Rooba
/
spaCy
mirror of https://github.com/explosion/spaCy.git
1
1
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Update NER config

This commit is contained in:
Matthw Honnibal 2020-07-09 19:33:54 +02:00
parent 77af0a6bb4
commit 0becc5954b
1 changed files with 6 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
examples/experiments/onto-ner.cfg
View File

@ -5,7 +5,7 @@
# data is passed in sentence-by-sentence via some prior preprocessing. # data is passed in sentence-by-sentence via some prior preprocessing.
gold_preproc = false gold_preproc = false
# Limitations on training document length or number of examples. # Limitations on training document length or number of examples.
max_length = 5000 max_length = 3000
limit = 0 limit = 0
# Data augmentation # Data augmentation
orth_variant_level = 0.0 orth_variant_level = 0.0
@ -17,20 +17,20 @@ max_steps = 0
eval_frequency = 1000 eval_frequency = 1000
# Other settings # Other settings
seed = 0 seed = 0
accumulate_gradient = 2 accumulate_gradient = 1
use_pytorch_for_gpu_memory = false use_pytorch_for_gpu_memory = false
# Control how scores are printed and checkpoints are evaluated. # Control how scores are printed and checkpoints are evaluated.
scores = ["speed", "ents_p", "ents_r", "ents_f"] scores = ["speed", "ents_p", "ents_r", "ents_f"]
score_weights = {"ents_f": 1.0} score_weights = {"ents_f": 1.0}
# These settings are invalid for the transformer models. # These settings are invalid for the transformer models.
init_tok2vec = null init_tok2vec = null
discard_oversize = true discard_oversize = false
omit_extra_lookups = false omit_extra_lookups = false
batch_by_words = true batch_by = "words"
[training.batch_size] [training.batch_size]
@schedules = "compounding.v1" @schedules = "compounding.v1"
start = 1000 start = 100
stop = 1000 stop = 1000
compound = 1.001 compound = 1.001
@ -45,12 +45,6 @@ use_averages = true
eps = 1e-8 eps = 1e-8
learn_rate = 0.001 learn_rate = 0.001
#[training.optimizer.learn_rate]
#@schedules = "warmup_linear.v1"
#warmup_steps = 1000
#total_steps = 50000
#initial_rate = 0.003
[nlp] [nlp]
lang = "en" lang = "en"
vectors = null vectors = null
@ -74,6 +68,6 @@ width = 96
depth = 4 depth = 4
window_size = 1 window_size = 1
embed_size = 2000 embed_size = 2000
maxout_pieces = 1 maxout_pieces = 3
subword_features = true subword_features = true
dropout = ${training:dropout} dropout = ${training:dropout}