How to use _c method in Molotov

Best Python code snippet using molotov_python

default.py

Source:default.py

1#!/usr/bin/env python32# Copyright (c) Facebook, Inc. and its affiliates.3# This source code is licensed under the MIT license found in the4# LICENSE file in the root directory of this source tree.5from typing import List, Optional, Union6import numpy as np7from habitat import get_config as get_task_config8from habitat.config import Config as CN9DEFAULT_CONFIG_DIR = "configs/"10CONFIG_FILE_SEPARATOR = ","11# -----------------------------------------------------------------------------12# EXPERIMENT CONFIG13# -----------------------------------------------------------------------------14_C = CN()15_C.BASE_TASK_CONFIG_PATH = "configs/tasks/pointnav.yaml"16_C.TASK_CONFIG = CN()  # task_config will be stored as a config node17_C.CMD_TRAILING_OPTS = []  # store command line options as list of strings18_C.TRAINER_NAME = "ppo"19_C.ENV_NAME = "NavRLEnv"20_C.SIMULATOR_GPU_ID = 021_C.TORCH_GPU_ID = 022_C.VIDEO_OPTION = ["disk", "tensorboard"]23_C.TENSORBOARD_DIR = "tb"24_C.VIDEO_DIR = "video_dir"25_C.TEST_EPISODE_COUNT = 226_C.EVAL_CKPT_PATH_DIR = "data/checkpoints"  # path to ckpt or path to ckpts dir27_C.NUM_PROCESSES = 1628_C.SENSORS = ["RGB_SENSOR", "DEPTH_SENSOR"]29_C.CHECKPOINT_FOLDER = "data/checkpoints"30_C.NUM_UPDATES = 1000031_C.LOG_INTERVAL = 1032_C.LOG_FILE = "train.log"33_C.CHECKPOINT_INTERVAL = 5034_C.VIS_INTERVAL = 20035# -----------------------------------------------------------------------------36# EVAL CONFIG37# -----------------------------------------------------------------------------38_C.EVAL = CN()39# The split to evaluate on40_C.EVAL.SPLIT = "val"41_C.EVAL.USE_CKPT_CONFIG = True42# -----------------------------------------------------------------------------43# REINFORCEMENT LEARNING (RL) ENVIRONMENT CONFIG44# -----------------------------------------------------------------------------45_C.RL = CN()46_C.RL.REWARD_MEASURE = "distance_to_goal"47_C.RL.SUCCESS_MEASURE = "spl"48_C.RL.SUCCESS_REWARD = 10.049_C.RL.SLACK_REWARD = -0.0150# -----------------------------------------------------------------------------51# PROXIMAL POLICY OPTIMIZATION (PPO)52# -----------------------------------------------------------------------------53_C.RL.PPO = CN()54_C.RL.PPO.clip_param = 0.255_C.RL.PPO.ppo_epoch = 456_C.RL.PPO.num_mini_batch = 1657_C.RL.PPO.value_loss_coef = 0.558_C.RL.PPO.entropy_coef = 0.0159_C.RL.PPO.lr = 7e-460_C.RL.PPO.eps = 1e-561_C.RL.PPO.max_grad_norm = 0.562_C.RL.PPO.num_steps = 563_C.RL.PPO.use_gae = True64_C.RL.PPO.use_linear_lr_decay = False65_C.RL.PPO.use_linear_clip_decay = False66_C.RL.PPO.gamma = 0.9967_C.RL.PPO.tau = 0.9568_C.RL.PPO.reward_window_size = 5069_C.RL.PPO.use_normalized_advantage = True70_C.RL.PPO.hidden_size = 51271# -----------------------------------------------------------------------------72# DECENTRALIZED DISTRIBUTED PROXIMAL POLICY OPTIMIZATION (DD-PPO)73# -----------------------------------------------------------------------------74_C.RL.DDPPO = CN()75_C.RL.DDPPO.sync_frac = 0.676_C.RL.DDPPO.distrib_backend = "GLOO"77_C.RL.DDPPO.rnn_type = "LSTM"78_C.RL.DDPPO.num_recurrent_layers = 279_C.RL.DDPPO.backbone = "resnet50"80_C.RL.DDPPO.pretrained_weights = "data/ddppo-models/gibson-2plus-resnet50.pth"81# Loads pretrained weights82_C.RL.DDPPO.pretrained = False83# Loads just the visual encoder backbone weights84_C.RL.DDPPO.pretrained_encoder = False85# Whether or not the visual encoder backbone will be trained86_C.RL.DDPPO.train_encoder = True87# Whether or not to reset the critic linear layer88_C.RL.DDPPO.reset_critic = True89# -----------------------------------------------------------------------------90# ORBSLAM2 BASELINE91# -----------------------------------------------------------------------------92_C.ORBSLAM2 = CN()93_C.ORBSLAM2.SLAM_VOCAB_PATH = "habitat_baselines/slambased/data/ORBvoc.txt"94_C.ORBSLAM2.SLAM_SETTINGS_PATH = (95    "habitat_baselines/slambased/data/mp3d3_small1k.yaml"96)97_C.ORBSLAM2.MAP_CELL_SIZE = 0.198_C.ORBSLAM2.MAP_SIZE = 4099_C.ORBSLAM2.CAMERA_HEIGHT = get_task_config().SIMULATOR.DEPTH_SENSOR.POSITION[100    1101]102_C.ORBSLAM2.BETA = 100103_C.ORBSLAM2.H_OBSTACLE_MIN = 0.3 * _C.ORBSLAM2.CAMERA_HEIGHT104_C.ORBSLAM2.H_OBSTACLE_MAX = 1.0 * _C.ORBSLAM2.CAMERA_HEIGHT105_C.ORBSLAM2.D_OBSTACLE_MIN = 0.1106_C.ORBSLAM2.D_OBSTACLE_MAX = 4.0107_C.ORBSLAM2.PREPROCESS_MAP = True108_C.ORBSLAM2.MIN_PTS_IN_OBSTACLE = (109    get_task_config().SIMULATOR.DEPTH_SENSOR.WIDTH / 2.0110)111_C.ORBSLAM2.ANGLE_TH = float(np.deg2rad(15))112_C.ORBSLAM2.DIST_REACHED_TH = 0.15113_C.ORBSLAM2.NEXT_WAYPOINT_TH = 0.5114_C.ORBSLAM2.NUM_ACTIONS = 3115_C.ORBSLAM2.DIST_TO_STOP = 0.05116_C.ORBSLAM2.PLANNER_MAX_STEPS = 500117_C.ORBSLAM2.DEPTH_DENORM = get_task_config().SIMULATOR.DEPTH_SENSOR.MAX_DEPTH118_C.attention = CN()119_C.attention.n_head = 4120_C.attention.d_model = 512 + 32 + 1121_C.attention.d_k = 512 + 32 + 1122_C.attention.d_v = 512 + 32 + 1123_C.attention.dropout = 0.1124_C.attention.lsh = CN()125_C.attention.lsh.bucket_size = 10126_C.attention.lsh.n_hashes = 4127_C.attention.lsh.add_local_attn_hash = False128_C.attention.lsh.causal = True129_C.attention.lsh.attn_chunks = 8130_C.attention.lsh.random_rotations_per_head = False131_C.attention.lsh.attend_across_buckets = True132_C.attention.lsh.allow_duplicate_attention = True133_C.attention.lsh.num_mem_kv = 0134_C.attention.lsh.one_value_head = False135_C.attention.lsh.full_attn_thres = 'none'136_C.attention.lsh.return_attn = False137_C.attention.lsh.post_attn_dropout = 0.1138_C.attention.lsh.dropout = 0.1139_C.attention.lsh.use_full_attn = False140_C.memory = CN()141_C.memory.embedding_size = 512 + 32142_C.memory.memory_size = 100143_C.memory.pose_dim = 5144def get_config(145    config_paths: Optional[Union[List[str], str]] = None,146    opts: Optional[list] = None,147) -> CN:148    r"""Create a unified config with default values overwritten by values from149    `config_paths` and overwritten by options from `opts`.150    Args:151        config_paths: List of config paths or string that contains comma152        separated list of config paths.153        opts: Config options (keys, values) in a list (e.g., passed from154        command line into the config. For example, `opts = ['FOO.BAR',155        0.5]`. Argument can be used for parameter sweeping or quick tests.156    """157    config = _C.clone()158    if config_paths:159        if isinstance(config_paths, str):160            if CONFIG_FILE_SEPARATOR in config_paths:161                config_paths = config_paths.split(CONFIG_FILE_SEPARATOR)162            else:163                config_paths = [config_paths]164        for config_path in config_paths:165            config.merge_from_file(config_path)166    config.TASK_CONFIG = get_task_config(config.BASE_TASK_CONFIG_PATH)167    if opts:168        config.CMD_TRAILING_OPTS = opts169        config.merge_from_list(opts)170    config.freeze()...

defaults.py

Source:defaults.py

1from yacs.config import CfgNode as CN2_C = CN()3_C.MODEL = CN()4_C.MODEL.META_ARCHITECTURE = 'SSDDetector'5_C.MODEL.DEVICE = "cuda"6# match default boxes to any ground truth with jaccard overlap higher than a threshold (0.5)7_C.MODEL.THRESHOLD = 0.58_C.MODEL.NUM_CLASSES = 219# Hard negative mining10_C.MODEL.NEG_POS_RATIO = 3 # negative : positive = 3:111_C.MODEL.CENTER_VARIANCE = 0.112_C.MODEL.SIZE_VARIANCE = 0.213# ---------------------------------------------------------------------------- #14# Backbone15# ---------------------------------------------------------------------------- #16_C.MODEL.BACKBONE = CN()17_C.MODEL.BACKBONE.NAME = 'vgg'18_C.MODEL.BACKBONE.OUT_CHANNELS = (512, 1024, 512, 256, 256, 256)19_C.MODEL.BACKBONE.PRETRAINED = True20_C.MODEL.BACKBONE.ISFREEZE = False21_C.MODEL.BACKBONE.RFBTYPE = 'NONE' # NONE or Advanced or Basic or Original22# for hrnet23_C.MODEL.BACKBONE.C = 3224_C.MODEL.BACKBONE.BN_MOMENTUM = 0.125# ---------------------------------------------------------------------------- #26# Neck27# ---------------------------------------------------------------------------- #28_C.MODEL.NECK = CN()29_C.MODEL.NECK.NAME = 'NONE'30# -----------------------------------------------------------------------------31# PRIORS32# -----------------------------------------------------------------------------33_C.MODEL.PRIORS = CN()34_C.MODEL.PRIORS.FEATURE_MAPS = [38, 19, 10, 5, 3, 1] # éè¦ä¾æ®ç½ç»çç¹å¾å¾å¤§å°æ¥è¿è¡è°è35_C.MODEL.PRIORS.STRIDES = [8, 16, 32, 64, 100, 300]36# MIN_SIZES å MAX_SIZES æ¯ç± ratio çåæ°è®¡ç®åºæ¥ç, ä½æ¯ MIN_SIZES å MAX æ¯ä¾æ®æ°æ®éçå°ºåº¦åå¸æ¥è®¡ç®ç, å³å®ä¸åç¹å¾å¾ä¸prior boxçè¾¹é¿37_C.MODEL.PRIORS.MIN_SIZES = [30, 60, 111, 162, 213, 264] # these parameters are for voc0712 dataset 38_C.MODEL.PRIORS.MAX_SIZES = [60, 111, 162, 213, 264, 315]39_C.MODEL.PRIORS.ASPECT_RATIOS = [[2], [2, 3], [2, 3], [2, 3], [2], [2]]40# When has 1 aspect ratio, every location has 4 boxes, 2 ratio 6 boxes.41# #boxes = 2 + #ratio * 242_C.MODEL.PRIORS.BOXES_PER_LOCATION = [4, 6, 6, 6, 4, 4]  # number of boxes per feature map location43_C.MODEL.PRIORS.CLIP = True44# -----------------------------------------------------------------------------45# Box Head46# -----------------------------------------------------------------------------47_C.MODEL.BOX_HEAD = CN()48_C.MODEL.BOX_HEAD.NAME = 'SSDBoxHead'49_C.MODEL.BOX_HEAD.PREDICTOR = 'SSDBoxPredictor'50# -----------------------------------------------------------------------------51# INPUT52# -----------------------------------------------------------------------------53_C.INPUT = CN()54# Image size55_C.INPUT.IMAGE_SIZE = 30056# Values to be used for image normalization, RGB layout57_C.INPUT.PIXEL_MEAN = [123, 117, 104]58# -----------------------------------------------------------------------------59# Dataset60# -----------------------------------------------------------------------------61_C.DATASETS = CN()62# List of the dataset names for training, as present in paths_catalog.py63_C.DATASETS.TRAIN = ()64# List of the dataset names for testing, as present in paths_catalog.py65_C.DATASETS.TEST = ()66# -----------------------------------------------------------------------------67# DataLoader68# -----------------------------------------------------------------------------69_C.DATA_LOADER = CN()70# Number of data loading threads71_C.DATA_LOADER.NUM_WORKERS = 872_C.DATA_LOADER.PIN_MEMORY = True73# ---------------------------------------------------------------------------- #74# Solver75# ---------------------------------------------------------------------------- #76_C.SOLVER = CN()77# train configs78_C.SOLVER.MAX_ITER = 12000079_C.SOLVER.LR_STEPS = [80000, 100000]80_C.SOLVER.GAMMA = 0.181_C.SOLVER.BATCH_SIZE = 3282_C.SOLVER.LR = 1e-3 # original 1e-383_C.SOLVER.MOMENTUM = 0.984_C.SOLVER.WEIGHT_DECAY = 5e-485_C.SOLVER.WARMUP_FACTOR = 1.0 / 3 # original 1.0 / 386_C.SOLVER.WARMUP_ITERS = 500 # original 50087# ---------------------------------------------------------------------------- #88# Specific test options89# ---------------------------------------------------------------------------- #90_C.TEST = CN()91_C.TEST.NMS_THRESHOLD = 0.4592_C.TEST.CONFIDENCE_THRESHOLD = 0.0193_C.TEST.MAX_PER_CLASS = -194_C.TEST.MAX_PER_IMAGE = 10095_C.TEST.BATCH_SIZE = 10...

Automation Testing Tutorials

Learn to execute automation testing from scratch with LambdaTest Learning Hub. Right from setting up the prerequisites to run your first automation test, to following best practices and diving deeper into advanced test scenarios. LambdaTest Learning Hubs compile a list of step-by-step guides to help you be proficient with different test automation frameworks i.e. Selenium, Cypress, TestNG etc.