issues
search
facebookresearch
/
generative-recommenders
Repository hosting code used to reproduce results in "Actions Speak Louder than Words: Trillion-Parameter Sequential Transducers for Generative Recommendations" (https://arxiv.org/abs/2402.17152).
Apache License 2.0
745
stars
141
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
copy complete_cumsum kernel to ops/cpp/
#127
jiyuanzFB
opened
1 day ago
1
Redefine FBGEMM targets with gpu_cpp_library (Re-land attempt of D64863809)
#126
q10
opened
1 day ago
1
num_stages=0 becomes num_stages=2
#125
nmacchioni
opened
1 day ago
3
Fix max_attn_len numerical
#124
hanli0612
opened
2 days ago
1
Change autotune key for _AddTimestampPositionEmbeddingsFunction to support dynamic batch size
#123
AlbertDachiChen
closed
1 day ago
3
fix overflow
#122
LinjianMa
closed
1 day ago
1
re-enable lint for hammer/oss
#121
LinjianMa
closed
2 days ago
1
: Remove attn_scale
#120
xing-liu
closed
1 day ago
4
: Add concat_1d_jagged_jagged
#119
xing-liu
closed
1 day ago
2
Fix lock in hstu kernel
#118
yunjiangster
closed
1 day ago
5
Triton autotuner also takes pre_hook/post_hook as args, changing the arg order.
#117
SamGinzburg
closed
1 day ago
1
Update README.md to include instructions on using synthetic datasets
#116
ruochen99
closed
1 day ago
1
How to install hammer package used in triton_ragged_hstu_attention?
#115
Lix1993
opened
1 week ago
1
Questions Regarding the Implementation of the GR Ranking Model
#114
shijieliu
opened
1 week ago
2
: Add permute_4d_jagged_2013 cufa kernel
#113
xing-liu
opened
1 week ago
1
Update input_features_preprocessors.py
#112
Lix1993
closed
1 week ago
3
Move triton_position to oss
#111
ruochen99
closed
1 week ago
3
Cleanup tl.constexpr HAS_ATTN_SCALE
#110
hanli0612
closed
1 day ago
5
add specs for amd_standalone_cint_v2
#109
cx-yin
closed
1 day ago
9
Move triton_hstu_linear to oss
#108
ruochen99
closed
1 week ago
7
Move triton_jagged to oss
#107
ruochen99
closed
1 week ago
3
Move triton_layer_norm to oss
#106
ruochen99
closed
1 week ago
4
use autotune_max_seq_len to choose hstu bwd path
#105
LinjianMa
closed
1 week ago
1
Could you provide a config file for ranking?
#104
Lix1993
closed
1 week ago
1
Ranking task
#103
bityigoss
closed
2 weeks ago
0
Create a directory for cuda and triton kernels
#102
xing-liu
closed
2 weeks ago
12
remove the second call to movielens_seq_features_from_row，
#101
Lix1993
closed
3 weeks ago
3
Redundant call to movielens_seq_features_from_row in training loop
#100
Lix1993
closed
3 weeks ago
0
Support sorting input by sequence lengths
#99
ruochen99
closed
2 weeks ago
2
Full Contextual Mask for HSTU
#98
lic225
closed
4 weeks ago
1
Update README.md to include tensorboard in pip install command
#97
Lix1993
closed
4 weeks ago
4
Do not pipeline last inner loop.
#96
htyu
closed
2 weeks ago
1
How to deal with unknown tokens in practice?
#95
isofun
opened
1 month ago
2
Evaluation on public dataset
#94
Blank-z0
opened
1 month ago
1
change addmm grid size to 2D to overcome uint16 limitation
#93
jiyuanzFB
closed
2 weeks ago
2
Enable HSTU SMEM for TW and PW with autotuning for both SMEM preload and maxnregs
#92
plotfi
opened
1 month ago
7
Integration with TorchRec
#91
sbhavani
closed
4 weeks ago
1
Try to use triton.language.extra.libdevice when possible
#90
danzimm
closed
1 month ago
5
Raise precision bar for mrn_compute_output
#89
mrshenli
closed
1 month ago
2
confused about "SampledSoftmaxLoss" func
#88
zhhu1996
opened
1 month ago
3
add arg AUTOTUNE_MAX_SEQ_LEN to attn kernel
#87
LinjianMa
closed
1 month ago
1
Add type error suppressions for upcoming upgrade
#86
MaggieMoss
closed
1 month ago
1
ItemFeatures not used in training
#85
rubydevelop
closed
2 months ago
2
Add type error suppressions for upcoming upgrade
#84
MaggieMoss
closed
1 month ago
1
Fix for persistent kernel issue with early return
#83
narayanan2004
closed
2 months ago
3
Add type error suppressions for upcoming upgrade
#82
MaggieMoss
closed
1 month ago
2
add fractal expansion algorithm + ml-3B dataset interface
#81
LinjianMa
closed
2 months ago
1
Add type error suppressions for upcoming upgrade
#80
MaggieMoss
closed
2 months ago
1
[Question] attention score scaling and gr_output_length
#79
JacoCheung
closed
2 months ago
1
Allow broadcasting in addmm
#78
ruochen99
closed
2 months ago
3
Next