Noun chunks returning empty list for the custom pipeline

[gunalanlakshmanan]

I have trained a custom pipeline and with new ner and Spancat component and used the parser and tagger from source.

Here is my config

[paths]
train = null
dev = null
vectors = null
init_tok2vec = null
raw_text = null
ner_labels = null
spancat_labels = null

[system]
gpu_allocator = null
seed = 0

[nlp]
lang = "en"
pipeline = ["tok2vec","tagger","parser","ner","spancat"]
batch_size = 1000
disabled = []
before_creation = null
after_creation = null
after_pipeline_creation = null
tokenizer = {"@tokenizers":"spacy.Tokenizer.v1"}

[components]

[components.tagger]
source = "en_core_web_sm"
replace_listeners = ["model.tok2vec"]

[components.parser]
source = "en_core_web_sm"
replace_listeners = ["model.tok2vec"]

[components.ner]
factory = "ner"
incorrect_spans_key = null
moves = null
scorer = {"@scorers":"spacy.ner_scorer.v1"}
update_with_oracle_cut_size = 100

[components.ner.model]
@architectures = "spacy.TransitionBasedParser.v2"
state_type = "ner"
extra_state_tokens = false
hidden_width = 64
maxout_pieces = 2
use_upper = true
nO = null

[components.ner.model.tok2vec]
@architectures = "spacy.Tok2VecListener.v1"
width = ${components.tok2vec.model.encode.width}
upstream = "*"

[components.spancat]
factory = "spancat"
max_positive = null
scorer = {"@scorers":"spacy.spancat_scorer.v1"}
spans_key = "sc"
threshold = 0.5

[components.spancat.model]
@architectures = "spacy.SpanCategorizer.v1"

[components.spancat.model.reducer]
@layers = "spacy.mean_max_reducer.v1"
hidden_size = 128

[components.spancat.model.scorer]
@layers = "spacy.LinearLogistic.v1"
nO = null
nI = null

[components.spancat.model.tok2vec]
@architectures = "spacy.Tok2VecListener.v1"
width = ${components.tok2vec.model.encode.width}
upstream = "*"

[components.spancat.suggester]
@misc = "spacy.ngram_suggester.v1"
sizes = [1, 2]

[components.tok2vec]
factory = "tok2vec"

[components.tok2vec.model]
@architectures = "spacy.Tok2Vec.v2"

[components.tok2vec.model.embed]
@architectures = "spacy.MultiHashEmbed.v2"
width = ${components.tok2vec.model.encode.width}
attrs = ["ORTH","SHAPE"]
rows = [5000,2500]
include_static_vectors = false

[components.tok2vec.model.encode]
@architectures = "spacy.MaxoutWindowEncoder.v2"
width = 96
depth = 4
window_size = 1
maxout_pieces = 3

[corpora]

[corpora.dev]
@readers = "spacy.Corpus.v1"
path = ${paths.dev}
max_length = 0
gold_preproc = false
limit = 0
augmenter = null

[corpora.pretrain]
@readers = "spacy.JsonlCorpus.v1"
path = ${paths.raw_text}
min_length = 5
max_length = 500
limit = 0

[corpora.train]
@readers = "spacy.Corpus.v1"
path = ${paths.train}
max_length = 0
gold_preproc = false
limit = 0
augmenter = null

[training]
dev_corpus = "corpora.dev"
train_corpus = "corpora.train"
seed = ${system.seed}
gpu_allocator = ${system.gpu_allocator}
dropout = 0.1
accumulate_gradient = 1
patience = 1600
max_epochs = 100
max_steps = 20000
eval_frequency = 200
frozen_components = ["tok2vec", "tagger", "parser"]
annotating_components = ["tok2vec"]
before_to_disk = null

[training.batcher]
@batchers = "spacy.batch_by_words.v1"
discard_oversize = false
tolerance = 0.2
get_length = null

[training.batcher.size]
@schedules = "compounding.v1"
start = 100
stop = 1000
compound = 1.001
t = 0.0

[training.logger]
@loggers = "spacy.ConsoleLogger.v1"
progress_bar = false

[training.optimizer]
@optimizers = "Adam.v1"
beta1 = 0.9
beta2 = 0.999
L2_is_weight_decay = true
L2 = 0.01
grad_clip = 1.0
use_averages = false
eps = 0.00000001
learn_rate = 0.001

[training.score_weights]
tag_acc = 0.25
dep_uas = 0.12
dep_las = 0.12
dep_las_per_type = null
sents_p = null
sents_r = null
sents_f = 0.0
ents_f = 0.25
ents_p = 0.0
ents_r = 0.0
ents_per_type = null
spans_sc_f = 0.25
spans_sc_p = 0.0
spans_sc_r = 0.0

[pretraining]
max_epochs = 1500
dropout = 0.2
n_save_every = null
n_save_epoch = null
component = "tok2vec"
layer = ""
corpus = "corpora.pretrain"

[pretraining.batcher]
@batchers = "spacy.batch_by_words.v1"
size = 3000
discard_oversize = false
tolerance = 0.2
get_length = null

[pretraining.objective]
@architectures = "spacy.PretrainCharacters.v1"
maxout_pieces = 3
hidden_size = 300
n_characters = 4

[pretraining.optimizer]
@optimizers = "Adam.v1"
beta1 = 0.9
beta2 = 0.999
L2_is_weight_decay = true
L2 = 0.01
grad_clip = 1.0
use_averages = true
eps = 0.00000001
learn_rate = 0.001

[initialize]
vectors = ${paths.vectors}
init_tok2vec = ${paths.init_tok2vec}
vocab_data = null
lookups = null
after_init = null

[initialize.before_init]
@callbacks = "sms_tokenizer"

[initialize.components]

[initialize.components.ner]

[initialize.components.ner.labels]
@readers = "spacy.read_labels.v1"
path = ${paths.ner_labels}

[initialize.components.spancat]

[initialize.components.spancat.labels]
@readers = "spacy.read_labels.v1"
path = ${paths.spancat_labels}
require = false

[initialize.tokenizer]

I have checked the syntax iterators in Default, it contains noun chucks. The default NLP model "en_core_web_sm" provides noun chunks for same sentence.

default_doc = nlp("It's a beautiful dog")
list(default_doc.noun_chunks)
Out[17]: [It, a beautiful dog]

custom_doc = custom_nlp("It's a beautiful dog")
list(custom_doc.noun_chunks)
Out[15]: []

The parser output is same for both default and custom model.

for token in default_doc:
    print(token.text, token.tag_)
    
It PRP
's VBZ
a DT
beautiful JJ
dog NN

for token in custom_doc:
    print(token.text, token.tag_)
    
It PRP
's VBZ
a DT
beautiful JJ
dog NN

I am not sure what I am missing here. I am not getting the noun chunks and pos_ from the doc created using the custom model. Can you please help me on finding the problem here?

Thanks in advance

[polm]

The noun chunks functionality depends on the UPOS tags (doc.pos), not the fine-grained tags (doc.tag). You need to include the AttributeRuler to get those. You can do that by sourcing the AttributeRuler from en_core_web_sm like you did the Tagger, though note it doesn't need replace_listeners.

For example, this config, which just sources components from the pretrained pipeline, works:

[paths]
train = null
dev = null
vectors = null
init_tok2vec = null
raw_text = null

[system]
gpu_allocator = null
seed = 0

[nlp]
lang = "en"
pipeline = ["tagger","parser", "attribute_ruler"]
batch_size = 1000
disabled = []
before_creation = null
after_creation = null
after_pipeline_creation = null
tokenizer = {"@tokenizers":"spacy.Tokenizer.v1"}

[components]

[components.tagger]
source = "en_core_web_sm"
replace_listeners = ["model.tok2vec"]

[components.parser]
source = "en_core_web_sm"
replace_listeners = ["model.tok2vec"]

[components.attribute_ruler]
source = "en_core_web_sm"

[corpora]

[corpora.dev]
@readers = "spacy.Corpus.v1"
path = ${paths.dev}
max_length = 0
gold_preproc = false
limit = 0
augmenter = null

[corpora.pretrain]
@readers = "spacy.JsonlCorpus.v1"
path = ${paths.raw_text}
min_length = 5
max_length = 500
limit = 0

[corpora.train]
@readers = "spacy.Corpus.v1"
path = ${paths.train}
max_length = 0
gold_preproc = false
limit = 0
augmenter = null

[training]
dev_corpus = "corpora.dev"
train_corpus = "corpora.train"
seed = ${system.seed}
gpu_allocator = ${system.gpu_allocator}
dropout = 0.1
accumulate_gradient = 1
patience = 1600
max_epochs = 100
max_steps = 20000
eval_frequency = 200
frozen_components = ["tok2vec", "tagger", "parser"]
annotating_components = ["tok2vec"]
before_to_disk = null

[training.batcher]
@batchers = "spacy.batch_by_words.v1"
discard_oversize = false
tolerance = 0.2
get_length = null

[training.batcher.size]
@schedules = "compounding.v1"
start = 100
stop = 1000
compound = 1.001
t = 0.0

[training.logger]
@loggers = "spacy.ConsoleLogger.v1"
progress_bar = false

[training.optimizer]
@optimizers = "Adam.v1"
beta1 = 0.9
beta2 = 0.999
L2_is_weight_decay = true
L2 = 0.01
grad_clip = 1.0
use_averages = false
eps = 0.00000001
learn_rate = 0.001

[training.score_weights]
tag_acc = 0.25
dep_uas = 0.12
dep_las = 0.12
dep_las_per_type = null
sents_p = null
sents_r = null
sents_f = 0.0
ents_f = 0.25
ents_p = 0.0
ents_r = 0.0
ents_per_type = null
spans_sc_f = 0.25
spans_sc_p = 0.0
spans_sc_r = 0.0

[pretraining]
max_epochs = 1500
dropout = 0.2
n_save_every = null
n_save_epoch = null
component = "tok2vec"
layer = ""
corpus = "corpora.pretrain"

[pretraining.batcher]
@batchers = "spacy.batch_by_words.v1"
size = 3000
discard_oversize = false
tolerance = 0.2
get_length = null

[pretraining.objective]
@architectures = "spacy.PretrainCharacters.v1"
maxout_pieces = 3
hidden_size = 300
n_characters = 4

[pretraining.optimizer]
@optimizers = "Adam.v1"
beta1 = 0.9
beta2 = 0.999
L2_is_weight_decay = true
L2 = 0.01
grad_clip = 1.0
use_averages = true
eps = 0.00000001
learn_rate = 0.001

[initialize]
vectors = ${paths.vectors}
init_tok2vec = ${paths.init_tok2vec}
vocab_data = null
lookups = null
after_init = null

[initialize.components]

[initialize.tokenizer]

[gunalanlakshmanan]

Thanks @polm. After adding the AttributeRuler, I could get the NounChunks.