File size: 3,203 Bytes
e0d0d76 |
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 |
import math
from hydra.core.config_store import ConfigStore
from dataclasses import dataclass
@dataclass
class PreProcess:
augm_size: int = 1
src_train: str = '/data/USPTO/src_train.txt'
tgt_train: str = '/data/USPTO/tgt_train.txt'
src_valid: str = '/data/USPTO/src_valid.txt'
tgt_valid: str = '/data/USPTO/tgt_valid.txt'
batch_size: int = 256
@dataclass
class ModelConfig:
dim_model: int = 512
num_encoder_layers: int = 6
num_decoder_layers: int = 6
nhead: int = 8
dropout: float = 0.1
dim_ff: int = 2048
ckpt:str = '/ckpts/Transformer/ckpt_conditional.pth'
@dataclass
class TrainConfig:
src_train: str = '/data/USPTO/src_train.txt'
tgt_train: str = '/data/USPTO/tgt_train.txt'
src_valid: str = '/data/USPTO/src_valid.txt'
tgt_valid: str = '/data/USPTO/tgt_valid.txt'
batch_size: int = 128
label_smoothing: float = 0.0
lr: float = 0.001
betas: tuple = (0.9, 0.998)
step_num: int = 500000 # set training steps
patience: int = 10
log_interval : int = 100
val_interval: int = 1000
save_interval: int = 10000
@dataclass
class TranslateConfig:
src_train: str = '/data/USPTO/src_train.txt'
tgt_train: str = '/data/USPTO/tgt_train.txt'
src_valid: str = '/data/USPTO/src_valid.txt'
tgt_valid: str = '/data/USPTO/tgt_valid.txt'
GCN_ckpt: str = '/ckpts/GCN/GCN.pth'
out_dir: str = '/translation'
src_test_path: str = '/data/input/test.txt'
annotated_templates: str = '/data/beamsearch_template_list.txt'
filename: str = 'test'
GCN_num_sampling: int = 10
inf_max_len: int = 256
nbest: int = 10
beam_size: int = 10
@dataclass
class GCN_TrainConfig:
train: str = '/data/USPTO/src_train.txt'
valid: str = '/data/USPTO/src_valid.txt'
test: str = '/data/USPTO/src_test.txt'
batch_size: int = 256
dim: int = 256
n_conv_hidden: int = 1
n_mlp_hidden: int = 3
dropout: float = 0.1
lr: float = 0.0004
epochs: int = 100
patience: int = 5
save_path: str = '/ckpts/GCN'
@dataclass
class MCTSConfig:
src_train: str = '/data/USPTO/src_train.txt'
tgt_train: str = '/data/USPTO/tgt_train.txt'
src_valid: str = '/data/USPTO/src_valid.txt'
tgt_valid: str = '/data/USPTO/tgt_valid.txt'
n_step: int = 200
max_depth: int = 10
in_smiles_file: str = '/data/input/init_smiles_drd2.txt'
out_dir: str = '/mcts_out'
ucb_c: float = 1/math.sqrt(2)
reward_name: str = 'DRD2' # 'DRD2' or 'QED'
ckpt_Transformer: str = '/ckpts/Transformer/ckpt_conditional.pth'
ckpt_GCN: str = '/ckpts/GCN/GCN.pth'
beam_width:int = 10
nbest:int = 10
exp_num_sampling:int = 10
rollout_depth:int = 2
roll_num_sampling:int = 5
@dataclass
class Config:
prep: PreProcess = PreProcess()
model: ModelConfig = ModelConfig()
train: TrainConfig = TrainConfig()
translate: TranslateConfig = TranslateConfig()
GCN_train: GCN_TrainConfig = GCN_TrainConfig()
mcts: MCTSConfig = MCTSConfig()
cs = ConfigStore.instance()
cs.store(name="config", node=Config)
|