Spacy pretrain [Config validation error] - attempting to pretrain new vector weights

[Kobena-Afful]

Hi,
We are using prodigy to train a model for Named entity recognition. The model is performing alright but it is also picking up numbers that it shouldn't be. I think the data space we are using is very specific so I figured if we pretrained a tok2vec vector weights it will help our models perform better. We are using spaCy v3, so I set up a new config file for the pretrain. Unfortunately when I run it I am getting a:
Spacy pretrain [Config validation error]
tokenizer field required
components field required

I can't seem to figure out which parameters I am missing from the config file and I can't find examples of config files with this section to guide me to what the config parameter should be. Any assistance will be very much appreciated.
thanks!

[Kobena-Afful]

my config file is below if that is of any use:

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

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

[system]
gpu_allocator = null
seed = 0

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

[components]

[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.textcat]
factory = "textcat"
scorer = {"@scorers":"spacy.textcat_scorer.v1"}
threshold = 0.5

[components.textcat.model]
@architectures = "spacy.TextCatEnsemble.v2"
nO = null

[components.textcat.model.linear_model]
@architectures = "spacy.TextCatBOW.v2"
exclusive_classes = true
ngram_size = 1
no_output_layer = false
nO = null

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

[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 = ["NORM","PREFIX","SUFFIX","SHAPE"]
rows = [5000,2500,2500,2500]
include_static_vectors = true

[components.tok2vec.model.encode]
@architectures = "spacy.MaxoutWindowEncoder.v2"
width = 256
depth = 8
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.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 = 0
max_steps = 20000
eval_frequency = 200
frozen_components = []
annotating_components = []
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]
ents_f = 0.5
ents_p = 0.0
ents_r = 0.0
ents_per_type = null
cats_score = 0.5
cats_score_desc = null
cats_micro_p = null
cats_micro_r = null
cats_micro_f = null
cats_macro_p = null
cats_macro_r = null
cats_macro_f = null
cats_macro_auc = null
cats_f_per_type = null
cats_macro_auc_per_type = null

[model]
@architectures = "spacy.TextCatCNN.v2"
exclusive_classes = false
nO = null

[model.tok2vec]
@architectures = "spacy.HashEmbedCNN.v2"
pretrained_vectors = null
width = 128
depth = 8
embed_size = 2000
window_size = 1
maxout_pieces = 3
subword_features = true

[pretraining]
max_epochs = 1000
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.PretrainVectors.v1"
maxout_pieces = 3
hidden_size = 300
loss = "cosine

[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 = 1e-8
learn_rate = 0.001


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

[polm]

Thank you for including your config, it would have been impossible to troubleshoot the issue without it.

It looks like you removed two lines from the end of your config, adding them back should resolve the issue.

[initialize.components]

[initialize.tokenizer]

The error message could definitely be more clear about this.

A separate issue is that you seem to be missing a quote after cosine, though I guess that was just a copy/paste issue?

[Kobena-Afful]

Thank you very much Polm! That resolved it. And thanks for catching the missing quote after the cosine.