71e9249ff4
Classifier objects will be removed in 5.0
2023-05-31 13:42:42 -07:00
97c4eef086
Move deserialize to Model object
2023-04-17 21:35:38 -07:00
457b569741
Update README
2023-04-17 21:33:03 -07:00
4546c4cffa
Fix profiler and benchmark
2023-04-17 21:28:24 -07:00
7b7ef39d0b
Merge compiler into model.py
2023-04-17 21:15:18 -07:00
a252a15e9d
Clean up code
2023-04-17 21:06:47 -07:00
9513025e60
Fix type annotations
2023-04-17 18:16:20 -07:00
2c3fc77ba6
Finish classification explanations
...
A couple things I missed in 7f68dc6fc6
2023-04-16 15:48:19 -07:00
d8f3d2e701
Bump model version
...
99ad07a876
2023-04-16 15:36:49 -07:00
7f68dc6fc6
Add classification explanations
...
Closes #17
2023-04-16 15:35:53 -07:00
99ad07a876
Casefold
...
Closes #14
2023-04-16 14:49:03 -07:00
f38f4ca801
Add profiler
2023-04-16 14:27:31 -07:00
56550ca457
Remove Classifier objects
...
Closes #16
2023-04-16 14:27:07 -07:00
75fdb5ba3c
Split compiler into two functions
2023-01-15 09:39:35 -08:00
071656c2d2
Bump version to 4.0.1
2022-12-24 12:49:12 -08:00
aad590636a
Fix type annotations
2022-12-24 12:48:43 -08:00
099e810a18
Fix check
2022-12-24 12:44:09 -08:00
822aa7d1fd
Bump version to 4.0.0
2022-12-24 12:18:51 -08:00
8417c8acda
Recompile model
2022-12-24 12:18:25 -08:00
ec7f4116fc
Include file name of output in arguments
2022-12-24 12:17:44 -08:00
f8dbc78b82
Allow hash algorithm selection
...
Closes #9
2022-12-24 11:18:05 -08:00
6f21e0d4e9
Remove debug print lines from compiler
2022-12-24 10:48:09 -08:00
41bba61410
Remove has_emoji and bump model version
...
Closes #11
2022-12-24 10:47:23 -08:00
10668691ea
Normalize characters
...
Closes #3
2022-12-24 10:46:40 -08:00
295a1189de
Include numbers in tokenized output
...
Closes #12
2022-12-24 10:42:50 -08:00
74b2ba81b9
Deserialize from file
2022-12-23 10:49:24 -08:00
9916744801
New type annotation for serialize
2022-12-23 10:33:56 -08:00
7e7b5f3e9c
Performance improvements
2022-12-22 18:01:37 -08:00
a76c6d3da8
Bump version to 3.1.1
2022-11-27 15:01:06 -08:00
c84758af56
list, not tuple
2022-11-27 15:00:37 -08:00
3a9c8d2bf2
Revert "Bump version to 3.1.1"
...
This reverts commit 12f97ae765
2022-11-27 14:56:10 -08:00
12f97ae765
Bump version to 3.1.1
2022-11-27 14:54:11 -08:00
c754293d69
Compiler performance improvements
2022-11-27 14:32:44 -08:00
8d42a92848
Add type annotation to Model.get()
2022-11-27 13:36:49 -08:00
e4eb322aa7
Bump version to 3.1.0
2022-11-26 18:37:11 -08:00
83ef71e8ce
Remove doc for gptc classify --category
2022-11-26 18:36:41 -08:00
991d3fd54a
Revert "Bump version to 3.1.0"
...
This reverts commit b3e6a13e65
2022-11-26 18:36:18 -08:00
b3e6a13e65
Bump version to 3.1.0
2022-11-26 18:34:04 -08:00
b1228edd9c
Add CLI for Model.get()
2022-11-26 18:28:44 -08:00
25192ffddf
Add ability to look up individual token
...
Closes #10
2022-11-26 18:17:02 -08:00
548d670960
Use Classifier for --category
2022-11-26 17:50:26 -08:00
b3a43150d8
Split hash function
2022-11-26 17:42:42 -08:00
08437a2696
Add normalize()
2022-11-26 17:17:28 -08:00
fc4665bb9e
Separate tokenization and hashing
2022-11-26 17:04:56 -08:00
30287288f2
Fix README issues
2022-11-26 16:45:30 -08:00
448f200923
Add confidence to Model; deprecate Classifier
2022-11-26 16:41:29 -08:00
b4766cb613
Bump version to 3.0.1
2022-11-25 19:44:32 -08:00
f1a1ed9e2a
Remove most emoji-optional code
...
Almost all of the code previously used to make the emoji module optional
is removed in this commit. It was always my intent to make the `emoji`
module a hard dependency in v3.0.0 and remove the code for making it
optional, but for some reason I remembered to do the former but not the
latter; in fact, I added emoji-optional code to the new model handling
code. I can't completely remove this code because 3.0.0 will not
successfully deserialize a model without the `has_emoji` field in the
JSON config options, but this commit removes as much as possible without
breaking the model format and API version.
See also issue #11
2022-11-25 19:39:31 -08:00
7ecb7dd90a
Bump version to 3.0.0
2022-11-23 17:48:46 -08:00
3340abbf15
Fix CLI tool
2022-11-23 17:47:27 -08:00
