Torch-GRU4Rec#

Out-of-the-box:
A single typo needed to be fixed, so the code is able to run.
The small differences (e.g. missing momentum parameter) add up to a noticeable difference when Torch-GRU4Rec is compared to the matching features version.
Sampling is performed after all item scores are computed, which slows down training. This bug is rooted so deep in the code that we did not fix it.

Rees46#

Metric

BPR-Max

Metrics#
Implementation	Variant	Recall@1	MRR@1	Recall@5	MRR@5	Recall@10	MRR@10	Recall@20	MRR@20
GRU4Rec Official	Best params	0.1027	0.1027	0.2897	0.1680	0.4027	0.1831	0.5206	0.1913
GRU4Rec Official	Torch-GRU4Rec params	0.0968	0.0968	0.2801	0.1607	0.3923	0.1757	0.5112	0.1839
Torch-GRU4Rec	OOB	0.0954	0.0954	0.2774	0.1588	0.3894	0.1737	0.5081	0.1820

Cross-Entropy

Metrics#
Implementation	Variant	Recall@1	MRR@1	Recall@5	MRR@5	Recall@10	MRR@10	Recall@20	MRR@20
GRU4Rec Official	Best params	0.1108	0.1108	0.3000	0.1772	0.4126	0.1922	0.5291	0.2003
GRU4Rec Official	Torch-GRU4Rec params	0.0825	0.0825	0.2484	0.1401	0.3551	0.1543	0.4716	0.1624
Torch-GRU4Rec	OOB	0.0814	0.0814	0.2459	0.1383	0.3531	0.1525	0.4688	0.1606

Metric Diff

BPR-Max

Metric difference compared to the “Best params” version with the corresponding loss#
Implementation	Variant	Recall@1 Diff	MRR@1 Diff	Recall@5 Diff	MRR@5 Diff	Recall@10 Diff	MRR@10 Diff	Recall@20 Diff	MRR@20 Diff
GRU4Rec Official	Best params
GRU4Rec Official	Torch-GRU4Rec params	-5.78%	-5.78%	-3.34%	-4.34%	-2.59%	-4.04%	-1.80%	-3.83%
Torch-GRU4Rec	OOB	-7.12%	-7.12%	-4.28%	-5.50%	-3.32%	-5.11%	-2.40%	-4.86%

Cross-Entropy

Metric difference compared to the “Best params” version with the corresponding loss#
Implementation	Variant	Recall@1 Diff	MRR@1 Diff	Recall@5 Diff	MRR@5 Diff	Recall@10 Diff	MRR@10 Diff	Recall@20 Diff	MRR@20 Diff
GRU4Rec Official	Best params
GRU4Rec Official	Torch-GRU4Rec params	-25.58%	-25.58%	-17.21%	-20.94%	-13.94%	-19.73%	-10.87%	-18.94%
Torch-GRU4Rec	OOB	-26.58%	-26.58%	-18.05%	-21.96%	-14.42%	-20.63%	-11.40%	-19.83%

Hyperparameters

BPR-Max

Hyperparameters used in the experiment#
	GRU4Rec Official	GRU4Rec Official	Torch-GRU4Rec
Variant	Best params	Torch-GRU4Rec params	OOB
loss	bpr-max	bpr-max	bpr-max
optim	adagrad	adagrad	adagrad
constrained_embedding	True	False	False
embedding	0	512	512
final_act	elu-0.5	elu-0.5	elu-0.5
layers	512	512	512
batch_size	32	32	32
dropout_p_embed	0.1	0.1	0.1
dropout_p_hidden	0	0	0
learning_rate	0.03	0.03	0.03
momentum	0.55	0	N/A
n_sample	2048	2048	2048
sample_alpha	0.2	0.2	0.2
bpreg	0.75	0.75	0.75
logq	0	0	N/A

Cross-Entropy

Hyperparameters used in the experiment#
	GRU4Rec Official	GRU4Rec Official	Torch-GRU4Rec
Variant	Best params	Torch-GRU4Rec params	OOB
loss	cross-entropy	cross-entropy	cross-entropy
optim	adagrad	adagrad	adagrad
constrained_embedding	True	False	False
embedding	0	512	512
final_act	softmax	softmax	softmax
layers	512	512	512
batch_size	240	240	240
dropout_p_embed	0.45	0.45	0.45
dropout_p_hidden	0	0	0
learning_rate	0.065	0.065	0.065
momentum	0	0	N/A
n_sample	2048	2048	2048
sample_alpha	0.5	0.5	0.5
bpreg	0	0	0
logq	1	0	N/A

Runtimes

BPR-Max

Runtime metrics#
Implementation	Variant	Avg. epoch time (s)	Avg. epoch time to Best	Avg. epoch time to Matching	Avg. mb/s	Avg. e/s
GRU4Rec Official	Best params	1956.80			916.45	29326.00
GRU4Rec Official	Torch-GRU4Rec params	1816.76	0.93 x		987.09	31587.00
Torch-GRU4Rec	OOB	30689.54	15.68 x	16.89 x	58.43	1869.86

Cross-Entropy

Runtime metrics#
Implementation	Variant	Avg. epoch time (s)	Avg. epoch time to Best	Avg. epoch time to Matching	Avg. mb/s	Avg. e/s
GRU4Rec Official	Best params	367.41			650.91	156189.00
GRU4Rec Official	Torch-GRU4Rec params	381.71	1.04 x		626.53	150339.00
Torch-GRU4Rec	OOB	7192.88	19.58 x	18.84 x	33.25	7978.05

Yoochoose#

Metric

BPR-Max

Metrics#
Implementation	Variant	Recall@1	MRR@1	Recall@5	MRR@5	Recall@10	MRR@10	Recall@20	MRR@20
GRU4Rec Official	Best params	0.1745	0.1745	0.4346	0.2675	0.5664	0.2851	0.6799	0.2931
GRU4Rec Official	Torch-GRU4Rec params	0.1748	0.1748	0.4298	0.2662	0.5603	0.2837	0.6769	0.2919
Torch-GRU4Rec	OOB	0.1755	0.1755	0.4271	0.2654	0.5560	0.2826	0.6711	0.2907

Cross-Entropy

Metrics#
Implementation	Variant	Recall@1	MRR@1	Recall@5	MRR@5	Recall@10	MRR@10	Recall@20	MRR@20
GRU4Rec Official	Best params	0.1797	0.1797	0.4457	0.2757	0.5698	0.2924	0.6804	0.3002
GRU4Rec Official	Torch-GRU4Rec params	0.1710	0.1710	0.4301	0.2633	0.5571	0.2803	0.6690	0.2882
Torch-GRU4Rec	OOB	0.1686	0.1686	0.4268	0.2598	0.5528	0.2768	0.6671	0.2847

Metric Diff

BPR-Max

Metric difference compared to the “Best params” version with the corresponding loss#
Implementation	Variant	Recall@1 Diff	MRR@1 Diff	Recall@5 Diff	MRR@5 Diff	Recall@10 Diff	MRR@10 Diff	Recall@20 Diff	MRR@20 Diff
GRU4Rec Official	Best params
GRU4Rec Official	Torch-GRU4Rec params	0.14%	0.14%	-1.09%	-0.49%	-1.07%	-0.48%	-0.45%	-0.39%
Torch-GRU4Rec	OOB	0.53%	0.53%	-1.72%	-0.77%	-1.84%	-0.89%	-1.30%	-0.82%

Cross-Entropy

Metric difference compared to the “Best params” version with the corresponding loss#
Implementation	Variant	Recall@1 Diff	MRR@1 Diff	Recall@5 Diff	MRR@5 Diff	Recall@10 Diff	MRR@10 Diff	Recall@20 Diff	MRR@20 Diff
GRU4Rec Official	Best params
GRU4Rec Official	Torch-GRU4Rec params	-4.81%	-4.81%	-3.50%	-4.52%	-2.23%	-4.14%	-1.67%	-4.00%
Torch-GRU4Rec	OOB	-6.14%	-6.14%	-4.25%	-5.77%	-2.99%	-5.35%	-1.95%	-5.14%

Hyperparameters

BPR-Max

Hyperparameters used in the experiment#
	GRU4Rec Official	GRU4Rec Official	Torch-GRU4Rec
Variant	Best params	Torch-GRU4Rec params	OOB
loss	bpr-max	bpr-max	bpr-max
optim	adagrad	adagrad	adagrad
constrained_embedding	True	False	False
embedding	0	448	448
final_act	linear	linear	linear
layers	448	448	448
batch_size	48	48	48
dropout_p_embed	0.25	0.25	0.25
dropout_p_hidden	0	0	0
learning_rate	0.075	0.075	0.075
momentum	0.1	0	N/A
n_sample	2048	2048	2048
sample_alpha	0.2	0.2	0.2
bpreg	0.5	0.5	0.5
logq	0	0	N/A

Cross-Entropy

Hyperparameters used in the experiment#
	GRU4Rec Official	GRU4Rec Official	Torch-GRU4Rec
Variant	Best params	Torch-GRU4Rec params	OOB
loss	cross-entropy	cross-entropy	cross-entropy
optim	adagrad	adagrad	adagrad
constrained_embedding	True	False	False
embedding	0	480	480
final_act	softmax	softmax	softmax
layers	480	480	480
batch_size	48	48	48
dropout_p_embed	0	0	0
dropout_p_hidden	0.2	0.2	0.2
learning_rate	0.07	0.07	0.07
momentum	0	0	N/A
n_sample	2048	2048	2048
sample_alpha	0.2	0.2	0.2
bpreg	0	0	0
logq	1	0	N/A

Runtimes

BPR-Max

Runtime metrics#
Implementation	Variant	Avg. epoch time (s)	Avg. epoch time to Best	Avg. epoch time to Matching	Avg. mb/s	Avg. e/s
GRU4Rec Official	Best params	487.51			919.23	44121.00
GRU4Rec Official	Torch-GRU4Rec params	461.22	0.95 x		971.64	46636.00
Torch-GRU4Rec	OOB	2164.21	4.44 x	4.69 x	207.09	9939.82

Cross-Entropy

Runtime metrics#
Implementation	Variant	Avg. epoch time (s)	Avg. epoch time to Best	Avg. epoch time to Matching	Avg. mb/s	Avg. e/s
GRU4Rec Official	Best params	451.75			991.99	47613.00
GRU4Rec Official	Torch-GRU4Rec params	439.19	0.97 x		1020.37	48975.00
Torch-GRU4Rec	OOB	2082.11	4.61 x	4.74 x	215.24	10330.85

Coveo#

Metric

BPR-Max

Metrics#
Implementation	Variant	Recall@1	MRR@1	Recall@5	MRR@5	Recall@10	MRR@10	Recall@20	MRR@20
GRU4Rec Official	Best params	0.0501	0.0501	0.1464	0.0835	0.2172	0.0928	0.3123	0.0994
GRU4Rec Official	Torch-GRU4Rec params	0.0484	0.0484	0.1430	0.0811	0.2107	0.0901	0.2995	0.0962
Torch-GRU4Rec	OOB	0.0479	0.0479	0.1400	0.0794	0.2067	0.0883	0.2960	0.0944

Cross-Entropy

Metrics#
Implementation	Variant	Recall@1	MRR@1	Recall@5	MRR@5	Recall@10	MRR@10	Recall@20	MRR@20
GRU4Rec Official	Best params	0.0489	0.0489	0.1418	0.0814	0.2085	0.0901	0.2947	0.0960
GRU4Rec Official	Torch-GRU4Rec params	0.0473	0.0473	0.1379	0.0786	0.2015	0.0871	0.2848	0.0928
Torch-GRU4Rec	OOB	0.0445	0.0445	0.1343	0.0755	0.1949	0.0835	0.2835	0.0897

Metric Diff

BPR-Max

Metric difference compared to the “Best params” version with the corresponding loss#
Implementation	Variant	Recall@1 Diff	MRR@1 Diff	Recall@5 Diff	MRR@5 Diff	Recall@10 Diff	MRR@10 Diff	Recall@20 Diff	MRR@20 Diff
GRU4Rec Official	Best params
GRU4Rec Official	Torch-GRU4Rec params	-3.52%	-3.52%	-2.27%	-2.86%	-2.98%	-2.99%	-4.09%	-3.21%
Torch-GRU4Rec	OOB	-4.43%	-4.43%	-4.35%	-4.88%	-4.84%	-4.92%	-5.23%	-5.01%

Cross-Entropy

Metric difference compared to the “Best params” version with the corresponding loss#
Implementation	Variant	Recall@1 Diff	MRR@1 Diff	Recall@5 Diff	MRR@5 Diff	Recall@10 Diff	MRR@10 Diff	Recall@20 Diff	MRR@20 Diff
GRU4Rec Official	Best params
GRU4Rec Official	Torch-GRU4Rec params	-3.29%	-3.29%	-2.74%	-3.34%	-3.35%	-3.38%	-3.37%	-3.37%
Torch-GRU4Rec	OOB	-8.98%	-8.98%	-5.30%	-7.14%	-6.55%	-7.33%	-3.80%	-6.61%

Hyperparameters

BPR-Max

Hyperparameters used in the experiment#
	GRU4Rec Official	GRU4Rec Official	Torch-GRU4Rec
Variant	Best params	Torch-GRU4Rec params	OOB
loss	bpr-max	bpr-max	bpr-max
optim	adagrad	adagrad	adagrad
constrained_embedding	True	False	False
embedding	0	512	512
final_act	elu-1	elu-1	elu-1
layers	512	512	512
batch_size	144	144	144
dropout_p_embed	0.35	0.35	0.35
dropout_p_hidden	0	0	0
learning_rate	0.05	0.05	0.05
momentum	0.4	0	N/A
n_sample	2048	2048	2048
sample_alpha	0.2	0.2	0.2
bpreg	1.85	1.85	1.85
logq	0	0	N/A

Cross-Entropy

Hyperparameters used in the experiment#
	GRU4Rec Official	GRU4Rec Official	Torch-GRU4Rec
Variant	Best params	Torch-GRU4Rec params	OOB
loss	cross-entropy	cross-entropy	cross-entropy
optim	adagrad	adagrad	adagrad
constrained_embedding	True	False	False
embedding	0	512	512
final_act	softmax	softmax	softmax
layers	512	512	512
batch_size	32	32	32
dropout_p_embed	0.4	0.4	0.4
dropout_p_hidden	0.15	0.15	0.15
learning_rate	0.03	0.03	0.03
momentum	0	0	N/A
n_sample	2048	2048	2048
sample_alpha	0	0	0
bpreg	0	0	0
logq	1	0	N/A

Runtimes

BPR-Max

Runtime metrics#
Implementation	Variant	Avg. epoch time (s)	Avg. epoch time to Best	Avg. epoch time to Matching	Avg. mb/s	Avg. e/s
GRU4Rec Official	Best params	12.38			704.14	100615.00
GRU4Rec Official	Torch-GRU4Rec params	12.23	0.99 x		712.45	101803.00
Torch-GRU4Rec	OOB	32.18	2.60 x	2.63 x	270.97	38719.00

Cross-Entropy

Runtime metrics#
Implementation	Variant	Avg. epoch time (s)	Avg. epoch time to Best	Avg. epoch time to Matching	Avg. mb/s	Avg. e/s
GRU4Rec Official	Best params	37.17			1047.47	33505.00
GRU4Rec Official	Torch-GRU4Rec params	36.81	0.99 x		1057.66	33830.00
Torch-GRU4Rec	OOB	91.28	2.46 x	2.48 x	426.78	13650.91

Retailrocket#

Metric

BPR-Max

Metrics#
Implementation	Variant	Recall@1	MRR@1	Recall@5	MRR@5	Recall@10	MRR@10	Recall@20	MRR@20
GRU4Rec Official	Best params	0.1224	0.1224	0.3196	0.1928	0.4181	0.2060	0.5187	0.2131
GRU4Rec Official	Torch-GRU4Rec params	0.0979	0.0979	0.2672	0.1576	0.3631	0.1703	0.4639	0.1773
Torch-GRU4Rec	OOB	0.0926	0.0926	0.2413	0.1451	0.3336	0.1574	0.4201	0.1633

Cross-Entropy

Metrics#
Implementation	Variant	Recall@1	MRR@1	Recall@5	MRR@5	Recall@10	MRR@10	Recall@20	MRR@20
GRU4Rec Official	Best params	0.1150	0.1150	0.3026	0.1814	0.4019	0.1947	0.4953	0.2012
GRU4Rec Official	Torch-GRU4Rec params	0.0907	0.0907	0.2462	0.1456	0.3331	0.1571	0.4249	0.1635
Torch-GRU4Rec	OOB	0.0806	0.0806	0.2206	0.1300	0.2994	0.1405	0.3834	0.1464

Metric Diff

BPR-Max

Metric difference compared to the “Best params” version with the corresponding loss#
Implementation	Variant	Recall@1 Diff	MRR@1 Diff	Recall@5 Diff	MRR@5 Diff	Recall@10 Diff	MRR@10 Diff	Recall@20 Diff	MRR@20 Diff
GRU4Rec Official	Best params
GRU4Rec Official	Torch-GRU4Rec params	-20.04%	-20.04%	-16.41%	-18.23%	-13.15%	-17.33%	-10.56%	-16.77%
Torch-GRU4Rec	OOB	-24.35%	-24.35%	-24.51%	-24.70%	-20.22%	-23.61%	-19.00%	-23.35%

Cross-Entropy

Metric difference compared to the “Best params” version with the corresponding loss#
Implementation	Variant	Recall@1 Diff	MRR@1 Diff	Recall@5 Diff	MRR@5 Diff	Recall@10 Diff	MRR@10 Diff	Recall@20 Diff	MRR@20 Diff
GRU4Rec Official	Best params
GRU4Rec Official	Torch-GRU4Rec params	-21.12%	-21.12%	-18.63%	-19.73%	-17.11%	-19.29%	-14.20%	-18.73%
Torch-GRU4Rec	OOB	-29.89%	-29.89%	-27.10%	-28.36%	-25.50%	-27.83%	-22.60%	-27.27%

Hyperparameters

BPR-Max

Hyperparameters used in the experiment#
	GRU4Rec Official	GRU4Rec Official	Torch-GRU4Rec
Variant	Best params	Torch-GRU4Rec params	OOB
loss	bpr-max	bpr-max	bpr-max
optim	adagrad	adagrad	adagrad
constrained_embedding	True	False	False
embedding	0	224	224
final_act	elu-0.5	elu-0.5	elu-0.5
layers	224	224	224
batch_size	80	80	80
dropout_p_embed	0.5	0.5	0.5
dropout_p_hidden	0.05	0.05	0.05
learning_rate	0.05	0.05	0.05
momentum	0.4	0	N/A
n_sample	2048	2048	2048
sample_alpha	0.4	0.4	0.4
bpreg	1.95	1.95	1.95
logq	0	0	N/A

Cross-Entropy

Hyperparameters used in the experiment#
	GRU4Rec Official	GRU4Rec Official	Torch-GRU4Rec
Variant	Best params	Torch-GRU4Rec params	OOB
loss	cross-entropy	cross-entropy	cross-entropy
optim	adagrad	adagrad	adagrad
constrained_embedding	True	False	False
embedding	0	192	192
final_act	softmax	softmax	softmax
layers	192	192	192
batch_size	240	240	240
dropout_p_embed	0.5	0.5	0.5
dropout_p_hidden	0.05	0.05	0.05
learning_rate	0.085	0.085	0.085
momentum	0.3	0	N/A
n_sample	2048	2048	2048
sample_alpha	0.3	0.3	0.3
bpreg	0	0	0
logq	1	0	N/A

Runtimes

BPR-Max

Runtime metrics#
Implementation	Variant	Avg. epoch time (s)	Avg. epoch time to Best	Avg. epoch time to Matching	Avg. mb/s	Avg. e/s
GRU4Rec Official	Best params	6.86			1019.34	80807.00
GRU4Rec Official	Torch-GRU4Rec params	6.53	0.95 x		1071.09	84909.00
Torch-GRU4Rec	OOB	26.39	3.85 x	4.04 x	265.19	21022.30

Cross-Entropy

Runtime metrics#
Implementation	Variant	Avg. epoch time (s)	Avg. epoch time to Best	Avg. epoch time to Matching	Avg. mb/s	Avg. e/s
GRU4Rec Official	Best params	2.77			880.71	199935.00
GRU4Rec Official	Torch-GRU4Rec params	2.45	0.88 x		996.78	226283.00
Torch-GRU4Rec	OOB	10.67	3.85 x	4.36 x	229.70	52144.58

Diginetica#

Metric

BPR-Max

Metrics#
Implementation	Variant	Recall@1	MRR@1	Recall@5	MRR@5	Recall@10	MRR@10	Recall@20	MRR@20
GRU4Rec Official	Best params	0.0688	0.0688	0.2304	0.1237	0.3533	0.1399	0.4995	0.1500
GRU4Rec Official	Torch-GRU4Rec params	0.0643	0.0643	0.2113	0.1143	0.3204	0.1287	0.4597	0.1383
Torch-GRU4Rec	OOB	0.0636	0.0636	0.2110	0.1137	0.3185	0.1278	0.4550	0.1372

Cross-Entropy

Metrics#
Implementation	Variant	Recall@1	MRR@1	Recall@5	MRR@5	Recall@10	MRR@10	Recall@20	MRR@20
GRU4Rec Official	Best params	0.0647	0.0647	0.2220	0.1181	0.3414	0.1339	0.4874	0.1440
GRU4Rec Official	Torch-GRU4Rec params	0.0552	0.0552	0.1927	0.1020	0.2967	0.1157	0.4255	0.1246
Torch-GRU4Rec	OOB	0.0541	0.0541	0.1894	0.1002	0.2921	0.1137	0.4245	0.1229

Metric Diff

BPR-Max

Metric difference compared to the “Best params” version with the corresponding loss#
Implementation	Variant	Recall@1 Diff	MRR@1 Diff	Recall@5 Diff	MRR@5 Diff	Recall@10 Diff	MRR@10 Diff	Recall@20 Diff	MRR@20 Diff
GRU4Rec Official	Best params
GRU4Rec Official	Torch-GRU4Rec params	-6.58%	-6.58%	-8.30%	-7.53%	-9.32%	-7.97%	-7.97%	-7.76%
Torch-GRU4Rec	OOB	-7.64%	-7.64%	-8.42%	-8.08%	-9.85%	-8.63%	-8.92%	-8.54%

Cross-Entropy

Metric difference compared to the “Best params” version with the corresponding loss#
Implementation	Variant	Recall@1 Diff	MRR@1 Diff	Recall@5 Diff	MRR@5 Diff	Recall@10 Diff	MRR@10 Diff	Recall@20 Diff	MRR@20 Diff
GRU4Rec Official	Best params
GRU4Rec Official	Torch-GRU4Rec params	-14.78%	-14.78%	-13.17%	-13.69%	-13.08%	-13.55%	-12.70%	-13.47%
Torch-GRU4Rec	OOB	-16.45%	-16.45%	-14.67%	-15.19%	-14.44%	-15.04%	-12.90%	-14.71%

Hyperparameters

BPR-Max

Hyperparameters used in the experiment#
	GRU4Rec Official	GRU4Rec Official	Torch-GRU4Rec
Variant	Best params	Torch-GRU4Rec params	OOB
loss	bpr-max	bpr-max	bpr-max
optim	adagrad	adagrad	adagrad
constrained_embedding	True	False	False
embedding	0	512	512
final_act	elu-1	elu-1	elu-1
layers	512	512	512
batch_size	128	128	128
dropout_p_embed	0.5	0.5	0.5
dropout_p_hidden	0.3	0.3	0.3
learning_rate	0.05	0.05	0.05
momentum	0.15	0	N/A
n_sample	2048	2048	2048
sample_alpha	0.3	0.3	0.3
bpreg	0.9	0.9	0.9
logq	0	0	N/A

Cross-Entropy

Hyperparameters used in the experiment#
	GRU4Rec Official	GRU4Rec Official	Torch-GRU4Rec
Variant	Best params	Torch-GRU4Rec params	OOB
loss	cross-entropy	cross-entropy	cross-entropy
optim	adagrad	adagrad	adagrad
constrained_embedding	True	False	False
embedding	0	192	192
final_act	softmax	softmax	softmax
layers	192	192	192
batch_size	128	128	128
dropout_p_embed	0.45	0.45	0.45
dropout_p_hidden	0.15	0.15	0.15
learning_rate	0.1	0.1	0.1
momentum	0	0	N/A
n_sample	2048	2048	2048
sample_alpha	0	0	0
bpreg	0	0	0
logq	1	0	N/A

Runtimes

BPR-Max

Runtime metrics#
Implementation	Variant	Avg. epoch time (s)	Avg. epoch time to Best	Avg. epoch time to Matching	Avg. mb/s	Avg. e/s
GRU4Rec Official	Best params	8.02			639.87	81757.00
GRU4Rec Official	Torch-GRU4Rec params	7.71	0.96 x		665.42	85021.00
Torch-GRU4Rec	OOB	36.89	4.60 x	4.78 x	139.19	17783.97

Cross-Entropy

Runtime metrics#
Implementation	Variant	Avg. epoch time (s)	Avg. epoch time to Best	Avg. epoch time to Matching	Avg. mb/s	Avg. e/s
GRU4Rec Official	Best params	4.52			1134.52	144959.00
GRU4Rec Official	Torch-GRU4Rec params	4.48	0.99 x		1146.91	146542.00
Torch-GRU4Rec	OOB	17.74	3.92 x	3.96 x	289.38	36974.80

Torch-GRU4Rec

Contents

Torch-GRU4Rec#

Rees46#

Yoochoose#

Coveo#

Retailrocket#

Diginetica#