docstrings ( #1084 )
2023-07-25 12:19:24 +03:00
Alex Burlacu
09363b0d30
Fix typo
2023-07-21 14:23:32 +03:00
alex-burlacu-clear-ml
05c048d0bd
Fix style and grammar mistakes
2023-07-19 22:55:08 +03:00
alex-burlacu-clear-ml
ac1ae7f447
Update errata, clarify a few points
2023-07-19 20:23:03 +03:00
Alex Burlacu
618a625b74
Bump version to 1.12.0
2023-07-19 17:33:28 +03:00
Alex Burlacu
9b680d1743
Add clarification about breaking changes in sdk 1.11.0 and 1.11.1
2023-07-19 17:29:41 +03:00
Alex Burlacu
cc875f0fbe
Fix linter issues
2023-07-19 17:28:58 +03:00
Alex Burlacu
028a835676
Deprecate Task.get_by_name
2023-07-19 02:11:40 +03:00
