spaCy/spacy/tests/test_models.py

271 lines
8.8 KiB
Python
Raw Normal View History

from typing import List
import pytest
from thinc.api import fix_random_seed, Adam, set_dropout_rate
Add SpanCategorizer component (#6747) * Draft spancat model * Add spancat model * Add test for extract_spans * Add extract_spans layer * Upd extract_spans * Add spancat model * Add test for spancat model * Upd spancat model * Update spancat component * Upd spancat * Update spancat model * Add quick spancat test * Import SpanCategorizer * Fix SpanCategorizer component * Import SpanGroup * Fix span extraction * Fix import * Fix import * Upd model * Update spancat models * Add scoring, update defaults * Update and add docs * Fix type * Update spacy/ml/extract_spans.py * Auto-format and fix import * Fix comment * Fix type * Fix type * Update website/docs/api/spancategorizer.md * Fix comment Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Better defense Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Fix labels list Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/ml/extract_spans.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/pipeline/spancat.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Set annotations during update * Set annotations in spancat * fix imports in test * Update spacy/pipeline/spancat.py * replace MaxoutLogistic with LinearLogistic * fix config * various small fixes * remove set_annotations parameter in update * use our beloved tupley format with recent support for doc.spans * bugfix to allow renaming the default span_key (scores weren't showing up) * use different key in docs example * change defaults to better-working parameters from project (WIP) * register spacy.extract_spans.v1 for legacy purposes * Upd dev version so can build wheel * layers instead of architectures for smaller building blocks * Update website/docs/api/spancategorizer.md Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Update website/docs/api/spancategorizer.md Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Include additional scores from overrides in combined score weights * Parameterize spans key in scoring Parameterize the `SpanCategorizer` `spans_key` for scoring purposes so that it's possible to evaluate multiple `spancat` components in the same pipeline. * Use the (intentionally very short) default spans key `sc` in the `SpanCategorizer` * Adjust the default score weights to include the default key * Adjust the scorer to use `spans_{spans_key}` as the prefix for the returned score * Revert addition of `attr_name` argument to `score_spans` and adjust the key in the `getter` instead. Note that for `spancat` components with a custom `span_key`, the score weights currently need to be modified manually in `[training.score_weights]` for them to be available during training. To suppress the default score weights `spans_sc_p/r/f` during training, set them to `null` in `[training.score_weights]`. * Update website/docs/api/scorer.md * Fix scorer for spans key containing underscore * Increment version * Add Spans to Evaluate CLI (#8439) * Add Spans to Evaluate CLI * Change to spans_key * Add spans per_type output Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Fix spancat GPU issues (#8455) * Fix GPU issues * Require thinc >=8.0.6 * Switch to glorot_uniform_init * Fix and test ngram suggester * Include final ngram in doc for all sizes * Fix ngrams for docs of the same length as ngram size * Handle batches of docs that result in no ngrams * Add tests Co-authored-by: Ines Montani <ines@ines.io> Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com> Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> Co-authored-by: Nirant <NirantK@users.noreply.github.com>
2021-06-24 13:35:27 +03:00
from thinc.api import Ragged, reduce_mean, Logistic, chain, Relu
from numpy.testing import assert_array_equal, assert_array_almost_equal
import numpy
2020-07-28 23:43:19 +03:00
from spacy.ml.models import build_Tok2Vec_model, MultiHashEmbed, MaxoutWindowEncoder
from spacy.ml.models import build_bow_text_classifier, build_simple_cnn_text_classifier
Add SpanCategorizer component (#6747) * Draft spancat model * Add spancat model * Add test for extract_spans * Add extract_spans layer * Upd extract_spans * Add spancat model * Add test for spancat model * Upd spancat model * Update spancat component * Upd spancat * Update spancat model * Add quick spancat test * Import SpanCategorizer * Fix SpanCategorizer component * Import SpanGroup * Fix span extraction * Fix import * Fix import * Upd model * Update spancat models * Add scoring, update defaults * Update and add docs * Fix type * Update spacy/ml/extract_spans.py * Auto-format and fix import * Fix comment * Fix type * Fix type * Update website/docs/api/spancategorizer.md * Fix comment Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Better defense Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Fix labels list Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/ml/extract_spans.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/pipeline/spancat.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Set annotations during update * Set annotations in spancat * fix imports in test * Update spacy/pipeline/spancat.py * replace MaxoutLogistic with LinearLogistic * fix config * various small fixes * remove set_annotations parameter in update * use our beloved tupley format with recent support for doc.spans * bugfix to allow renaming the default span_key (scores weren't showing up) * use different key in docs example * change defaults to better-working parameters from project (WIP) * register spacy.extract_spans.v1 for legacy purposes * Upd dev version so can build wheel * layers instead of architectures for smaller building blocks * Update website/docs/api/spancategorizer.md Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Update website/docs/api/spancategorizer.md Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Include additional scores from overrides in combined score weights * Parameterize spans key in scoring Parameterize the `SpanCategorizer` `spans_key` for scoring purposes so that it's possible to evaluate multiple `spancat` components in the same pipeline. * Use the (intentionally very short) default spans key `sc` in the `SpanCategorizer` * Adjust the default score weights to include the default key * Adjust the scorer to use `spans_{spans_key}` as the prefix for the returned score * Revert addition of `attr_name` argument to `score_spans` and adjust the key in the `getter` instead. Note that for `spancat` components with a custom `span_key`, the score weights currently need to be modified manually in `[training.score_weights]` for them to be available during training. To suppress the default score weights `spans_sc_p/r/f` during training, set them to `null` in `[training.score_weights]`. * Update website/docs/api/scorer.md * Fix scorer for spans key containing underscore * Increment version * Add Spans to Evaluate CLI (#8439) * Add Spans to Evaluate CLI * Change to spans_key * Add spans per_type output Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Fix spancat GPU issues (#8455) * Fix GPU issues * Require thinc >=8.0.6 * Switch to glorot_uniform_init * Fix and test ngram suggester * Include final ngram in doc for all sizes * Fix ngrams for docs of the same length as ngram size * Handle batches of docs that result in no ngrams * Add tests Co-authored-by: Ines Montani <ines@ines.io> Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com> Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> Co-authored-by: Nirant <NirantK@users.noreply.github.com>
2021-06-24 13:35:27 +03:00
from spacy.ml.models import build_spancat_model
from spacy.ml.staticvectors import StaticVectors
Add SpanCategorizer component (#6747) * Draft spancat model * Add spancat model * Add test for extract_spans * Add extract_spans layer * Upd extract_spans * Add spancat model * Add test for spancat model * Upd spancat model * Update spancat component * Upd spancat * Update spancat model * Add quick spancat test * Import SpanCategorizer * Fix SpanCategorizer component * Import SpanGroup * Fix span extraction * Fix import * Fix import * Upd model * Update spancat models * Add scoring, update defaults * Update and add docs * Fix type * Update spacy/ml/extract_spans.py * Auto-format and fix import * Fix comment * Fix type * Fix type * Update website/docs/api/spancategorizer.md * Fix comment Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Better defense Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Fix labels list Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/ml/extract_spans.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/pipeline/spancat.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Set annotations during update * Set annotations in spancat * fix imports in test * Update spacy/pipeline/spancat.py * replace MaxoutLogistic with LinearLogistic * fix config * various small fixes * remove set_annotations parameter in update * use our beloved tupley format with recent support for doc.spans * bugfix to allow renaming the default span_key (scores weren't showing up) * use different key in docs example * change defaults to better-working parameters from project (WIP) * register spacy.extract_spans.v1 for legacy purposes * Upd dev version so can build wheel * layers instead of architectures for smaller building blocks * Update website/docs/api/spancategorizer.md Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Update website/docs/api/spancategorizer.md Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Include additional scores from overrides in combined score weights * Parameterize spans key in scoring Parameterize the `SpanCategorizer` `spans_key` for scoring purposes so that it's possible to evaluate multiple `spancat` components in the same pipeline. * Use the (intentionally very short) default spans key `sc` in the `SpanCategorizer` * Adjust the default score weights to include the default key * Adjust the scorer to use `spans_{spans_key}` as the prefix for the returned score * Revert addition of `attr_name` argument to `score_spans` and adjust the key in the `getter` instead. Note that for `spancat` components with a custom `span_key`, the score weights currently need to be modified manually in `[training.score_weights]` for them to be available during training. To suppress the default score weights `spans_sc_p/r/f` during training, set them to `null` in `[training.score_weights]`. * Update website/docs/api/scorer.md * Fix scorer for spans key containing underscore * Increment version * Add Spans to Evaluate CLI (#8439) * Add Spans to Evaluate CLI * Change to spans_key * Add spans per_type output Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Fix spancat GPU issues (#8455) * Fix GPU issues * Require thinc >=8.0.6 * Switch to glorot_uniform_init * Fix and test ngram suggester * Include final ngram in doc for all sizes * Fix ngrams for docs of the same length as ngram size * Handle batches of docs that result in no ngrams * Add tests Co-authored-by: Ines Montani <ines@ines.io> Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com> Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> Co-authored-by: Nirant <NirantK@users.noreply.github.com>
2021-06-24 13:35:27 +03:00
from spacy.ml.extract_spans import extract_spans, _get_span_indices
from spacy.lang.en import English
from spacy.lang.en.examples import sentences as EN_SENTENCES
def get_textcat_bow_kwargs():
2020-07-28 23:43:19 +03:00
return {
"exclusive_classes": True,
2020-07-28 23:43:19 +03:00
"ngram_size": 1,
"no_output_layer": False,
"nO": 34,
2020-07-28 23:43:19 +03:00
}
2020-08-05 17:00:59 +03:00
2020-07-28 23:43:19 +03:00
def get_textcat_cnn_kwargs():
return {"tok2vec": test_tok2vec(), "exclusive_classes": False, "nO": 13}
2020-07-28 23:43:19 +03:00
2020-08-05 17:00:59 +03:00
def get_all_params(model):
params = []
for node in model.walk():
for name in node.param_names:
params.append(node.get_param(name).ravel())
return node.ops.xp.concatenate(params)
def get_docs():
nlp = English()
return list(nlp.pipe(EN_SENTENCES + [" ".join(EN_SENTENCES)]))
def get_gradient(model, Y):
if isinstance(Y, model.ops.xp.ndarray):
dY = model.ops.alloc(Y.shape, dtype=Y.dtype)
dY += model.ops.xp.random.uniform(-1.0, 1.0, Y.shape)
return dY
elif isinstance(Y, List):
return [get_gradient(model, y) for y in Y]
else:
raise ValueError(f"Could not get gradient for type {type(Y)}")
2020-07-28 23:43:19 +03:00
def get_tok2vec_kwargs():
# This actually creates models, so seems best to put it in a function.
return {
"embed": MultiHashEmbed(
width=32,
2020-10-05 20:57:45 +03:00
rows=[500, 500, 500],
attrs=["NORM", "PREFIX", "SHAPE"],
2020-10-05 22:58:18 +03:00
include_static_vectors=False,
2020-07-28 23:43:19 +03:00
),
"encode": MaxoutWindowEncoder(
2020-09-29 22:39:28 +03:00
width=32, depth=2, maxout_pieces=2, window_size=1
2020-08-05 17:00:59 +03:00
),
2020-07-28 23:43:19 +03:00
}
Refactor pipeline components, config and language data (#5759) * Update with WIP * Update with WIP * Update with pipeline serialization * Update types and pipe factories * Add deep merge, tidy up and add tests * Fix pipe creation from config * Don't validate default configs on load * Update spacy/language.py Co-authored-by: Ines Montani <ines@ines.io> * Adjust factory/component meta error * Clean up factory args and remove defaults * Add test for failing empty dict defaults * Update pipeline handling and methods * provide KB as registry function instead of as object * small change in test to make functionality more clear * update example script for EL configuration * Fix typo * Simplify test * Simplify test * splitting pipes.pyx into separate files * moving default configs to each component file * fix batch_size type * removing default values from component constructors where possible (TODO: test 4725) * skip instead of xfail * Add test for config -> nlp with multiple instances * pipeline.pipes -> pipeline.pipe * Tidy up, document, remove kwargs * small cleanup/generalization for Tok2VecListener * use DEFAULT_UPSTREAM field * revert to avoid circular imports * Fix tests * Replace deprecated arg * Make model dirs require config * fix pickling of keyword-only arguments in constructor * WIP: clean up and integrate full config * Add helper to handle function args more reliably Now also includes keyword-only args * Fix config composition and serialization * Improve config debugging and add visual diff * Remove unused defaults and fix type * Remove pipeline and factories from meta * Update spacy/default_config.cfg Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/default_config.cfg * small UX edits * avoid printing stack trace for debug CLI commands * Add support for language-specific factories * specify the section of the config which holds the model to debug * WIP: add Language.from_config * Update with language data refactor WIP * Auto-format * Add backwards-compat handling for Language.factories * Update morphologizer.pyx * Fix morphologizer * Update and simplify lemmatizers * Fix Japanese tests * Port over tagger changes * Fix Chinese and tests * Update to latest Thinc * WIP: xfail first Russian lemmatizer test * Fix component-specific overrides * fix nO for output layers in debug_model * Fix default value * Fix tests and don't pass objects in config * Fix deep merging * Fix lemma lookup data registry Only load the lookups if an entry is available in the registry (and if spacy-lookups-data is installed) * Add types * Add Vocab.from_config * Fix typo * Fix tests * Make config copying more elegant * Fix pipe analysis * Fix lemmatizers and is_base_form * WIP: move language defaults to config * Fix morphology type * Fix vocab * Remove comment * Update to latest Thinc * Add morph rules to config * Tidy up * Remove set_morphology option from tagger factory * Hack use_gpu * Move [pipeline] to top-level block and make [nlp.pipeline] list Allows separating component blocks from component order – otherwise, ordering the config would mean a changed component order, which is bad. Also allows initial config to define more components and not use all of them * Fix use_gpu and resume in CLI * Auto-format * Remove resume from config * Fix formatting and error * [pipeline] -> [components] * Fix types * Fix tagger test: requires set_morphology? Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com> Co-authored-by: Matthew Honnibal <honnibal+gh@gmail.com>
2020-07-22 14:42:59 +03:00
def test_tok2vec():
2020-07-28 23:43:19 +03:00
return build_Tok2Vec_model(**get_tok2vec_kwargs())
def test_multi_hash_embed():
embed = MultiHashEmbed(
width=32,
2020-10-05 20:57:45 +03:00
rows=[500, 500, 500],
attrs=["NORM", "PREFIX", "SHAPE"],
2020-10-05 22:58:18 +03:00
include_static_vectors=False,
)
hash_embeds = [node for node in embed.walk() if node.name == "hashembed"]
assert len(hash_embeds) == 3
# Check they look at different columns.
assert list(sorted(he.attrs["column"] for he in hash_embeds)) == [0, 1, 2]
# Check they use different seeds
assert len(set(he.attrs["seed"] for he in hash_embeds)) == 3
# Check they all have the same number of rows
assert [he.get_dim("nV") for he in hash_embeds] == [500, 500, 500]
# Now try with different row factors
embed = MultiHashEmbed(
width=32,
2020-10-05 20:57:45 +03:00
rows=[1000, 50, 250],
attrs=["NORM", "PREFIX", "SHAPE"],
2020-10-05 22:58:18 +03:00
include_static_vectors=False,
)
hash_embeds = [node for node in embed.walk() if node.name == "hashembed"]
assert [he.get_dim("nV") for he in hash_embeds] == [1000, 50, 250]
2020-10-05 22:58:18 +03:00
@pytest.mark.parametrize(
"seed,model_func,kwargs",
[
2020-07-28 23:43:19 +03:00
(0, build_Tok2Vec_model, get_tok2vec_kwargs()),
(0, build_bow_text_classifier, get_textcat_bow_kwargs()),
2020-07-28 23:43:19 +03:00
(0, build_simple_cnn_text_classifier, get_textcat_cnn_kwargs()),
],
)
def test_models_initialize_consistently(seed, model_func, kwargs):
fix_random_seed(seed)
model1 = model_func(**kwargs)
model1.initialize()
fix_random_seed(seed)
model2 = model_func(**kwargs)
model2.initialize()
params1 = get_all_params(model1)
params2 = get_all_params(model2)
assert_array_equal(model1.ops.to_numpy(params1), model2.ops.to_numpy(params2))
@pytest.mark.parametrize(
"seed,model_func,kwargs,get_X",
[
2020-07-28 23:43:19 +03:00
(0, build_Tok2Vec_model, get_tok2vec_kwargs(), get_docs),
(0, build_bow_text_classifier, get_textcat_bow_kwargs(), get_docs),
2020-07-28 23:43:19 +03:00
(0, build_simple_cnn_text_classifier, get_textcat_cnn_kwargs(), get_docs),
],
)
def test_models_predict_consistently(seed, model_func, kwargs, get_X):
fix_random_seed(seed)
model1 = model_func(**kwargs).initialize()
Y1 = model1.predict(get_X())
fix_random_seed(seed)
model2 = model_func(**kwargs).initialize()
Y2 = model2.predict(get_X())
if model1.has_ref("tok2vec"):
tok2vec1 = model1.get_ref("tok2vec").predict(get_X())
tok2vec2 = model2.get_ref("tok2vec").predict(get_X())
for i in range(len(tok2vec1)):
for j in range(len(tok2vec1[i])):
Refactor pipeline components, config and language data (#5759) * Update with WIP * Update with WIP * Update with pipeline serialization * Update types and pipe factories * Add deep merge, tidy up and add tests * Fix pipe creation from config * Don't validate default configs on load * Update spacy/language.py Co-authored-by: Ines Montani <ines@ines.io> * Adjust factory/component meta error * Clean up factory args and remove defaults * Add test for failing empty dict defaults * Update pipeline handling and methods * provide KB as registry function instead of as object * small change in test to make functionality more clear * update example script for EL configuration * Fix typo * Simplify test * Simplify test * splitting pipes.pyx into separate files * moving default configs to each component file * fix batch_size type * removing default values from component constructors where possible (TODO: test 4725) * skip instead of xfail * Add test for config -> nlp with multiple instances * pipeline.pipes -> pipeline.pipe * Tidy up, document, remove kwargs * small cleanup/generalization for Tok2VecListener * use DEFAULT_UPSTREAM field * revert to avoid circular imports * Fix tests * Replace deprecated arg * Make model dirs require config * fix pickling of keyword-only arguments in constructor * WIP: clean up and integrate full config * Add helper to handle function args more reliably Now also includes keyword-only args * Fix config composition and serialization * Improve config debugging and add visual diff * Remove unused defaults and fix type * Remove pipeline and factories from meta * Update spacy/default_config.cfg Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/default_config.cfg * small UX edits * avoid printing stack trace for debug CLI commands * Add support for language-specific factories * specify the section of the config which holds the model to debug * WIP: add Language.from_config * Update with language data refactor WIP * Auto-format * Add backwards-compat handling for Language.factories * Update morphologizer.pyx * Fix morphologizer * Update and simplify lemmatizers * Fix Japanese tests * Port over tagger changes * Fix Chinese and tests * Update to latest Thinc * WIP: xfail first Russian lemmatizer test * Fix component-specific overrides * fix nO for output layers in debug_model * Fix default value * Fix tests and don't pass objects in config * Fix deep merging * Fix lemma lookup data registry Only load the lookups if an entry is available in the registry (and if spacy-lookups-data is installed) * Add types * Add Vocab.from_config * Fix typo * Fix tests * Make config copying more elegant * Fix pipe analysis * Fix lemmatizers and is_base_form * WIP: move language defaults to config * Fix morphology type * Fix vocab * Remove comment * Update to latest Thinc * Add morph rules to config * Tidy up * Remove set_morphology option from tagger factory * Hack use_gpu * Move [pipeline] to top-level block and make [nlp.pipeline] list Allows separating component blocks from component order – otherwise, ordering the config would mean a changed component order, which is bad. Also allows initial config to define more components and not use all of them * Fix use_gpu and resume in CLI * Auto-format * Remove resume from config * Fix formatting and error * [pipeline] -> [components] * Fix types * Fix tagger test: requires set_morphology? Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com> Co-authored-by: Matthew Honnibal <honnibal+gh@gmail.com>
2020-07-22 14:42:59 +03:00
assert_array_equal(
numpy.asarray(model1.ops.to_numpy(tok2vec1[i][j])),
numpy.asarray(model2.ops.to_numpy(tok2vec2[i][j])),
Refactor pipeline components, config and language data (#5759) * Update with WIP * Update with WIP * Update with pipeline serialization * Update types and pipe factories * Add deep merge, tidy up and add tests * Fix pipe creation from config * Don't validate default configs on load * Update spacy/language.py Co-authored-by: Ines Montani <ines@ines.io> * Adjust factory/component meta error * Clean up factory args and remove defaults * Add test for failing empty dict defaults * Update pipeline handling and methods * provide KB as registry function instead of as object * small change in test to make functionality more clear * update example script for EL configuration * Fix typo * Simplify test * Simplify test * splitting pipes.pyx into separate files * moving default configs to each component file * fix batch_size type * removing default values from component constructors where possible (TODO: test 4725) * skip instead of xfail * Add test for config -> nlp with multiple instances * pipeline.pipes -> pipeline.pipe * Tidy up, document, remove kwargs * small cleanup/generalization for Tok2VecListener * use DEFAULT_UPSTREAM field * revert to avoid circular imports * Fix tests * Replace deprecated arg * Make model dirs require config * fix pickling of keyword-only arguments in constructor * WIP: clean up and integrate full config * Add helper to handle function args more reliably Now also includes keyword-only args * Fix config composition and serialization * Improve config debugging and add visual diff * Remove unused defaults and fix type * Remove pipeline and factories from meta * Update spacy/default_config.cfg Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/default_config.cfg * small UX edits * avoid printing stack trace for debug CLI commands * Add support for language-specific factories * specify the section of the config which holds the model to debug * WIP: add Language.from_config * Update with language data refactor WIP * Auto-format * Add backwards-compat handling for Language.factories * Update morphologizer.pyx * Fix morphologizer * Update and simplify lemmatizers * Fix Japanese tests * Port over tagger changes * Fix Chinese and tests * Update to latest Thinc * WIP: xfail first Russian lemmatizer test * Fix component-specific overrides * fix nO for output layers in debug_model * Fix default value * Fix tests and don't pass objects in config * Fix deep merging * Fix lemma lookup data registry Only load the lookups if an entry is available in the registry (and if spacy-lookups-data is installed) * Add types * Add Vocab.from_config * Fix typo * Fix tests * Make config copying more elegant * Fix pipe analysis * Fix lemmatizers and is_base_form * WIP: move language defaults to config * Fix morphology type * Fix vocab * Remove comment * Update to latest Thinc * Add morph rules to config * Tidy up * Remove set_morphology option from tagger factory * Hack use_gpu * Move [pipeline] to top-level block and make [nlp.pipeline] list Allows separating component blocks from component order – otherwise, ordering the config would mean a changed component order, which is bad. Also allows initial config to define more components and not use all of them * Fix use_gpu and resume in CLI * Auto-format * Remove resume from config * Fix formatting and error * [pipeline] -> [components] * Fix types * Fix tagger test: requires set_morphology? Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com> Co-authored-by: Matthew Honnibal <honnibal+gh@gmail.com>
2020-07-22 14:42:59 +03:00
)
try:
Y1 = model1.ops.to_numpy(Y1)
Y2 = model2.ops.to_numpy(Y2)
except Exception:
pass
if isinstance(Y1, numpy.ndarray):
assert_array_equal(Y1, Y2)
elif isinstance(Y1, List):
assert len(Y1) == len(Y2)
for y1, y2 in zip(Y1, Y2):
try:
y1 = model1.ops.to_numpy(y1)
y2 = model2.ops.to_numpy(y2)
except Exception:
pass
assert_array_equal(y1, y2)
else:
raise ValueError(f"Could not compare type {type(Y1)}")
@pytest.mark.parametrize(
"seed,dropout,model_func,kwargs,get_X",
[
2020-07-28 23:43:19 +03:00
(0, 0.2, build_Tok2Vec_model, get_tok2vec_kwargs(), get_docs),
(0, 0.2, build_bow_text_classifier, get_textcat_bow_kwargs(), get_docs),
2020-07-28 23:43:19 +03:00
(0, 0.2, build_simple_cnn_text_classifier, get_textcat_cnn_kwargs(), get_docs),
],
)
def test_models_update_consistently(seed, dropout, model_func, kwargs, get_X):
def get_updated_model():
fix_random_seed(seed)
optimizer = Adam(0.001)
model = model_func(**kwargs).initialize()
initial_params = get_all_params(model)
set_dropout_rate(model, dropout)
for _ in range(5):
Y, get_dX = model.begin_update(get_X())
dY = get_gradient(model, Y)
Refactor pipeline components, config and language data (#5759) * Update with WIP * Update with WIP * Update with pipeline serialization * Update types and pipe factories * Add deep merge, tidy up and add tests * Fix pipe creation from config * Don't validate default configs on load * Update spacy/language.py Co-authored-by: Ines Montani <ines@ines.io> * Adjust factory/component meta error * Clean up factory args and remove defaults * Add test for failing empty dict defaults * Update pipeline handling and methods * provide KB as registry function instead of as object * small change in test to make functionality more clear * update example script for EL configuration * Fix typo * Simplify test * Simplify test * splitting pipes.pyx into separate files * moving default configs to each component file * fix batch_size type * removing default values from component constructors where possible (TODO: test 4725) * skip instead of xfail * Add test for config -> nlp with multiple instances * pipeline.pipes -> pipeline.pipe * Tidy up, document, remove kwargs * small cleanup/generalization for Tok2VecListener * use DEFAULT_UPSTREAM field * revert to avoid circular imports * Fix tests * Replace deprecated arg * Make model dirs require config * fix pickling of keyword-only arguments in constructor * WIP: clean up and integrate full config * Add helper to handle function args more reliably Now also includes keyword-only args * Fix config composition and serialization * Improve config debugging and add visual diff * Remove unused defaults and fix type * Remove pipeline and factories from meta * Update spacy/default_config.cfg Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/default_config.cfg * small UX edits * avoid printing stack trace for debug CLI commands * Add support for language-specific factories * specify the section of the config which holds the model to debug * WIP: add Language.from_config * Update with language data refactor WIP * Auto-format * Add backwards-compat handling for Language.factories * Update morphologizer.pyx * Fix morphologizer * Update and simplify lemmatizers * Fix Japanese tests * Port over tagger changes * Fix Chinese and tests * Update to latest Thinc * WIP: xfail first Russian lemmatizer test * Fix component-specific overrides * fix nO for output layers in debug_model * Fix default value * Fix tests and don't pass objects in config * Fix deep merging * Fix lemma lookup data registry Only load the lookups if an entry is available in the registry (and if spacy-lookups-data is installed) * Add types * Add Vocab.from_config * Fix typo * Fix tests * Make config copying more elegant * Fix pipe analysis * Fix lemmatizers and is_base_form * WIP: move language defaults to config * Fix morphology type * Fix vocab * Remove comment * Update to latest Thinc * Add morph rules to config * Tidy up * Remove set_morphology option from tagger factory * Hack use_gpu * Move [pipeline] to top-level block and make [nlp.pipeline] list Allows separating component blocks from component order – otherwise, ordering the config would mean a changed component order, which is bad. Also allows initial config to define more components and not use all of them * Fix use_gpu and resume in CLI * Auto-format * Remove resume from config * Fix formatting and error * [pipeline] -> [components] * Fix types * Fix tagger test: requires set_morphology? Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com> Co-authored-by: Matthew Honnibal <honnibal+gh@gmail.com>
2020-07-22 14:42:59 +03:00
get_dX(dY)
model.finish_update(optimizer)
updated_params = get_all_params(model)
with pytest.raises(AssertionError):
assert_array_equal(
model.ops.to_numpy(initial_params), model.ops.to_numpy(updated_params)
)
return model
model1 = get_updated_model()
model2 = get_updated_model()
assert_array_almost_equal(
model1.ops.to_numpy(get_all_params(model1)),
model2.ops.to_numpy(get_all_params(model2)),
)
2020-10-10 20:14:48 +03:00
@pytest.mark.parametrize("model_func,kwargs", [(StaticVectors, {"nO": 128, "nM": 300})])
def test_empty_docs(model_func, kwargs):
nlp = English()
model = model_func(**kwargs).initialize()
# Test the layer can be called successfully with 0, 1 and 2 empty docs.
for n_docs in range(3):
docs = [nlp("") for _ in range(n_docs)]
# Test predict
2020-10-10 20:14:48 +03:00
model.predict(docs)
# Test backprop
output, backprop = model.begin_update(docs)
2020-10-10 20:14:48 +03:00
backprop(output)
Add SpanCategorizer component (#6747) * Draft spancat model * Add spancat model * Add test for extract_spans * Add extract_spans layer * Upd extract_spans * Add spancat model * Add test for spancat model * Upd spancat model * Update spancat component * Upd spancat * Update spancat model * Add quick spancat test * Import SpanCategorizer * Fix SpanCategorizer component * Import SpanGroup * Fix span extraction * Fix import * Fix import * Upd model * Update spancat models * Add scoring, update defaults * Update and add docs * Fix type * Update spacy/ml/extract_spans.py * Auto-format and fix import * Fix comment * Fix type * Fix type * Update website/docs/api/spancategorizer.md * Fix comment Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Better defense Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Fix labels list Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/ml/extract_spans.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/pipeline/spancat.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Set annotations during update * Set annotations in spancat * fix imports in test * Update spacy/pipeline/spancat.py * replace MaxoutLogistic with LinearLogistic * fix config * various small fixes * remove set_annotations parameter in update * use our beloved tupley format with recent support for doc.spans * bugfix to allow renaming the default span_key (scores weren't showing up) * use different key in docs example * change defaults to better-working parameters from project (WIP) * register spacy.extract_spans.v1 for legacy purposes * Upd dev version so can build wheel * layers instead of architectures for smaller building blocks * Update website/docs/api/spancategorizer.md Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Update website/docs/api/spancategorizer.md Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Include additional scores from overrides in combined score weights * Parameterize spans key in scoring Parameterize the `SpanCategorizer` `spans_key` for scoring purposes so that it's possible to evaluate multiple `spancat` components in the same pipeline. * Use the (intentionally very short) default spans key `sc` in the `SpanCategorizer` * Adjust the default score weights to include the default key * Adjust the scorer to use `spans_{spans_key}` as the prefix for the returned score * Revert addition of `attr_name` argument to `score_spans` and adjust the key in the `getter` instead. Note that for `spancat` components with a custom `span_key`, the score weights currently need to be modified manually in `[training.score_weights]` for them to be available during training. To suppress the default score weights `spans_sc_p/r/f` during training, set them to `null` in `[training.score_weights]`. * Update website/docs/api/scorer.md * Fix scorer for spans key containing underscore * Increment version * Add Spans to Evaluate CLI (#8439) * Add Spans to Evaluate CLI * Change to spans_key * Add spans per_type output Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Fix spancat GPU issues (#8455) * Fix GPU issues * Require thinc >=8.0.6 * Switch to glorot_uniform_init * Fix and test ngram suggester * Include final ngram in doc for all sizes * Fix ngrams for docs of the same length as ngram size * Handle batches of docs that result in no ngrams * Add tests Co-authored-by: Ines Montani <ines@ines.io> Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com> Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> Co-authored-by: Nirant <NirantK@users.noreply.github.com>
2021-06-24 13:35:27 +03:00
def test_init_extract_spans():
2021-07-18 08:44:56 +03:00
extract_spans().initialize()
Add SpanCategorizer component (#6747) * Draft spancat model * Add spancat model * Add test for extract_spans * Add extract_spans layer * Upd extract_spans * Add spancat model * Add test for spancat model * Upd spancat model * Update spancat component * Upd spancat * Update spancat model * Add quick spancat test * Import SpanCategorizer * Fix SpanCategorizer component * Import SpanGroup * Fix span extraction * Fix import * Fix import * Upd model * Update spancat models * Add scoring, update defaults * Update and add docs * Fix type * Update spacy/ml/extract_spans.py * Auto-format and fix import * Fix comment * Fix type * Fix type * Update website/docs/api/spancategorizer.md * Fix comment Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Better defense Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Fix labels list Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/ml/extract_spans.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Update spacy/pipeline/spancat.py Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> * Set annotations during update * Set annotations in spancat * fix imports in test * Update spacy/pipeline/spancat.py * replace MaxoutLogistic with LinearLogistic * fix config * various small fixes * remove set_annotations parameter in update * use our beloved tupley format with recent support for doc.spans * bugfix to allow renaming the default span_key (scores weren't showing up) * use different key in docs example * change defaults to better-working parameters from project (WIP) * register spacy.extract_spans.v1 for legacy purposes * Upd dev version so can build wheel * layers instead of architectures for smaller building blocks * Update website/docs/api/spancategorizer.md Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Update website/docs/api/spancategorizer.md Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Include additional scores from overrides in combined score weights * Parameterize spans key in scoring Parameterize the `SpanCategorizer` `spans_key` for scoring purposes so that it's possible to evaluate multiple `spancat` components in the same pipeline. * Use the (intentionally very short) default spans key `sc` in the `SpanCategorizer` * Adjust the default score weights to include the default key * Adjust the scorer to use `spans_{spans_key}` as the prefix for the returned score * Revert addition of `attr_name` argument to `score_spans` and adjust the key in the `getter` instead. Note that for `spancat` components with a custom `span_key`, the score weights currently need to be modified manually in `[training.score_weights]` for them to be available during training. To suppress the default score weights `spans_sc_p/r/f` during training, set them to `null` in `[training.score_weights]`. * Update website/docs/api/scorer.md * Fix scorer for spans key containing underscore * Increment version * Add Spans to Evaluate CLI (#8439) * Add Spans to Evaluate CLI * Change to spans_key * Add spans per_type output Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> * Fix spancat GPU issues (#8455) * Fix GPU issues * Require thinc >=8.0.6 * Switch to glorot_uniform_init * Fix and test ngram suggester * Include final ngram in doc for all sizes * Fix ngrams for docs of the same length as ngram size * Handle batches of docs that result in no ngrams * Add tests Co-authored-by: Ines Montani <ines@ines.io> Co-authored-by: Sofie Van Landeghem <svlandeg@users.noreply.github.com> Co-authored-by: svlandeg <sofie.vanlandeghem@gmail.com> Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com> Co-authored-by: Nirant <NirantK@users.noreply.github.com>
2021-06-24 13:35:27 +03:00
def test_extract_spans_span_indices():
model = extract_spans().initialize()
spans = Ragged(
model.ops.asarray([[0, 3], [2, 3], [5, 7]], dtype="i"),
model.ops.asarray([2, 1], dtype="i"),
)
x_lengths = model.ops.asarray([5, 10], dtype="i")
indices = _get_span_indices(model.ops, spans, x_lengths)
assert list(indices) == [0, 1, 2, 2, 10, 11]
def test_extract_spans_forward_backward():
model = extract_spans().initialize()
X = Ragged(model.ops.alloc2f(15, 4), model.ops.asarray([5, 10], dtype="i"))
spans = Ragged(
model.ops.asarray([[0, 3], [2, 3], [5, 7]], dtype="i"),
model.ops.asarray([2, 1], dtype="i"),
)
Y, backprop = model.begin_update((X, spans))
assert list(Y.lengths) == [3, 1, 2]
assert Y.dataXd.shape == (6, 4)
dX, spans2 = backprop(Y)
assert spans2 is spans
assert dX.dataXd.shape == X.dataXd.shape
assert list(dX.lengths) == list(X.lengths)
def test_spancat_model_init():
model = build_spancat_model(
build_Tok2Vec_model(**get_tok2vec_kwargs()), reduce_mean(), Logistic()
)
model.initialize()
def test_spancat_model_forward_backward(nO=5):
tok2vec = build_Tok2Vec_model(**get_tok2vec_kwargs())
docs = get_docs()
spans_list = []
lengths = []
for doc in docs:
spans_list.append(doc[:2])
spans_list.append(doc[1:4])
lengths.append(2)
spans = Ragged(
tok2vec.ops.asarray([[s.start, s.end] for s in spans_list], dtype="i"),
tok2vec.ops.asarray(lengths, dtype="i"),
)
model = build_spancat_model(
tok2vec, reduce_mean(), chain(Relu(nO=nO), Logistic())
).initialize(X=(docs, spans))
Y, backprop = model((docs, spans), is_train=True)
assert Y.shape == (spans.dataXd.shape[0], nO)
backprop(Y)