Spaces:
Running
Running
File size: 5,300 Bytes
03f6091 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 |
# -*- coding: utf-8 -*-
r"""
Polos command line interface (CLI)
==============
Composed by 4 main commands:
train Used to train a machine translation metric.
score Uses Polos to score a list of MT outputs.
download Used to download corpora or pretrained metric.
"""
import json
import os
import click
import yaml
from pytorch_lightning import seed_everything
from polos.corpora import corpus2download, download_corpus
from polos.models import download_model, load_checkpoint, model2download, str2model
from polos.trainer import TrainerConfig, build_trainer
@click.group()
def polos():
pass
@polos.command()
@click.option(
"--config",
"-f",
type=click.Path(exists=True),
required=True,
help="Path to the configure YAML file",
)
@click.option(
"--resume",
"-r",
type=click.Path(exists=True),
required=False,
help="Path to the configure YAML file",
)
def train(config,resume):
yaml_file = yaml.load(open(config).read(), Loader=yaml.FullLoader)
# Build Trainer
train_configs = TrainerConfig(yaml_file)
seed_everything(train_configs.seed)
trainer = build_trainer(train_configs.namespace(),resume)
# Print Trainer parameters into terminal
result = "Hyperparameters:\n"
for k, v in train_configs.namespace().__dict__.items():
result += "{0:30}| {1}\n".format(k, v)
click.secho(f"{result}", fg="green", nl=False)
# Build Model
try:
model_config = str2model[train_configs.model].ModelConfig(yaml_file)
print(str2model[train_configs.model].ModelConfig)
print(model_config.namespace())
model = str2model[train_configs.model](model_config.namespace())
except KeyError:
raise Exception(f"Invalid model {train_configs.model}!")
result = ""
for k, v in model_config.namespace().__dict__.items():
result += "{0:30}| {1}\n".format(k, v)
click.secho(f"{result}", fg="cyan")
# Train model
click.secho(f"{model.__class__.__name__} train starting:", fg="yellow")
trainer.fit(model)
# test model
# trainer.test(model)
def check_model(ctx, param, reference):
""" Helper function that checks if the model requires references or not. """
if reference is None and "wmt-large-qe-estimator-1719" not in ctx.params["model"]:
raise click.ClickException("Error: Missing option '--reference' / '-r'.")
return reference
@polos.command()
@click.option(
"--model",
default="wmt-large-da-estimator-1719",
help="Name of the pretrained model OR path to a model checkpoint.",
show_default=True,
type=str,
is_eager=True,
)
@click.option(
"--source",
"-s",
required=True,
help="Source segments.",
type=click.File(),
)
@click.option(
"--hypothesis",
"-h",
required=True,
help="MT outputs.",
type=click.File(),
)
@click.option(
"--reference",
"-r",
required=False,
help="Reference segments.",
type=click.File(),
callback=check_model,
)
@click.option(
"--cuda/--cpu",
default=True,
help="Flag that either runs inference on cuda or in cpu.",
show_default=True,
)
@click.option(
"--batch_size",
default=-1,
help="Batch size used during inference. By default uses the same batch size used during training.",
type=int,
)
@click.option(
"--to_json",
default=False,
help="Creates and exports model predictions to a JSON file.",
type=str,
show_default=True,
)
def score(model, source, hypothesis, reference, cuda, batch_size, to_json):
source = [s.strip() for s in source.readlines()]
hypothesis = [s.strip() for s in hypothesis.readlines()]
if reference:
reference = [s.strip() for s in reference.readlines()]
data = {"src": source, "mt": hypothesis, "ref": reference}
else:
data = {"src": source, "mt": hypothesis}
data = [dict(zip(data, t)) for t in zip(*data.values())]
model = load_checkpoint(model) if os.path.exists(model) else download_model(model)
data, scores = model.predict(data, cuda, show_progress=True, batch_size=batch_size)
if isinstance(to_json, str):
with open(to_json, "w") as outfile:
json.dump(data, outfile, ensure_ascii=False, indent=4)
click.secho(f"Predictions saved in: {to_json}.", fg="yellow")
for i in range(len(scores)):
click.secho("Segment {} score: {:.3f}".format(i, scores[i]), fg="yellow")
click.secho(
"Polos system score: {:.3f}".format(sum(scores) / len(scores)), fg="yellow"
)
@polos.command()
@click.option(
"--data",
"-d",
type=click.Choice(corpus2download.keys(), case_sensitive=False),
multiple=True,
help="Public corpora to download.",
)
@click.option(
"--model",
"-m",
type=click.Choice(model2download().keys(), case_sensitive=False),
multiple=True,
help="Pretrained models to download.",
)
@click.option(
"--saving_path",
type=str,
help="Relative path to save the downloaded files.",
required=True,
)
def download(data, model, saving_path):
print("Download ...")
print(data,model)
for d in data:
download_corpus(d, saving_path)
for m in model:
download_model(m, saving_path)
if __name__ == '__main__':
polos() |