--- title: MovieLens Dataset keywords: fastai sidebar: home_sidebar summary: "Implementation of MovieLens datasets." description: "Implementation of MovieLens datasets." nb_path: "nbs/datasets/movielens.ipynb" ---
{% raw %}
{% endraw %} {% raw %}
{% endraw %} {% raw %}

class ML1MDataset[source]

ML1MDataset(args)

{% endraw %} {% raw %}
{% endraw %} {% raw %}
class Args:
    min_rating = 0
    min_uc = 5
    min_sc = 5
    split = 'leave_one_out'
    data_raw_path = 'data/bronze'
    data_save_path = 'data/silver'

args = Args()

os.makedirs(args.data_raw_path, exist_ok=True)
os.makedirs(args.data_save_path, exist_ok=True)

dataset = ML1MDataset(args)
dataset.preprocess()
Raw file doesn't exist. Downloading...

Turning into implicit ratings
Filtering triplets
Densifying index
Splitting
100%|██████████| 6040/6040 [00:02<00:00, 2707.83it/s]
{% endraw %} {% raw %}
!tree --du -h -C ./data
./data
├── [ 24M]  bronze
│   └── [ 24M]  ml-1m
│       ├── [167K]  movies.dat
│       ├── [ 23M]  ratings.dat
│       ├── [5.4K]  README
│       └── [131K]  users.dat
└── [3.0M]  silver
    └── [3.0M]  ml-1m_min_rating0-min_uc5-min_sc5-splitleave_one_out
        └── [3.0M]  dataset.pkl

  27M used in 4 directories, 5 files
{% endraw %}