deepmd.tf.nvnmd.utils

deepmd.tf.nvnmd.utils#

Submodules#

Attributes#

nvnmd_cfg

Classes#

`Encode`	Encoding value as hex, bin, and dec format.
`FioBin`	Input and output for binary file.
`FioDic`	Input and output for dict class data
`FioTxt`	Input and output for .txt file with string.

Functions#

`nvnmd_args`(→ dargs.Argument)
`one_layer`(→ deepmd.tf.env.tf.Tensor)	Build one layer with continuous or quantized value.
`map_nvnmd`(→ float)	Mapping function implemented by numpy.
`get_filter_weight`(→ tuple)	Get weight and bias of embedding network.
`get_fitnet_weight`(→ tuple)	Get weight and bias of fitting network.

Package Contents#

deepmd.tf.nvnmd.utils.nvnmd_args(fold_subdoc: bool = False) → dargs.Argument[source]#

deepmd.tf.nvnmd.utils.nvnmd_cfg[source]#

class deepmd.tf.nvnmd.utils.Encode[source]#

Encoding value as hex, bin, and dec format.

qr(v: numpy.ndarray | float, nbit: int = 14) → numpy.ndarray | float[source]#: Quantize value using round.

qf(v: numpy.ndarray | float, nbit: int = 14) → numpy.ndarray | float[source]#: Quantize value using floor.

qc(v: numpy.ndarray | float, nbit: int = 14) → numpy.ndarray | float[source]#: Quantize value using ceil.

split_expo_mant(v: numpy.ndarray | float, min: int = -1000) → tuple[numpy.ndarray | float, numpy.ndarray | float][source]#

find_max_expo(v: numpy.ndarray | float, expo_min: int = -1000) → float[source]#

norm_expo(v: numpy.ndarray | float, nbit_frac: int = 20, expo_min: int = -1000) → tuple[numpy.ndarray | float, float][source]#

flt2bin_one(v: float, nbit_expo: int, nbit_frac: int) → str[source]#

flt2bin(data: numpy.ndarray, nbit_expo: int, nbit_frac: int) → list[str][source]#: Convert float into binary string list.

byte2hex(bs: bytes, nbyte: int) → list[str][source]#: Convert byte into hex bs: low byte in the first hex: low byte in the right.

check_dec(idec: numpy.ndarray, nbit: int, signed: bool = False, name: str = '') → None[source]#: Check whether the data (idec) is in the range range is \([0, 2^nbit-1]\) for unsigned range is \([-2^{nbit-1}, 2^{nbit-1}-1]\) for signed.

extend_list(slbin: list[str], nfull: int) → list[str][source]#

Extend the list (slbin) to the length (nfull) the attached element of list is 0.

such as, when

slbin = [‘10010’,’10100’],

nfull = 4

extent it to

[‘10010’,’10100’,’00000’,’00000]

extend_bin(slbin: list[str], nfull: int) → list[str][source]#

Extend the element of list (slbin) to the length (nfull).

such as, when

slbin = [‘10010’,’10100’],

nfull = 6

extent to

[‘010010’,’010100’]

extend_hex(slhex: list[str], nfull: int) → list[str][source]#: Extend the element of list (slhex) to the length (nfull).

split_bin(sbin: str | list[str], nbit: int) → list[str][source]#: Split sbin into many segment with the length nbit.

reverse_bin(slbin: list[str], nreverse: int) → list[str][source]#: Reverse binary string list per nreverse value.

merge_bin(slbin: list[str], nmerge: int) → list[str][source]#: Merge binary string list per nmerge value.

dec2bin(idec: numpy.ndarray | float, nbit: int = 10, signed: bool = False, name: str = '') → list[str][source]#: Convert dec array to binary string list.

hex2bin_str(shex: str) → str[source]#: Convert hex string to binary string.

hex2bin(data: numpy.ndarray | list[str]) → list[str][source]#: Convert hex string list to binary string list.

bin2hex_str(sbin: str) → str[source]#: Convert binary string to hex string.

bin2hex(data: numpy.ndarray | list[str]) → list[str][source]#: Convert binary string list to hex string list.

class deepmd.tf.nvnmd.utils.FioBin[source]#

Input and output for binary file.

load(file_name: str = '', default_value: str = '') → str | bytes[source]#: Load binary file into bytes value.

save(file_name: str, data: list[str]) → None[source]#: Save hex string into binary file.

class deepmd.tf.nvnmd.utils.FioDic[source]#

Input and output for dict class data the file can be .json or .npy file containing a dictionary.

load(file_name: str = '', default_value: dict = {}) → dict[source]#

save(file_name: str = '', dic: dict = {}) → None[source]#

get(jdata: dict, key: str, default_value: Any) → Any[source]#

update(jdata: dict, jdata_o: dict) → dict[source]#

Update key-value pair is key in jdata_o.keys().

Parameters:

jdata: new jdata
jdata_o: origin jdata

class deepmd.tf.nvnmd.utils.FioTxt[source]#

Input and output for .txt file with string.

load(file_name: str = '', default_value: list = []) → list[str][source]#: Load .txt file into string list.

save(file_name: str = '', data: list = []) → None[source]#: Save string list into .txt file.

deepmd.tf.nvnmd.utils.one_layer(inputs: deepmd.tf.env.tf.Tensor, outputs_size: int, activation_fn: Callable | None = tf.nn.tanh, precision: deepmd.tf.env.tf.DType = GLOBAL_TF_FLOAT_PRECISION, stddev: float = 1.0, bavg: float = 0.0, name: str = 'linear', reuse: bool | None = None, seed: int | None = None, use_timestep: bool = False, trainable: bool = True, useBN: bool = False, uniform_seed: bool = False, initial_variables: dict | None = None, mixed_prec: dict | None = None, final_layer: bool = False) → deepmd.tf.env.tf.Tensor[source]#: Build one layer with continuous or quantized value. Its weight and bias can be initialed with random or constant value.

deepmd.tf.nvnmd.utils.map_nvnmd(x: float, map_y: numpy.ndarray, map_dy: numpy.ndarray, prec: float, nbit: int | None = None) → float[source]#: Mapping function implemented by numpy.

deepmd.tf.nvnmd.utils.get_filter_weight(weights: int, spe_j: int, layer_l: int) → tuple[source]#

Get weight and bias of embedding network.

Parameters:

weightsdict: weights
spe_jint: special order of neighbor atom j 0~ntype-1
layer_l: layer order in embedding network 1~nlayer

deepmd.tf.nvnmd.utils.get_fitnet_weight(weights: dict, spe_i: int, layer_l: int, nlayer: int = 10) → tuple[source]#

Get weight and bias of fitting network.

Parameters:

weightsdict: weights
spe_iint: special order of central atom i 0~ntype-1
layer_lint: layer order in embedding network 0~nlayer-1
nlayerint: number of layers