Merge pull request flairNLP#1335 from flairNLP/optimizations

relax version requirements for Colab installation

.travis.yml

@@ -10,5 +10,4 @@ before_script: cd tests
 script:
   - pip freeze
   - 'if [ "$TRAVIS_PULL_REQUEST" != "false" ]; then pytest --runintegration; fi'
-  - 'if [ "$TRAVIS_PULL_REQUEST" = "false" ]; then pytest; fi'
-  - 'if [ "$TRAVIS_PULL_REQUEST" != "false" ]; then pip install black; black --check .; fi'
+  - 'if [ "$TRAVIS_PULL_REQUEST" = "false" ]; then pytest; fi'

flair/visual/training_curves.py

@@ -6,18 +6,8 @@
 import numpy as np
 import csv
 
-import matplotlib
 import math
 
-
-# change from Agg to TkAgg for interactive mode
-try:
-    # change from Agg to TkAgg for interactive mode
-    matplotlib.use("TkAgg")
-except:
-    pass
-
-
 import matplotlib.pyplot as plt
 
 

requirements.txt

@@ -1,12 +1,12 @@
-python-dateutil==2.8.0
+python-dateutil>=2.8.1
 torch>=1.1.0
 gensim>=3.4.0
-pytest>=3.6.4
+pytest>=5.3.2
 tqdm>=4.26.0
 segtok>=1.5.7
 matplotlib>=2.2.3
 mpld3==0.3
-scikit-learn==0.21.3
+scikit-learn>=0.21.3
 sqlitedict>=1.6.0
 deprecated>=1.2.4
 hyperopt>=0.1.1
@@ -15,4 +15,4 @@ bpemb>=0.2.9
 regex
 tabulate
 urllib3<1.25,>=1.20
-langdetect
+langdetect

setup.py

@@ -10,8 +10,8 @@
     long_description=open("README.md", encoding="utf-8").read(),
     long_description_content_type="text/markdown",
     author="Alan Akbik",
-    author_email="alan.akbik@zalando.de",
-    url="https://github.com/zalandoresearch/flair",
+    author_email="alan.akbik@gmail.com",
+    url="https://github.com/flairNLP/flair",
     packages=find_packages(exclude="tests"),  # same as name
     license="MIT",
     install_requires=required,

tests/test_embeddings.py

@@ -42,6 +42,7 @@ def test_keep_batch_order():
 
     assert torch.norm(sentences_1[0].embedding - sentences_2[1].embedding) == 0.0
     assert torch.norm(sentences_1[0].embedding - sentences_2[1].embedding) == 0.0
+    del embeddings
 
 
 @pytest.mark.integration
@@ -58,6 +59,7 @@ def test_stacked_embeddings():
         token.clear_embeddings()
 
         assert len(token.get_embedding()) == 0
+    del embeddings
 
 
 @pytest.mark.integration
@@ -97,6 +99,7 @@ def test_fine_tunable_flair_embedding():
     sentence.clear_embeddings()
 
     assert len(sentence.get_embedding()) == 0
+    del embeddings
 
 
 @pytest.mark.integration
@@ -115,6 +118,7 @@ def test_document_lstm_embeddings():
     sentence.clear_embeddings()
 
     assert len(sentence.get_embedding()) == 0
+    del embeddings
 
 
 @pytest.mark.integration
@@ -133,6 +137,7 @@ def test_document_bidirectional_lstm_embeddings():
     sentence.clear_embeddings()
 
     assert len(sentence.get_embedding()) == 0
+    del embeddings
 
 
 @pytest.mark.integration
@@ -151,6 +156,7 @@ def test_document_pool_embeddings():
         sentence.clear_embeddings()
 
         assert len(sentence.get_embedding()) == 0
+        del embeddings
 
 
 @pytest.mark.integration
@@ -169,6 +175,7 @@ def test_document_pool_embeddings_nonlinear():
         sentence.clear_embeddings()
 
         assert len(sentence.get_embedding()) == 0
+        del embeddings
 
 
 def init_document_embeddings():
@@ -193,6 +200,7 @@ def load_and_apply_word_embeddings(emb_type: str):
         token.clear_embeddings()
 
         assert len(token.get_embedding()) == 0
+    del embeddings
 
 
 def load_and_apply_char_lm_embeddings(emb_type: str):
@@ -207,3 +215,4 @@ def load_and_apply_char_lm_embeddings(emb_type: str):
         token.clear_embeddings()
 
         assert len(token.get_embedding()) == 0
+    del embeddings

tests/test_hyperparameter.py

@@ -13,6 +13,8 @@
 )
 import flair.datasets
 
+glove_embedding: WordEmbeddings = WordEmbeddings("glove")
+
 
 @pytest.mark.integration
 def test_sequence_tagger_param_selector(results_base_path, tasks_base_path):
@@ -27,16 +29,7 @@ def test_sequence_tagger_param_selector(results_base_path, tasks_base_path):
     search_space.add(
         Parameter.EMBEDDINGS,
         hp.choice,
-        options=[
-            StackedEmbeddings([WordEmbeddings("glove")]),
-            StackedEmbeddings(
-                [
-                    WordEmbeddings("glove"),
-                    FlairEmbeddings("news-forward-fast"),
-                    FlairEmbeddings("news-backward-fast"),
-                ]
-            ),
-        ],
+        options=[StackedEmbeddings([glove_embedding])],
     )
     search_space.add(Parameter.USE_CRF, hp.choice, options=[True, False])
     search_space.add(Parameter.DROPOUT, hp.uniform, low=0.25, high=0.75)
@@ -63,14 +56,13 @@ def test_sequence_tagger_param_selector(results_base_path, tasks_base_path):
 
     # clean up results directory
     shutil.rmtree(results_base_path)
+    del optimizer, search_space
 
 
 @pytest.mark.integration
 def test_text_classifier_param_selector(results_base_path, tasks_base_path):
     corpus = flair.datasets.ClassificationCorpus(tasks_base_path / "imdb")
 
-    glove_embedding: WordEmbeddings = WordEmbeddings("glove")
-
     search_space = SearchSpace()
 
     # document embeddings parameter
@@ -97,3 +89,4 @@ def test_text_classifier_param_selector(results_base_path, tasks_base_path):
 
     # clean up results directory
     shutil.rmtree(results_base_path)
+    del param_selector, search_space

tests/test_language_model.py

@@ -35,6 +35,7 @@ def test_generate_text_with_small_temperatures():
     )
     assert text is not None
     assert len(text) >= 100
+    del language_model
 
 
 def test_compute_perplexity():
@@ -70,3 +71,4 @@ def test_compute_perplexity():
     print(f'"{ungrammatical}" - perplexity is {perplexity_ungramamtical_sentence}')
 
     assert perplexity_gramamtical_sentence < perplexity_ungramamtical_sentence
+    del language_model