def predict_and_eval(self, model_path):
predictor = Predictor.from_checkpoint(default_predictor_params(),
checkpoint=model_path)
for sample in predictor.predict(file_dataset()):
self.assertGreater(
sample.outputs.avg_char_probability,
0.95) # The model was trained and should yield good results
def test_raw_dataset_prediction(self):
args = PredictionAttrs()
predictor = Predictor.from_checkpoint(PredictorParams(
progress_bar=False, silent=True),
checkpoint=args.checkpoint[0])
params = PipelineParams(type=DataSetType.FILE, files=args.files)
for sample in predictor.predict(params):
pass
def test_raw_prediction(self):
args = PredictionAttrs()
predictor = Predictor.from_checkpoint(PredictorParams(
progress_bar=False, silent=True),
checkpoint=args.checkpoint[0])
images = [load_image(file) for file in args.files]
for result in predictor.predict_raw(images):
self.assertGreater(result.outputs.avg_char_probability, 0)
def test_white_image_raw_prediction(self):
args = PredictionAttrs()
predictor = Predictor.from_checkpoint(PredictorParams(
progress_bar=False, silent=True),
checkpoint=args.checkpoint[0])
images = [np.zeros(shape=(200, 50))]
for result in predictor.predict_raw(images):
print(result.outputs.sentence)
def test_raw_prediction(self):
args = PredictionAttrs()
predictor = Predictor.from_checkpoint(PredictorParams(
progress_bar=False, silent=True),
checkpoint=args.checkpoint[0])
images = [load_image(file) for file in args.files]
for file, image in zip(args.files, images):
_, prediction, _ = list(predictor.predict_raw([image]))[0]
print(file, prediction.sentence)
def benchmark_prediction(model, batch_size, processes, n_examples, runs=10):
params = PredictorParams(silent=True)
predictor = Predictor.from_checkpoint(params, model)
data = (np.random.random((400, 48)) * 255).astype(np.uint8)
print("Running with bs={}, proc={}, n={}".format(batch_size, processes,
n_examples))
start = time.time()
for i in range(runs):
list(predictor.predict_raw([data] * n_examples, batch_size=batch_size))
end = time.time()
return (end - start) / runs
def create_single_model_predictor():
checkpoint = os.path.join(this_dir, "models", "best.ckpt")
predictor = Predictor.from_checkpoint(default_predictor_params(),
checkpoint=checkpoint)
return predictor
