def __init__(self, use_tflite: bool = True, top=OBJECTIVE_TOP_LIVE) -> None:
self.top = top
if use_tflite:
from overtrack_cv.core.tflite import TFLiteModel
self.model = TFLiteModel(os.path.join(os.path.dirname(__file__), "data", "parse_objective.tflite"))
else:
from overtrack_cv.core.tf import load_model
self.model = load_model(os.path.join(os.path.dirname(__file__), "data", "v18"))
def __init__(self, use_tflite: bool = True):
if use_tflite:
from overtrack_cv.core.tflite import TFLiteModel
self.model = TFLiteModel(
os.path.join(os.path.dirname(__file__), "data",
"big_noodle.tflite"))
else:
from overtrack_cv.core.tf import load_model
self.model = load_model(
os.path.join(os.path.dirname(__file__), "data", "big_noodle"))
def __init__(self):
import tensorflow as tf
from tensorflow.python.keras import Model
from overtrack_cv.core.tf import (
all_custom_objects,
decode_ctc,
deserialize,
load_model,
)
self.extract: Model = load_model(
os.path.join(os.path.dirname(__file__), "data", "extract_v14"))
model_path = os.path.join(os.path.dirname(__file__), "data",
"parse_v6")
with open(model_path + "/assets/saved_model.json") as f:
self.parse_config = json.load(f)
self.parse_config["config"]["layers"][0]["config"][
"batch_input_shape"][2] = None
del self.parse_config["config"]["layers"][0]["config"]["ragged"]
self.parse: Model = deserialize(self.parse_config,
custom_objects=all_custom_objects)
self.parse.load_weights(model_path + "/variables/variables")
self.parse.trainable = False
with open(model_path + "/outputs.json") as f:
self.outputs = json.load(f)["outputs"]
self.outputs_dict = {o["name"]: o for o in self.outputs}
self.parse_layer_defs = {
layer["name"]: layer
for layer in self.parse_config["config"]["layers"]
}
# self.parse.outputs += [
# tf.keras.layers.Activation('softmax', name='heroes_softmax')(self.parse.outputs[1]),
# # tf.nn.softmax(self.parse.outputs[1]), # hero_softmax
# ]
# self.parse.build((None, 50, 600, 3))
self._decode_ctc = decode_ctc
self._softmax = tf.nn.softmax
def __init__(self, use_tflite: bool = True):
self.map_rotated = deque(maxlen=10)
self.map_rotate_in_config = None
if use_tflite:
from overtrack_cv.core.tflite import TFLiteModel
self.model = TFLiteModel(
os.path.join(os.path.dirname(__file__), "data",
"minimap_filter.tflite"))
else:
from overtrack_cv.core.tf import load_model
self.model = load_model(
os.path.join(os.path.dirname(__file__), "data",
"minimap_filter")
# "C:/Users/simon/overtrack_2/training/apex_minimap/v3/v15/checkpoint"
)
# from tensorflow.python.keras.saving import export_saved_model
# export_saved_model(self.model, os.path.join(os.path.dirname(__file__), 'data', 'minimap_filter'), serving_only=True)
self.current_game: Optional[CurrentGame] = None
self.current_composite: Optional[RingsComposite] = None
def __init__(self):
self.model: Model = load_model(
os.path.join(os.path.dirname(__file__), "data", "big_noodle_ctc"))