wassname/multifit
1
0
Fork 0
mirror of https://github.com/wassname/multifit.git synced 2026-09-11 12:20:41 +08:00
Code Issues Packages Projects Releases Wiki Activity

Training with noise & label smoothing

Browse Source
This commit is contained in:
Piotr Czapla
2019-02-22 12:02:18 +01:00
parent c69d31c420
commit 015f04ec08
4 changed files with 90 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/test_end_to_end.py
+20
View File
@@ -115,6 +115,26 @@ def test_ulmfit_fastai_end_to_end():
exp2 = ulmfit.train_clas.CLSHyperParams.from_lm(test_data / 'imdb', exp.model_dir)
exp2.train_cls(num_lm_epochs=0, unfreeze=False, bs=4, )
def test_ulmfit_fastai_end_to_end_label_smoothing():
""" Test ulmfit with sentencepiece tokenizer on small wikipedia dataset.
"""
test_data, wt2 = get_test_data()
lm_name = 'end-to-end-test-fastai'
exp = ulmfit.pretrain_lm.LMHyperParams(
dataset_path=wt2,
lang='en',
cuda_id=cuda_id,
qrnn=False,
tokenizer='f',
max_vocab=100,
name=lm_name,
)
exp.train_lm(num_epochs=1, bs=2, label_smoothing_eps=0.1)
exp2 = ulmfit.train_clas.CLSHyperParams.from_lm(test_data / 'imdb', exp.model_dir)
exp2.train_cls(num_lm_epochs=0, unfreeze=False, bs=4, label_smoothing_eps=0.1 )
def test_ulmfit_fastai_bidir_end_to_end():
""" Test ulmfit with sentencepiece tokenizer on small wikipedia dataset.
"""