9a9b77ec17
Almond: load data from the TSV files directly
...
Almond is not a translation task, it's a semantic parsing task.
And it has its own established on-disk file format, and we should
respect that, instead of messing with .tt files
2019-03-01 09:36:14 -08:00
73d94f187b
SummaryWriter: replace add_scalars with add_scalar
...
tensorboardX's add_scalars is bonkers, and will create a different
file for each tag, which makes the tensorboard completely unreadable
(See https://github.com/lanpa/tensorboardX/issues/366 )
Instead, use add_scalar(), which puts everything in the same file,
as tensorboard is designed to do.
2019-03-01 08:44:06 -08:00
ae4bbfea3a
arguments: remove bad ugly log dir messing
...
Stick with what the user says as the --save directory. If the user
needs, they can provide the timestamp on the command line.
2019-03-01 08:30:35 -08:00
7738b66e72
updates
2019-02-28 19:52:13 -08:00
d2656bb53a
update post_process_decoded_results.py
2019-02-27 12:01:42 -08:00
26f57a3059
don't shuffle dataset during prediction
2019-02-27 11:17:38 -08:00
eff58ba1eb
merge updates
2019-02-27 10:54:01 -08:00
2ef63c660e
export results to csv files
2019-02-27 10:52:38 -08:00
fede42e471
add cached option for predict
2019-02-20 11:22:32 -08:00
b8f4c1b0e5
fix max-margin loss implementation
2019-02-20 11:06:49 -08:00
f47da57330
fix bug for joining paths
2019-02-19 16:21:34 -08:00
ad5c9c6dc1
save cached files to an assigned path
...
useful when you don't have write permissions to dataset directory
2019-02-19 15:55:20 -08:00
f3545ad9d3
add max-margin loss
2019-02-19 13:48:52 -08:00
f2335e456b
Merge pull request #1 from Stanford-Mobisocial-IoT-Lab/wip/field
...
Remove hacks needed to use a different tokenizer for Almond
2019-01-24 10:44:33 -08:00
e4afb21928
minor fix
2019-01-23 19:13:26 -08:00
8c54dc1391
updates and fixes
2019-01-23 16:41:37 -08:00
fdfdd154c4
Remove hacks needed to use a different tokenizer for Almond
...
Almond data is always pretokenized (because we must preprocess
numbers/quoted strings/etc.).
Previously, we used a bad HACK of hardcoding almond in the generic
Field subclass. Now, we instead thread through the tokenizer/detokenizer
arguments from the right places.
In the future, we will probably want task classes, to clean up
the mess of hacks and hardcoded task-specific tweaks everywhere.
2019-01-23 10:48:49 -08:00
91ad08211f
Fix duplicate elmo option
2019-01-23 10:45:30 -08:00
9ec8d7cf2d
Merge remote-tracking branch 'upstream/master'
2019-01-23 10:44:20 -08:00
09bcae5dd5
Merge branch 'master' into mehrad/master
2019-01-23 10:34:17 -08:00
cd997f257f
prepend root before save
2019-01-10 21:24:43 +00:00
eed545bbf7
log_dir needs root prefix
2019-01-09 18:06:04 +00:00
81b7ea7e72
options for sgd
2019-01-09 01:13:46 +00:00
27bb192249
unique val tasks
2019-01-09 00:28:11 +00:00
287b911d9c
separate out root dir; add option for experiment name
2019-01-08 02:05:55 +00:00
873236860a
fix cuda error
2018-12-18 14:56:29 -08:00
b4bb1ad661
working version of differentiable BLEU loss
2018-12-17 16:43:06 -08:00
07dd886f9a
Update README.md
2018-12-12 12:31:44 -08:00
c096a1f5ba
bugs in overwrite; new best mqan model
2018-12-12 20:29:46 +00:00
1dc5f7d28e
moving schema raw files to s3; updated pretrained models
2018-12-10 21:52:34 +00:00
1bd922b5bc
back compat predict.py
2018-12-10 19:41:27 +00:00
9a4493b655
more dynamic elmo use cases
2018-12-10 19:41:27 +00:00
37003d2a0a
elmo works better w/o dropout
2018-12-10 19:41:27 +00:00
6260c9f527
adding elmo naively as a replacement
2018-12-10 19:41:27 +00:00
3aecae06d9
better to have tensorboard logging on as default
2018-12-10 19:41:27 +00:00
a3ede17b3e
bug fixes
2018-12-06 08:29:11 +00:00
e8d8ae0eab
add ELMo embeddings
2018-12-05 17:07:07 -08:00
a9b48d2b97
update error analysis
2018-11-29 14:48:45 -08:00
787b0887dc
adding device accuracy
2018-11-29 10:17:25 -08:00
cd4782574d
more updates
2018-11-28 18:04:04 -08:00
827e27d472
updates on post_processing
2018-11-28 18:00:05 -08:00
616b2833d3
updates
2018-11-28 16:36:23 -08:00
6b31d8f4cb
Remove CUDA from the command to build docker image
2018-11-28 07:40:44 -08:00
17c09d977b
Add differentiable BLEU loss
...
-Using differentiable BLEU loss instead of cross_entropy loss
-it helps decreasing train-test evaluation gap
2018-11-27 15:22:38 -08:00
39720d0713
Update README.md
2018-11-16 18:31:36 -08:00
42f1bad639
Update README.md
2018-11-16 18:23:23 -08:00
ae8d64b898
CoVe results
2018-11-16 18:20:57 -08:00
fe580047ce
better predict; better MQAN+CoVe
2018-11-17 02:17:29 +00:00
358ed621fd
bug fixes
2018-11-08 13:52:48 -08:00
f9cba2b427
Updates
...
1) Use CPU automatically if cuda is not available
2) Use a different directory to store eval results
2018-11-08 13:23:12 -08:00
Giovanni Campagna