[WARNING] [W112] The model specified to use for initial vectors (./ner_obj) has no vectors. This is almost certainly a mistake.

[shrinidhin]

Hi!I am working on building an entity based sentiment analyzer with custom trained NER and Entity Linker components. I have a custom trained NER component added to my pipeline and i am training my entity linker component. Now, When I train my Entity Linker, It gives the following warning
[WARNING] [W112] The model specified to use for initial vectors (./ner_obj) has no vectors. This is almost certainly a mistake.

ner_obj is the directory which holds the trained ner component model files.

This warning comes even even I add any other component to the pipeline with my custom trained component. Any way to resolve this?

[thomashacker]

Hello, is your ner_obj model using Transformer embeddings? If yes, the spaCy Transformers pipeline doesn't include word vectors. You can read a bit more about this here: #7643

[thomashacker]

Hello,
you can't use the ner_obj model to initialize the vectors in the EntityLinker config because it has no static vectors. You need to source the tok2vec component from the ner_obj model.

[components.tok2vec][components.tok2vec]
source = "${paths.ner}"
component = "tok2vec"
source = "${paths.ner}"
component = "tok2vec"

[shrinidhin]

@thomashacker Hi!Thank you for your reply!I kind of made a mistake, Printed the incorrect config file. Following is the config file with the suggested changes:

[paths]
train = "./data/train.spacy"
dev = "./data/dev.spacy"
raw = null
init_tok2vec = null
kb = "./stocks-sentiment_kb"
vectors = "${paths.ner_model}"
ner_model = "./ner_obj"

[system]
seed = 342
gpu_allocator = null

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

[components]

[components.entity_linker]
factory = "entity_linker"
entity_vector_length = 64
get_candidates = {"@misc":"spacy.CandidateGenerator.v1"}
incl_context = true
incl_prior = true
labels_discard = []
n_sents = 0
overwrite = true
scorer = {"@scorers":"spacy.entity_linker_scorer.v1"}

[components.entity_linker.model]
@architectures = "spacy.EntityLinker.v1"
nO = null

[components.entity_linker.model.tok2vec]
@architectures = "spacy.HashEmbedCNN.v1"
pretrained_vectors = null
width = 96
depth = 2
embed_size = 2000
window_size = 1
maxout_pieces = 3
subword_features = true

[components.ner]
source = ${paths.ner_model}
component = "ner"

[components.sentencizer]
factory = "sentencizer"
overwrite = false
punct_chars = null
scorer = {"@scorers":"spacy.senter_scorer.v1"}

[components.tok2vec]
source = ${paths.ner_model}
component = "tok2vec"

[corpora]

[corpora.dev]
@readers = "MyCorpus.v1"
file = ${paths.dev}

[corpora.train]
@readers = "MyCorpus.v1"
file = ${paths.train}

[training]
train_corpus = "corpora.train"
dev_corpus = "corpora.dev"
seed = ${system.seed}
gpu_allocator = ${system.gpu_allocator}
dropout = 0.2
patience = 10000
eval_frequency = 200
accumulate_gradient = 2
max_epochs = 0
max_steps = 600
frozen_components = ["sentencizer","tok2vec","ner"]
before_to_disk = null
annotating_components = []

[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]
sents_f = 0.33
sents_p = 0.0
sents_r = 0.0
ents_f = 0.33
ents_p = 0.0
ents_r = 0.0
ents_per_type = null
nel_micro_f = 0.33
nel_micro_r = 0.0
nel_micro_p = 0.0

[pretraining]

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

[initialize.components]

[initialize.components.entity_linker]

[initialize.components.entity_linker.kb_loader]
@misc = "spacy.KBFromFile.v1"
kb_path = ${paths.kb}

[initialize.tokenizer]

[optimizer]
@optimizers = "Adam.v1"
learn_rate = 0.001
beta1 = 0.9
beta2 = 0.999
L2 = 0.0
eps = 0.00000001
grad_clip = 1.0
L2_is_weight_decay = true
use_averages = true

I am still getting the same warning.

[thomashacker]

You're getting the warning because you can't use the ner_obj model to initialize the vectors in the EntityLinker config since it has no static vectors.

[shrinidhin]

Then adding the following statement to the config makes no difference right

[components.tok2vec]
source = "${paths.ner}"
component = "tok2vec"

Since the ner_obj model has no static vectors?

[thomashacker]

For the warning to disappear you need to change this line in the config:

[paths]
vectors = "${paths.ner_model}"

to

[paths]
vectors = null