Not getting the folder structure inside /model-best/transformer/model while training Custom NER model

[Chetan8000]

Not getting the folder structure inside the model while training the Custom NER model, right now just getting model file instead of a folder /model-best/transformer/model

when I tried this previously that time I'll get the folder structure in the above location ( /model-best/transformer/model)
right now I'm using the same config.cfg file and configuration but right now I'll just get the model file.

here is the folder structure I'll get previously,

/model-best/transformer/model
                             | ner
                             | transformer
                             |_______model
                             |       | config.json
                             |       | pytorch_model.bin
                             |       | special_tokens_map.json
                             |       | tokenizer.json
                             |       | tokenizer_config.json
                             |       | vocab.json
                             |_______config
                             | vocab
                             | meta.json
                             | tokenizer

and right now I'm getting this

/model-best/transformer/model
                             | ner
                             | transformer
                             |_______model
                             |_______config
                             | vocab
                             | meta.json
                             | tokenizer

here is my config.cfg

[paths]
train = "/content/drive/MyDrive/Flue_Model/corpus/train_df_bio_150_doc.spacy"
dev = "/content/drive/MyDrive/Flue_Model/corpus/valid_df_bio_100_doc.spacy"
vectors = null
init_tok2vec = null

[system]
gpu_allocator = "pytorch"
seed = 0

[nlp]
lang = "en"
pipeline = ["transformer","ner"]
batch_size = 128
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
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 = false
nO = null

[components.ner.model.tok2vec]
@architectures = "spacy-transformers.TransformerListener.v1"
grad_factor = 1.0
pooling = {"@layers":"reduce_mean.v1"}
upstream = "*"

[components.transformer]
factory = "transformer"
max_batch_items = 1024
set_extra_annotations = {"@annotation_setters":"spacy-transformers.null_annotation_setter.v1"}


[components.transformer.model]
@architectures = "spacy-transformers.TransformerModel.v1"
name = "studio-ousia/luke-base"

[components.transformer.model.get_spans]
@span_getters = "spacy-transformers.strided_spans.v1"
window = 128
stride = 96

[components.transformer.model.tokenizer_config]
use_fast = true

[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]
accumulate_gradient = 10
dev_corpus = "corpora.dev"
train_corpus = "corpora.train"
seed = ${system.seed}
gpu_allocator = ${system.gpu_allocator}
dropout = 0.15
patience = 1600
max_epochs = 0
max_steps = 130000
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

[training.optimizer.learn_rate]
@schedules = "warmup_linear.v1"
warmup_steps = 250
total_steps = 20000
initial_rate = 0.00005

[training.score_weights]
ents_f = 1.0
ents_p = 0.0
ents_r = 0.0
ents_per_type = null

[pretraining]

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

[initialize.components]

[initialize.tokenizer]

Your Environment

• Operating System: Google Colab
• Python Version Used: 3.7.12
• spaCy Version Used: 3.1.2
• Environment Information: Colab

[adrianeboyd]

The serialization format changed in spacy-transformers v1.1. The model file contains all the same data as in the directory in the v1.0. Some additional details here: https://github.com/explosion/spacy-transformers/releases/tag/v1.1.0

[Chetan8000]

Hey, thanks for the reply,

I tried this for version spacy-transformers==1.0.6

but I'll get the below error, (using same config.cfg)

✔ Created output directory: output_updated
ℹ Saving to output directory: output_updated
ℹ Using GPU: 0

=========================== Initializing pipeline ===========================
[2022-02-16 08:43:02,028] [INFO] Set up nlp object from config
[2022-02-16 08:43:02,070] [INFO] Pipeline: ['transformer', 'ner']
[2022-02-16 08:43:02,079] [INFO] Created vocabulary
[2022-02-16 08:43:02,100] [INFO] Finished initializing nlp object
Downloading: 100% 761/761 [00:00<00:00, 729kB/s]
Traceback (most recent call last):
  File "/usr/lib/python3.7/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/usr/lib/python3.7/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/usr/local/lib/python3.7/dist-packages/spacy/__main__.py", line 4, in <module>
    setup_cli()
  File "/usr/local/lib/python3.7/dist-packages/spacy/cli/_util.py", line 69, in setup_cli
    command(prog_name=COMMAND)
  File "/usr/local/lib/python3.7/dist-packages/click/core.py", line 829, in __call__
    return self.main(*args, **kwargs)
  File "/usr/local/lib/python3.7/dist-packages/click/core.py", line 782, in main
    rv = self.invoke(ctx)
  File "/usr/local/lib/python3.7/dist-packages/click/core.py", line 1259, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/usr/local/lib/python3.7/dist-packages/click/core.py", line 1066, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/usr/local/lib/python3.7/dist-packages/click/core.py", line 610, in invoke
    return callback(*args, **kwargs)
  File "/usr/local/lib/python3.7/dist-packages/typer/main.py", line 497, in wrapper
    return callback(**use_params)  # type: ignore
  File "/usr/local/lib/python3.7/dist-packages/spacy/cli/train.py", line 60, in train_cli
    nlp = init_nlp(config, use_gpu=use_gpu)
  File "/usr/local/lib/python3.7/dist-packages/spacy/training/initialize.py", line 84, in init_nlp
    nlp.initialize(lambda: train_corpus(nlp), sgd=optimizer)
  File "/usr/local/lib/python3.7/dist-packages/spacy/language.py", line 1273, in initialize
    proc.initialize(get_examples, nlp=self, **p_settings)
  File "/usr/local/lib/python3.7/dist-packages/spacy_transformers/pipeline_component.py", line 343, in initialize
    self.model.initialize(X=docs)
  File "/usr/local/lib/python3.7/dist-packages/thinc/model.py", line 299, in initialize
    self.init(self, X=X, Y=Y)
  File "/usr/local/lib/python3.7/dist-packages/spacy_transformers/layers/transformer_model.py", line 81, in init
    tokenizer, transformer = huggingface_from_pretrained(name, tok_cfg)
  File "/usr/local/lib/python3.7/dist-packages/spacy_transformers/util.py", line 31, in huggingface_from_pretrained
    tokenizer = AutoTokenizer.from_pretrained(str_path, **config)
  File "/usr/local/lib/python3.7/dist-packages/transformers/models/auto/tokenization_auto.py", line 352, in from_pretrained
    config = AutoConfig.from_pretrained(pretrained_model_name_or_path, **kwargs)
  File "/usr/local/lib/python3.7/dist-packages/transformers/models/auto/configuration_auto.py", line 363, in from_pretrained
    config_class = CONFIG_MAPPING[config_dict["model_type"]]
KeyError: 'luke'

[adrianeboyd]

The serialization was broken for a lot of use cases in v1.0.x. We'd strongly recommend using v1.1+ instead.

That might be unrelated to this particular error, which looks more like your version of transformers is too old for this model type.

[Chetan8000]

so the reason behind, I'm asking for serialization (folder structure ) is that if I want further finetune my model which is saved in my local

[components.transformer.model]
@architectures = "spacy-transformers.TransformerModel.v1"
name = "/content/drive/MyDrive/flu_model/best-model/transformer/model"

then it asks me for the config.json

getting below error,

- '/content/drive/MyDrive/flu_model/best-model/transformer/model' is a correct model identifier listed on 'https://huggingface.co/models'

- or '/content/drive/MyDrive/flu_model/best-model/transformer/model' is the correct path to a directory containing a config.json file

if model file contains all the same data as in the directory why this ask for the config.json

[adrianeboyd]

The model file contains all the necessary information, but in a spacy-specific format, so you can't load it with transformers anymore.

You want to finish fine-tuning the model with transformers before loading it into your spacy pipeline. In the spacy pipeline, it's fine-tuned for the specific downstream tasks in the pipeline like the specific ner component, so it doesn't make sense to fine-tune this modified version further with transformers directly. Once it's in the spacy pipeline, you would want to train with spacy instead.