Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / facebookresearch/generative-recommenders issues and pull requests

#179 - Environment issue

Issue - State: closed - Opened by mefor44 10 days ago - 4 comments

#178 - HSTU streaming training

Issue - State: open - Opened by WangPengGG 15 days ago - 1 comment

#177 - Installing requirements results in Python version conflict.

Issue - State: open - Opened by huzhiq94 17 days ago - 2 comments

#176 - Could not find TensorRT

Issue - State: open - Opened by X-yang03 23 days ago - 1 comment

#175 - How to use cache?

Issue - State: open - Opened by liang23333 25 days ago - 1 comment

#174 - Autocast support for half precision

Issue - State: open - Opened by majing921201 about 1 month ago - 3 comments

#172 - add more ops

Pull Request - State: closed - Opened by xing-liu about 2 months ago
Labels: CLA Signed

#171 - Sequence format

Issue - State: closed - Opened by Xinhui-Zhu about 2 months ago - 2 comments

#170 - TritonCC for the new hstu_attention kernel

Pull Request - State: closed - Opened by ruochen99 about 2 months ago - 2 comments
Labels: CLA Signed, fb-exported

#169 - remove triton specs from oss

Pull Request - State: closed - Opened by LinjianMa about 2 months ago - 2 comments
Labels: CLA Signed, fb-exported

#168 - Add back has_contextual_seq_len, add TritonCC support for UC mask

Pull Request - State: closed - Opened by ruochen99 about 2 months ago - 3 comments
Labels: CLA Signed, fb-exported

#167 - move triton wrappers to gr

Pull Request - State: closed - Opened by LinjianMa about 2 months ago - 5 comments
Labels: CLA Signed, fb-exported

#166 - add cpp ops

Pull Request - State: closed - Opened by LinjianMa about 2 months ago - 3 comments
Labels: CLA Signed, fb-exported

#165 - Clean up HSTU attention ops

Pull Request - State: closed - Opened by xing-liu about 2 months ago - 9 comments
Labels: CLA Signed, fb-exported

#164 - OSS HSTU attention ops

Pull Request - State: closed - Opened by xing-liu about 2 months ago - 3 comments
Labels: CLA Signed, fb-exported

#163 - remove oss

Pull Request - State: closed - Opened by LinjianMa about 2 months ago - 2 comments
Labels: CLA Signed, fb-exported

#162 - Use default warmpup and rep

Pull Request - State: closed - Opened by zoranzhao about 2 months ago - 1 comment
Labels: CLA Signed, fb-exported

#161 - Replace Fbgemm related operators with Pytorch aten operators

Issue - State: open - Opened by majing921201 2 months ago - 2 comments

#160 - fix lint

Pull Request - State: closed - Opened by LinjianMa 2 months ago - 1 comment
Labels: CLA Signed, fb-exported

#159 - tritoncc fp16 support

Pull Request - State: closed - Opened by LinjianMa 2 months ago - 2 comments
Labels: CLA Signed, fb-exported

#158 - Replace tl.exp() with tl.exp2() for better attention kernel performance

Pull Request - State: closed - Opened by yoyoyocmu 2 months ago - 1 comment
Labels: CLA Signed, fb-exported

#157 - add fp16 to hstu kernel tests

Pull Request - State: closed - Opened by LinjianMa 2 months ago - 1 comment
Labels: CLA Signed, fb-exported

#156 - Enable UC Mask on TritonCC

Pull Request - State: closed - Opened by lic225 2 months ago - 2 comments
Labels: CLA Signed, fb-exported, Reverted

#155 - : Fix contextual_seq_len and unit tests

Pull Request - State: closed - Opened by xing-liu 2 months ago - 3 comments
Labels: CLA Signed, fb-exported

#154 - simplify hstu bwd autotune

Pull Request - State: closed - Opened by yaoyj11 2 months ago - 1 comment
Labels: CLA Signed, fb-exported

#153 - Auto-tune config for HSTU attention forward

Pull Request - State: closed - Opened by yoyoyocmu 2 months ago - 1 comment
Labels: CLA Signed, fb-exported

#152 - Fix OSS import in triton_ragged_hstu_attention

Pull Request - State: closed - Opened by bertmaher 2 months ago - 2 comments
Labels: CLA Signed, fb-exported

#150 - Update README.md to reflect latest codebase status

Pull Request - State: closed - Opened by jiaqizhai 2 months ago - 2 comments
Labels: CLA Signed, Merged

#148 - What is gr_output_length specifically used for?

Issue - State: open - Opened by yudonglee 2 months ago - 1 comment

#147 - (2/2) Simplifies similarity fn-related code in GR based on the latest MoL/RAILS implementations

Pull Request - State: closed - Opened by jiaqizhai 2 months ago - 4 comments
Labels: CLA Signed, Merged

#146 - Fix arguments for triton:cc to match upstream

Pull Request - State: closed - Opened by SamGinzburg 2 months ago - 6 comments
Labels: CLA Signed, fb-exported

#145 - update block size for standalone_cint_v4

Pull Request - State: closed - Opened by zhaozhul 3 months ago - 2 comments
Labels: CLA Signed, fb-exported

#144 - standalone_cint_v4

Pull Request - State: closed - Opened by zhaozhul 3 months ago - 4 comments
Labels: CLA Signed, fb-exported

#143 - loop unroll for hstu attn bwd

Pull Request - State: closed - Opened by LinjianMa 3 months ago - 2 comments
Labels: CLA Signed, fb-exported

#142 - Convert directory fbcode/hammer to use the Ruff Formatter

Pull Request - State: closed - Opened by tpolasek 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#141 - Prepare for "Fix type-safety of `torch.nn.Module` instances": fbcode/h*

Pull Request - State: closed - Opened by ezyang 3 months ago - 3 comments
Labels: CLA Signed, fb-exported

#140 - Add complete_cumsum cpu and meta ops

Pull Request - State: closed - Opened by jiyuanzFB 3 months ago - 2 comments
Labels: CLA Signed, fb-exported

#139 - Cleanup tl.constexpr HAS_CONTEXTUAL_SEQ_LEN

Pull Request - State: closed - Opened by hanli0612 3 months ago - 2 comments
Labels: CLA Signed, fb-exported

#138 - restore missing fbgemm_gpu import

Pull Request - State: closed - Opened by Lix1993 3 months ago - 2 comments
Labels: CLA Signed, Merged

#137 - add pytorch implementations for jagged operations

Pull Request - State: closed - Opened by zhaozhul 3 months ago - 4 comments
Labels: CLA Signed, fb-exported

#136 - restore missing fbgemm_gpu import

Pull Request - State: closed - Opened by Lix1993 3 months ago
Labels: CLA Signed

#135 - (1/2) Simplifies similarity fn-related code in the GR codebase based on the latest MoL/RAILS implementations.

Pull Request - State: closed - Opened by jiaqizhai 3 months ago - 4 comments
Labels: CLA Signed, Merged

#134 - run error after remove fbgemm_gpu import

Issue - State: closed - Opened by Lix1993 3 months ago

#133 - Replace Triton addmm with` torch.addmm` for AMD to achieve better training performance

Pull Request - State: closed - Opened by yoyoyocmu 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#132 - Change autotune key for ragged_hstu_attention to support dynamic batch size

Pull Request - State: closed - Opened by AlbertDachiChen 3 months ago - 10 comments
Labels: CLA Signed, fb-exported

#131 - add pytorch implementation of layer_norm

Pull Request - State: closed - Opened by zhaozhul 3 months ago - 4 comments
Labels: CLA Signed, fb-exported, Merged

#130 - Bug fix: detecting contextual length in triton hstu attn code

Pull Request - State: closed - Opened by lic225 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#130 - Bug fix: detecting contextual length in triton hstu attn code

Pull Request - State: open - Opened by lic225 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#129 - Fix type-safety of `torch.nn.Module` instances

Pull Request - State: closed - Opened by ezyang 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#128 - move sorted_kv_pairs from hammer/ops/cuda/ to hammer/ops/cpp/

Pull Request - State: closed - Opened by jiyuanzFB 3 months ago - 4 comments
Labels: CLA Signed, fb-exported

#127 - copy complete_cumsum kernel to ops/cpp/

Pull Request - State: closed - Opened by jiyuanzFB 3 months ago - 3 comments
Labels: CLA Signed, fb-exported

#126 - Redefine FBGEMM targets with gpu_cpp_library (Re-land attempt of D64863809)

Pull Request - State: closed - Opened by q10 3 months ago - 2 comments
Labels: CLA Signed, fb-exported

#126 - Redefine FBGEMM targets with gpu_cpp_library (Re-land attempt of D64863809)

Pull Request - State: open - Opened by q10 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#125 - num_stages=0 becomes num_stages=2

Pull Request - State: open - Opened by nmacchioni 3 months ago - 3 comments
Labels: CLA Signed, fb-exported

#125 - num_stages=0 becomes num_stages=2

Pull Request - State: closed - Opened by nmacchioni 3 months ago - 3 comments
Labels: CLA Signed, fb-exported

#124 - Fix max_attn_len numerical

Pull Request - State: open - Opened by hanli0612 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#124 - Fix max_attn_len numerical

Pull Request - State: closed - Opened by hanli0612 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#123 - Change autotune key for _AddTimestampPositionEmbeddingsFunction to support dynamic batch size

Pull Request - State: closed - Opened by AlbertDachiChen 3 months ago - 3 comments
Labels: CLA Signed, fb-exported

#122 - fix overflow

Pull Request - State: closed - Opened by LinjianMa 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#122 - fix overflow

Pull Request - State: closed - Opened by LinjianMa 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#121 - re-enable lint for hammer/oss

Pull Request - State: closed - Opened by LinjianMa 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#120 - : Remove attn_scale

Pull Request - State: closed - Opened by xing-liu 3 months ago - 4 comments
Labels: CLA Signed, fb-exported

#119 - : Add concat_1d_jagged_jagged

Pull Request - State: closed - Opened by xing-liu 3 months ago - 2 comments
Labels: CLA Signed, fb-exported

#118 - Fix lock in hstu kernel

Pull Request - State: closed - Opened by yunjiangster 3 months ago - 5 comments
Labels: CLA Signed, fb-exported

#117 - Triton autotuner also takes pre_hook/post_hook as args, changing the arg order.

Pull Request - State: closed - Opened by SamGinzburg 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#116 - Update README.md to include instructions on using synthetic datasets

Pull Request - State: open - Opened by ruochen99 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#116 - Update README.md to include instructions on using synthetic datasets

Pull Request - State: closed - Opened by ruochen99 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#115 - How to install hammer package used in triton_ragged_hstu_attention?

Issue - State: closed - Opened by Lix1993 3 months ago - 1 comment

#115 - How to install hammer package used in triton_ragged_hstu_attention?

Issue - State: open - Opened by Lix1993 3 months ago - 1 comment

#114 - Questions Regarding the Implementation of the GR Ranking Model

Issue - State: closed - Opened by shijieliu 3 months ago - 3 comments

#113 - : Add permute_4d_jagged_2013 cufa kernel

Pull Request - State: closed - Opened by xing-liu 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#112 - Update input_features_preprocessors.py

Pull Request - State: closed - Opened by Lix1993 3 months ago - 3 comments
Labels: CLA Signed, Merged

#111 - Move triton_position to oss

Pull Request - State: closed - Opened by ruochen99 3 months ago - 3 comments
Labels: CLA Signed, fb-exported

#110 - Cleanup tl.constexpr HAS_ATTN_SCALE

Pull Request - State: closed - Opened by hanli0612 3 months ago - 5 comments
Labels: CLA Signed, fb-exported

#109 - add specs for amd_standalone_cint_v2

Pull Request - State: closed - Opened by cx-yin 3 months ago - 9 comments
Labels: CLA Signed, fb-exported

#108 - Move triton_hstu_linear to oss

Pull Request - State: closed - Opened by ruochen99 3 months ago - 7 comments
Labels: CLA Signed, fb-exported

#107 - Move triton_jagged to oss

Pull Request - State: closed - Opened by ruochen99 3 months ago - 3 comments
Labels: CLA Signed, fb-exported

#106 - Move triton_layer_norm to oss

Pull Request - State: closed - Opened by ruochen99 3 months ago - 4 comments
Labels: CLA Signed, fb-exported

#105 - use autotune_max_seq_len to choose hstu bwd path

Pull Request - State: closed - Opened by LinjianMa 3 months ago - 1 comment
Labels: CLA Signed, fb-exported

#104 - Could you provide a config file for ranking?

Issue - State: closed - Opened by Lix1993 3 months ago - 1 comment

#103 - Ranking task

Issue - State: closed - Opened by bityigoss 4 months ago

#102 - Create a directory for cuda and triton kernels

Pull Request - State: closed - Opened by xing-liu 4 months ago - 12 comments
Labels: CLA Signed, fb-exported

#101 - remove the second call to movielens_seq_features_from_row，

Pull Request - State: closed - Opened by Lix1993 4 months ago - 3 comments
Labels: CLA Signed, Merged

#99 - Support sorting input by sequence lengths

Pull Request - State: closed - Opened by ruochen99 4 months ago - 2 comments
Labels: CLA Signed, fb-exported

#98 - Full Contextual Mask for HSTU

Pull Request - State: closed - Opened by lic225 4 months ago - 1 comment
Labels: CLA Signed, fb-exported

#97 - Update README.md to include tensorboard in pip install command

Pull Request - State: closed - Opened by Lix1993 4 months ago - 4 comments
Labels: CLA Signed, Merged

#96 - Do not pipeline last inner loop.

Pull Request - State: closed - Opened by htyu 4 months ago - 1 comment
Labels: CLA Signed, fb-exported

#95 - How to deal with unknown tokens in practice?

Issue - State: closed - Opened by isofun 4 months ago - 2 comments

#94 - Evaluation on public dataset

Issue - State: closed - Opened by Blank-z0 4 months ago - 1 comment

#93 - change addmm grid size to 2D to overcome uint16 limitation

Pull Request - State: closed - Opened by jiyuanzFB 4 months ago - 2 comments
Labels: CLA Signed, fb-exported

#92 - Enable HSTU SMEM for TW and PW with autotuning for both SMEM preload and maxnregs

Pull Request - State: closed - Opened by plotfi 4 months ago - 7 comments
Labels: CLA Signed, fb-exported

#91 - Integration with TorchRec

Issue - State: closed - Opened by sbhavani 4 months ago - 1 comment

#90 - Try to use triton.language.extra.libdevice when possible

Pull Request - State: closed - Opened by danzimm 4 months ago - 5 comments
Labels: CLA Signed, fb-exported

#89 - Raise precision bar for mrn_compute_output

Pull Request - State: closed - Opened by mrshenli 5 months ago - 2 comments
Labels: CLA Signed, fb-exported