Datasets:

valeriol29
/

MovieLens-3B

item_id int64	rating float64	list_index int64	source_row int64	dataset_source string	synthetic bool
3,233	5	1	1	movielens-3b	true
3,798	4	2	1	movielens-3b	true
4,609	4	3	1	movielens-3b	true
5,349	4.5	4	1	movielens-3b	true
12,270	2.5	5	1	movielens-3b	true
14,694	5	6	1	movielens-3b	true
14,838	3.5	7	1	movielens-3b	true
16,968	3.5	8	1	movielens-3b	true
17,822	3.5	9	1	movielens-3b	true
19,774	5	10	1	movielens-3b	true
19,907	2.5	11	1	movielens-3b	true
20,223	3	12	1	movielens-3b	true
23,653	1.5	13	1	movielens-3b	true
23,752	5	14	1	movielens-3b	true
24,347	4	15	1	movielens-3b	true
27,963	0.5	16	1	movielens-3b	true
28,321	0.5	17	1	movielens-3b	true
30,170	0.5	18	1	movielens-3b	true
32,376	0.5	19	1	movielens-3b	true
32,626	0.5	20	1	movielens-3b	true
35,504	0.5	21	1	movielens-3b	true
37,231	0.5	22	1	movielens-3b	true
37,934	0.5	23	1	movielens-3b	true
40,273	0.5	24	1	movielens-3b	true
42,681	0.5	25	1	movielens-3b	true
43,530	0.5	26	1	movielens-3b	true
45,099	0.5	27	1	movielens-3b	true
45,785	0.5	28	1	movielens-3b	true
45,842	0.5	29	1	movielens-3b	true
48,219	0.5	30	1	movielens-3b	true
49,983	0.5	31	1	movielens-3b	true
50,032	0.5	32	1	movielens-3b	true
53,231	0.5	33	1	movielens-3b	true
53,463	0.5	34	1	movielens-3b	true
53,549	4	35	1	movielens-3b	true
53,578	3	36	1	movielens-3b	true
54,021	4	37	1	movielens-3b	true
55,467	2.5	38	1	movielens-3b	true
55,613	4	39	1	movielens-3b	true
56,035	3.5	40	1	movielens-3b	true
56,125	3	41	1	movielens-3b	true
56,239	3.5	42	1	movielens-3b	true
56,420	4	43	1	movielens-3b	true
56,889	3.5	44	1	movielens-3b	true
57,076	3.5	45	1	movielens-3b	true
57,574	4.5	46	1	movielens-3b	true
57,959	3	47	1	movielens-3b	true
60,011	3.5	48	1	movielens-3b	true
60,351	3.5	49	1	movielens-3b	true
62,074	4	50	1	movielens-3b	true
62,470	3	51	1	movielens-3b	true
63,149	3.5	52	1	movielens-3b	true
64,138	4	53	1	movielens-3b	true
64,268	3.5	54	1	movielens-3b	true
65,017	3.5	55	1	movielens-3b	true
65,278	3.5	56	1	movielens-3b	true
65,503	3	57	1	movielens-3b	true
65,727	4	58	1	movielens-3b	true
67,008	3	59	1	movielens-3b	true
68,285	4	60	1	movielens-3b	true
68,604	3	61	1	movielens-3b	true
68,637	4	62	1	movielens-3b	true
69,267	4	63	1	movielens-3b	true
69,520	4	64	1	movielens-3b	true
70,219	4	65	1	movielens-3b	true
71,309	3.5	66	1	movielens-3b	true
71,319	2.5	67	1	movielens-3b	true
71,400	3.5	68	1	movielens-3b	true
72,214	3	69	1	movielens-3b	true
72,605	3	70	1	movielens-3b	true
74,003	2	71	1	movielens-3b	true
74,442	4	72	1	movielens-3b	true
74,470	3.5	73	1	movielens-3b	true
74,650	4	74	1	movielens-3b	true
74,717	4	75	1	movielens-3b	true
74,955	2.5	76	1	movielens-3b	true
75,030	2	77	1	movielens-3b	true
75,328	4	78	1	movielens-3b	true
75,922	4	79	1	movielens-3b	true
75,946	3	80	1	movielens-3b	true
75,998	4.5	81	1	movielens-3b	true
76,008	4.5	82	1	movielens-3b	true
76,894	3.5	83	1	movielens-3b	true
76,920	3	84	1	movielens-3b	true
77,230	3.5	85	1	movielens-3b	true
77,811	5	86	1	movielens-3b	true
77,872	3.5	87	1	movielens-3b	true
78,174	3.5	88	1	movielens-3b	true
78,784	2.5	89	1	movielens-3b	true
78,834	4.5	90	1	movielens-3b	true
79,045	3	91	1	movielens-3b	true
79,253	3.5	92	1	movielens-3b	true
79,321	3.5	93	1	movielens-3b	true
79,826	4	94	1	movielens-3b	true
80,130	4.5	95	1	movielens-3b	true
80,594	4.5	96	1	movielens-3b	true
80,991	4.5	97	1	movielens-3b	true
82,466	4	98	1	movielens-3b	true
83,387	4.5	99	1	movielens-3b	true
84,097	4	100	1	movielens-3b	true

End of preview. Expand in Data Studio

MovieLens-3B Synthetic

Parquet republication of the local 16x32_*.csv synthetic MovieLens expansion. This is not an official GroupLens release. The 16 source shards are one split, train. There is no author-defined train/test cut.

Original sources

Expansion pattern: MLPerf recommendation data generation, ml-20m x 16 x 32, https://github.com/mlcommons/training
GroupLens description of that expansion: https://grouplens.org/datasets/movielens/movielens-1b/
Base ratings: MovieLens 20M, https://grouplens.org/datasets/movielens/20m/
MovieLens context paper: https://doi.org/10.1145/2827872

Each source row is one user, a comma-separated movie-id list, and an aligned rating list. The lists are unnested. list_index is the 1-based position in that row. user_id is already global across shards. source_shard is the file index 0–15. source_row is the 1-based line number within that file. There is no timestamp. synthetic is true.


shards	16
interactions	3,366,896,162

License

The underlying MovieLens ratings are covered by the GroupLens research license: research use with acknowledgement, no implied endorsement, redistribution only under the same conditions, and no commercial or revenue-bearing use without permission from GroupLens. See https://grouplens.org/datasets/movielens/20m/.

Citation

@article{harper2015movielens,
  author = {Harper, F. Maxwell and Konstan, Joseph A.},
  title = {The MovieLens Datasets: History and Context},
  journal = {ACM Transactions on Interactive Intelligent Systems},
  volume = {5},
  number = {4},
  pages = {19:1--19:19},
  year = {2015},
  doi = {10.1145/2827872},
  url = {https://doi.org/10.1145/2827872}
}

Downloads last month: 38