himkt

$ git                                                                                                                                                                                                                                             ± [ ● ][master]
add                -- add file contents to index
am                 -- apply patches from a mailbox
apply              -- apply patch to files and/or to index
archimport         -- import an Arch repository into git
archive            -- create archive of files from named tree
bisect             -- find, by binary search, change that introduced a bug
blame              -- show what revision and author last modified each line of a file
branch             -- list, create, or delete branches
bundle             -- move objects and refs by archive

himkt

# include <cmath>
# include <iostream>
# include <queue>
# include <set>
# include <stack>
# include <string>
# include <vector>


using namespace std;

himkt

"""
One of the design principles of AllenNLP is the use of a modular,
declarative language (JSON) for defining experiments and models.

This is implemented by giving each AllenNLP class a method

.. code-block
    @classmethod
    def from_params(cls, params: Params, **extras) -> 'ClassName':
        ...

himkt

  
{
  "dataset_reader":{
    "type": "sst_tokens",
    "use_subtrees": true,
    "granularity": "5-class"
  },
  "validation_dataset_reader":{
    "type": "sst_tokens",
    "use_subtrees": false,

himkt

study = optuna.create_study(direction="maximize")
study.optimize(
    objective,
    n_jobs=1,  # number of processes in parallel execution
    n_trials=30,  # number of trials to train a model
)

optuna.integration.allennlp.dump_best_config(
    "./config/imdb_optuna.jsonnet",
    "best.imdb1.json",

himkt

import sklearn
import sklearn.datasets
import sklearn.ensemble
import sklearn.model_selection
import sklearn.svm

import optuna

# Define an objective function to be minimized.
def objective(trial):

himkt

def objective(trial: optuna.Trial) -> float:
    trial.suggest_int("embedding_dim", 32, 256)
    trial.suggest_int("max_filter_size", 2, 6)
    trial.suggest_int("num_filters", 32, 256)
    trial.suggest_int("output_dim", 32, 256)
    trial.suggest_float("dropout", 0.0, 0.8)
    trial.suggest_float("lr", 5e-3, 5e-1, log=True)

    executor = optuna.integration.allennlp.AllenNLPExecutor(
        trial=trial,  # trial object

himkt

study = optuna.create_study(
    storage="sqlite:///result/trial.db",  # save results in DB
    sampler=optuna.samplers.TPESampler(seed=24),
    study_name="optuna_allennlp",
    direction="maximize",
)

timeout = 60 * 60 * 10  # timeout (sec): 60*60*10 sec => 10 hours
study.optimize(
    objective,

himkt

{
    "data_loader": {
        "batch_size": 64,
        "shuffle": true
    },
    "dataset_reader": {
        "lazy": false,
        "token_indexers": {
            "tokens": {
                "lowercase_tokens": true,

himkt

Creating virtualenv test in /Users/makoto-hiramatsu/Desktop/test/.venv
Using virtualenv: /Users/makoto-hiramatsu/Desktop/test/.venv
PyPI: 77 packages found for allennlp *
Using version ^0.9.0 for allennlp

Updating dependencies
Resolving dependencies...
   1: fact: test is 0.1.0
   1: derived: test
   1: fact: test depends on allennlp (^0.9.0)