Piotr Czapla
be117abac4
Make the end to end test run correctly
2018-11-24 23:51:23 +01:00
Piotr Czapla
8da47324c2
Clean ups and fixes
2018-11-22 15:32:40 +01:00
Piotr Czapla
dbd4884228
Fixes after mergin with master and updateing to newset fastai
2018-11-22 01:14:38 +01:00
Piotr Czapla
934fc79384
Merge branch 'master' into bilm
2018-11-21 23:48:23 +01:00
Piotr Czapla
9aa877dcd0
Share trained LM between different classfiication runs
2018-11-21 18:51:42 +01:00
Piotr Czapla
979eb196d8
Change the classfication training learning rate to the one that was working te best in my exp. on bidirectional clasification
2018-11-21 18:46:05 +01:00
Piotr Czapla
cbed02d5e0
Merge branch 'master' into bilm
2018-11-21 18:23:54 +01:00
NAUSICAA\Julian
8cb867b066
Compatibility with new fastai version
2018-11-20 19:21:02 -03:00
Aayush
429a3fe4c2
Merge pull request #13 from n-waves/models_path_fix
...
Models path fix, default_rules and sentencepiece support for train_clas
2018-11-20 12:25:25 +05:30
Aayush
5ffbe8ba5c
add vocab_size to sentencepiece
2018-11-19 19:13:55 +05:30
Piotr Czapla
7f1f8efcc3
Working version of biclassfier
2018-11-19 12:58:58 +01:00
Piotr Czapla
c821d2e783
first version of bi classifier
2018-11-19 09:59:08 +01:00
Nirant
56a9feec7d
Removed dependency note, use requirements.txt
2018-11-19 12:22:41 +05:30
Piotr Czapla
37b73e262f
Fix bug where de-all was de-100
2018-11-17 17:14:13 +01:00
Piotr Czapla
2674a713fc
fix resuming training of classifier
2018-11-17 17:13:15 +01:00
Piotr Czapla
e97085337e
Fix dropout and classification accuracy. 0.91 on imdb
2018-11-17 16:47:17 +01:00
Sebastian
7fe059c9a2
Fixed models path for vocabulary
2018-11-17 14:47:27 +00:00
Piotr Czapla
e1418b2114
fix model dir mkdir
2018-11-17 00:11:56 +01:00
Piotr Czapla
5cc4bfd444
Add loading weights and itos, so that we can extend training
2018-11-16 23:48:22 +01:00
Piotr Czapla
895a58c88d
Merge branch 'master' into bilm
2018-11-16 23:39:15 +01:00
Piotr Czapla
36b056a465
Fix issues discovered during execution of end-to-end test.
2018-11-16 23:31:16 +01:00
Piotr Czapla
a21f100ab2
Merge branch 'master' into sentencepiece
2018-11-16 23:23:49 +01:00
Piotr Czapla
aa6b59a0b3
Fix vocab size so that it remains 60k as in case of pretrain_lm + add longer lm training
2018-11-16 23:19:15 +01:00
Piotr Czapla
ebbf09775f
Merge branch 'master' into bilm
2018-11-16 20:09:41 +01:00
Piotr Czapla
45389be413
Use drop_mult properly
2018-11-16 20:02:32 +01:00
Piotr Czapla
23b17da61e
Add end to end test and improve the train_clas params
2018-11-16 19:51:29 +01:00
Piotr Czapla
977a506cfd
Fix bug where LM was always intialized as QRNN despite the paramters.
2018-11-16 19:50:36 +01:00
Aayush
e72b19499c
fix indent
2018-11-16 23:54:48 +05:30
aayush
59c8852b5d
sentencepiece for pretraining
...
modified: fastai_contrib/utils.py
modified: ulmfit/pretrain_lm.py
2018-11-16 23:14:24 +05:30
Piotr Czapla
f0e538f621
Add support for QRNN training to train_clas
2018-11-16 17:04:52 +01:00
Piotr Czapla
363130bcff
Update to the docs
2018-11-16 00:17:41 +01:00
Piotr Czapla
9a60ef2fbd
Add assertions to train_clas and ability to limit the dataset size
2018-11-15 23:40:10 +01:00
Piotr Czapla
d8b95430b8
Chain all scripts needed to prepare a wiki dump, put .unk to new folder
2018-11-15 23:39:27 +01:00
Nirant K
43340d03bc
Fix minor typos in train_clas
2018-11-15 18:04:15 +00:00
Nirant K
56fc150eb1
Merge xnli
2018-11-15 15:53:37 +00:00
Nirant K
653f6e96a0
Classifier fails at fit_one_cycle
2018-11-15 15:03:52 +00:00
Sebastian
94fa1a3ecb
Encapsulated data reading in utils method, removed fastai processing, added doc string
2018-11-15 14:36:50 +00:00
Sebastian
d5e6849cd8
Shortened XNLI download script, removed unnecessary parts; extracts now to data folder
2018-11-15 14:33:21 +00:00
Piotr Czapla
33f9eb2cc7
Reuse RNNCore in implementation of BiLM, add accuracy
2018-11-14 21:16:16 +01:00
Nirant K
62e6777952
Add download and unzip script
2018-11-14 15:53:59 +00:00
Piotr Czapla
cf7d93c378
Changes to pretrain_lm so that it works with bilm
2018-11-14 14:59:23 +01:00
Nirant K
0fb8161061
Merge branch 'master' of https://github.com/n-waves/ulmfit-multilingual
2018-11-14 13:56:58 +00:00
Nirant K
70680c21ce
Add better logs, add multi-language flag to pretrain
...
- Added lang variable to pretrain_lm
- Tried the QRNN model with Hindi
- Added better logging and consistent model saving paths
- Fixed minor bug in validate(model, ...) to to validate(learn.model, ...)
2018-11-14 13:54:31 +00:00
Nirant
5f845724d0
Add dependencies
2018-11-14 17:13:40 +05:30
Nirant
820ff231bd
Add Download and Process Docs
2018-11-14 17:04:47 +05:30
Piotr Czapla
81519a6066
Merge branch 'master' into bilm
2018-11-14 11:41:50 +01:00
Piotr Czapla
38ca35c044
Fix pretrain_lm to work with yesterdays changes to fastai text API
2018-11-14 11:41:26 +01:00
Sebastian
cba86ab5d4
Added initial classes and changes for BiLM implementation
2018-11-13 17:46:28 +00:00
Piotr Czapla
e13dbea1e1
Remove missleading docs.
2018-11-12 13:14:44 +01:00
Piotr Czapla
5ad6c2ace1
Update to new fastai.master with fixed qrnn issue
2018-11-12 13:08:00 +01:00
Piotr Czapla