Reproducibility advice

[sataketatsuya]

I trained the agent model from scrach, but i could not reproduce the result in your paper. I would like to get some advice about reproducibility.
I put some information on my environment bellow. Thank you.

in ./FirstTextWorldProblems/utils.py Saver Class

def _load_from_checkpoint(self):
    load_from = self.pretrained_model_path
    # print("Trying to load model from {}.".format(load_from))
    try:
        if self.device == 'cpu':
            state_dict = torch.load(load_from, map_location='cpu')
        else:
            state_dict = torch.load(load_from, map_location=self.device)
        # self.model.load_state_dict(state_dict, strict=True)        # comment out because I don't want to load any trained weights about the model
        self.model.to(self.device)
        print("Loaded model from '{}'".format(load_from))
    except:
        print("Failed to load checkpoint {} ...".format(load_from))
        IPython.embed()

Linux OS version

Ubuntu 20.04.2 LTS \n \l

the package lists pip installed

Package                  Version
------------------------ ---------
allennlp                 2.9.3
asttokens                2.0.5
async-generator          1.10
attrs                    21.4.0
backcall                 0.2.0
base58                   2.1.1
blessings                1.7
blis                     0.7.7
boto3                    1.24.22
botocore                 1.27.22
cached-path              1.1.4
cachetools               5.2.0
catalogue                2.0.7
certifi                  2022.6.15
click                    8.0.4
cloudpickle              2.0.0
commonmark               0.9.1
cycler                   0.11.0
cymem                    2.0.6
debugpy                  1.6.0
decorator                5.1.1
dill                     0.3.5.1
distlib                  0.3.4
docker-pycreds           0.4.0
entrypoints              0.4
executing                0.8.3
fairscale                0.4.6
filelock                 3.6.0
Flask                    2.1.1
fonttools                4.31.2
gdown                    4.4.0
gevent                   1.3.5
gitdb                    4.0.9
GitPython                3.1.27
gluonnlp                 0.7.1
google-api-core          2.8.2
google-auth              2.9.0
google-cloud-core        2.3.1
google-cloud-storage     2.4.0
google-crc32c            1.3.0
google-resumable-media   2.3.3
googleapis-common-protos 1.56.3
gpustat                  0.6.0
graphviz                 0.8.4
greenlet                 0.4.13
gym                      0.23.1
gym-notices              0.0.6
h11                      0.13.0
h5py                     3.7.0
hashids                  1.3.1
huggingface-hub          0.8.1
importlib-metadata       4.12.0
iniconfig                1.1.1
ipykernel                6.15.0
ipython                  8.4.0
itsdangerous             2.1.2
jedi                     0.18.1
jericho                  3.1.0
Jinja2                   3.1.2
jmespath                 1.0.1
joblib                   1.1.0
jsonnet                  0.18.0
jupyter-client           7.3.4
jupyter-core             4.10.0
kiwisolver               1.4.0
langcodes                3.3.0
lmdb                     1.3.0
MarkupSafe               2.1.1
matplotlib               3.5.1
matplotlib-inline        0.1.3
mementos                 1.3.1
more-itertools           8.12.0
murmurhash               1.0.6
mxnet                    1.9.1
nest-asyncio             1.5.5
networkx                 2.7.1
nltk                     3.7
numpy                    1.19.0
nvidia-ml-py3            7.352.0
outcome                  1.1.0
packaging                21.3
pandas                   1.4.1
parso                    0.8.3
pathtools                0.1.2
pathy                    0.6.1
pexpect                  4.8.0
pickleshare              0.7.5
pip                      21.2.4
platformdirs             2.5.2
plotly                   5.6.0
pluggy                   1.0.0
preshed                  3.0.6
prettytable              3.3.0
promise                  2.3
prompt-toolkit           2.0.10
protobuf                 3.20.1
psutil                   5.9.1
ptyprocess               0.7.0
pure-eval                0.2.2
py                       1.11.0
pyasn1                   0.4.8
pyasn1-modules           0.2.8
pybars3                  0.9.7
pydantic                 1.8.2
pydot                    1.4.2
pygame                   2.1.2
pyglet                   1.5.26
Pygments                 2.12.0
PyMeta3                  0.5.1
pyparsing                3.0.9
PySocks                  1.7.1
pytest                   7.1.2
python-dateutil          2.8.2
pytz                     2022.1
PyYAML                   6.0
pyzmq                    23.2.0
recordclass              0.17.2
regex                    2022.6.2
rich                     12.4.4
rsa                      4.8
s3transfer               0.6.0
sacremoses               0.0.53
scikit-learn             1.0.2
scipy                    1.8.0
seaborn                  0.11.2
selenium                 4.1.3
sentencepiece            0.1.96
sentry-sdk               1.6.0
setproctitle             1.2.3
setuptools               61.2.0
sh                       1.14.2
shortuuid                1.0.9
sklearn                  0.0
smart-open               5.2.1
smmap                    5.0.0
sniffio                  1.2.0
sortedcontainers         2.4.0
spacy                    3.2.3
spacy-legacy             3.0.9
spacy-loggers            1.0.1
srsly                    2.4.2
stack-data               0.3.0
TatSu                    5.6.1
tenacity                 8.0.1
tensorboardX             2.5.1
termcolor                1.1.0
textworld                1.5.2
thinc                    8.0.15
threadpoolctl            3.1.0
tokenizers               0.12.1
tomli                    2.0.1
torch                    1.11.0
torchsummary             1.5.1
torchvision              0.12.0
tornado                  6.2
tqdm                     4.63.1
traitlets                5.3.0
transformers             4.18.0
trio                     0.20.0
trio-websocket           0.9.2
typer                    0.4.2
typing_extensions        4.1.1
urllib3                  1.26.9
urwid                    2.1.2
virtualenv               20.15.1
wandb                    0.12.20
wasabi                   0.9.0
wcwidth                  0.2.5
Werkzeug                 2.1.1
wheel                    0.37.1
wsproto                  1.1.0
zipp                     3.8.0

Downloaded Dataset from here. It contains 4,440 different training games, 222 validation games, 514 test games.
https://competitions.codalab.org/competitions/21557#learn_the_details-data

The result which was runned 3 epoch on training dataset(4,400 games).

The training result on yout paper

I am looking forward to your good response. Thank you.