Ecosyste.ms: Issues

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

GitHub / NVIDIA/FasterTransformer issues and pull requests

#793 - An error occurred for the specific cuda version

Issue - State: open - Opened by CSLiuPeng 7 months ago
Labels: bug

#791 - fix: fix position_encoding_table memory error.

Pull Request - State: open - Opened by johnson-magic 8 months ago

#790 - bug: memory of position_encoding_table is not malloced correctly.

Issue - State: open - Opened by johnson-magic 8 months ago
Labels: bug

#787 - what is the mean of EFF-FT?

Issue - State: open - Opened by johnson-magic 9 months ago

#785 - M2M

Pull Request - State: closed - Opened by sfc-gh-ybsat 11 months ago - 1 comment

#783 - Confidence is not returned in the decoding example？

Issue - State: open - Opened by liuzhuang1024 11 months ago
Labels: bug

#782 - multi_block_mode performance issue

Issue - State: closed - Opened by akhoroshev 12 months ago - 1 comment

#782 - multi_block_mode performance issue

Issue - State: closed - Opened by akhoroshev 12 months ago - 1 comment

#780 - Free memory buffer - Llama

Pull Request - State: closed - Opened by sfc-gh-ybsat about 1 year ago

#780 - Free memory buffer - Llama

Pull Request - State: closed - Opened by sfc-gh-ybsat about 1 year ago

#779 - error You need C++17 to compile PyTorch

Issue - State: open - Opened by ranggihwang about 1 year ago
Labels: bug

#779 - error You need C++17 to compile PyTorch

Issue - State: open - Opened by ranggihwang about 1 year ago
Labels: bug

#778 - can support decoder only bart? such as MBartForCausalLM

Issue - State: open - Opened by sjtu-cz about 1 year ago

#778 - can support decoder only bart? such as MBartForCausalLM

Issue - State: open - Opened by sjtu-cz about 1 year ago

#777 - repetition_penalty logic in FT has bug

Issue - State: closed - Opened by hezeli123 about 1 year ago - 1 comment
Labels: bug

#777 - repetition_penalty logic in FT has bug

Issue - State: closed - Opened by hezeli123 about 1 year ago - 1 comment
Labels: bug

#776 - Update README.md

Pull Request - State: open - Opened by eltociear about 1 year ago

#776 - Update README.md

Pull Request - State: open - Opened by eltociear about 1 year ago

#775 - Sparsity support

Issue - State: open - Opened by zhang662817 about 1 year ago
Labels: bug

#775 - Sparsity support

Issue - State: open - Opened by zhang662817 about 1 year ago
Labels: bug

#774 - How to get started?

Issue - State: open - Opened by turbobuilt about 1 year ago

#774 - How to get started?

Issue - State: open - Opened by turbobuilt about 1 year ago

#772 - How to serving multi-gpu inference？

Issue - State: closed - Opened by Alone-wl about 1 year ago - 1 comment

#772 - How to serving multi-gpu inference？

Issue - State: closed - Opened by Alone-wl about 1 year ago - 1 comment

#771 - Is llama2 70b supported? Do you know minimal configuration?

Issue - State: open - Opened by ChristineSeven about 1 year ago - 1 comment

#771 - Is llama2 70b supported? Do you know minimal configuration?

Issue - State: open - Opened by ChristineSeven about 1 year ago - 1 comment

#770 - Include stdio.h

Pull Request - State: open - Opened by JihaoXin about 1 year ago - 2 comments

#770 - Include stdio.h

Pull Request - State: open - Opened by JihaoXin about 1 year ago - 2 comments

#769 - Supporting for expert parallelism in MoE inference

Issue - State: open - Opened by iteratorlee about 1 year ago

#769 - Supporting for expert parallelism in MoE inference

Issue - State: open - Opened by iteratorlee about 1 year ago

#767 - cuSPARSELt is slower?

Issue - State: open - Opened by BDHU about 1 year ago - 1 comment

#766 - Incorrect inline ptx device assembly code usage

Issue - State: open - Opened by zhiweij1 about 1 year ago
Labels: bug

#766 - Incorrect inline ptx device assembly code usage

Issue - State: open - Opened by zhiweij1 about 1 year ago
Labels: bug

#764 - How to calculate local batch size?

Issue - State: open - Opened by fotstrt about 1 year ago

#764 - How to calculate local batch size?

Issue - State: open - Opened by fotstrt about 1 year ago

#762 - Ft llama opt

Pull Request - State: open - Opened by dypshong about 1 year ago

#762 - Ft llama opt

Pull Request - State: open - Opened by dypshong about 1 year ago

#760 - fastertransformer/utils/nccl_utils.cc:62 'unhandled cuda error'

Issue - State: open - Opened by wangweiwei1188 about 1 year ago
Labels: bug

#760 - fastertransformer/utils/nccl_utils.cc:62 'unhandled cuda error'

Issue - State: open - Opened by wangweiwei1188 about 1 year ago
Labels: bug

#759 - Support for "no_repeat_ngram_size" parameter for generation

Issue - State: open - Opened by shreysingla11 about 1 year ago - 2 comments

#759 - Support for "no_repeat_ngram_size" parameter for generation

Issue - State: open - Opened by shreysingla11 about 1 year ago - 2 comments

#758 - Does FasterTransformer use FlashAttention?

Issue - State: open - Opened by niyunsheng about 1 year ago

#758 - Does FasterTransformer use FlashAttention?

Issue - State: open - Opened by niyunsheng about 1 year ago

#756 - Support Seq length up to 8K

Pull Request - State: open - Opened by zhen-jia about 1 year ago

#756 - Support Seq length up to 8K

Pull Request - State: open - Opened by zhen-jia about 1 year ago

#755 - [cmake] fix cmake policy for ENABLE_FP8

Pull Request - State: closed - Opened by DefTruth about 1 year ago

#755 - [cmake] fix cmake policy for ENABLE_FP8

Pull Request - State: closed - Opened by DefTruth about 1 year ago

#754 - flashattention only enabled for gpt-styled models

Issue - State: open - Opened by flexwang about 1 year ago - 7 comments

#754 - flashattention only enabled for gpt-styled models

Issue - State: open - Opened by flexwang about 1 year ago - 7 comments

#753 - How to get a npz file that satisfy the input requirement?

Issue - State: open - Opened by jy00161yang about 1 year ago - 1 comment
Labels: bug

#753 - How to get a npz file that satisfy the input requirement?

Issue - State: open - Opened by jy00161yang about 1 year ago - 1 comment
Labels: bug

#752 - [Long seq length] GPT Seq length constrain

Issue - State: open - Opened by zhen-jia over 1 year ago - 14 comments

#752 - [Long seq length] GPT Seq length constrain

Issue - State: open - Opened by zhen-jia over 1 year ago - 14 comments

#751 - specify the recognition language for Whisper

Issue - State: open - Opened by echodjx over 1 year ago

#751 - specify the recognition language for Whisper

Issue - State: open - Opened by echodjx over 1 year ago

#748 - [feature request] transformer on orin

Issue - State: open - Opened by superpigforever over 1 year ago

#748 - [feature request] transformer on orin

Issue - State: open - Opened by superpigforever over 1 year ago

#747 - How to run multi_gpu_gpt_examples.py without mpirun/mpiexe

Issue - State: closed - Opened by ZZWHU over 1 year ago

#746 - core dumped of swin model

Issue - State: open - Opened by chiemon over 1 year ago - 1 comment
Labels: bug

#746 - core dumped of swin model

Issue - State: open - Opened by chiemon over 1 year ago - 1 comment
Labels: bug

#744 - Failed building t5 model in FastTransformer (Reached 82% then stopped)

Issue - State: open - Opened by EmanElrefai12 over 1 year ago - 3 comments
Labels: bug

#744 - Failed building t5 model in FastTransformer (Reached 82% then stopped)

Issue - State: open - Opened by EmanElrefai12 over 1 year ago - 3 comments
Labels: bug

#736 - Using faster transformers to infer the bloom model, the accuracy rate is 0

Issue - State: open - Opened by hurun over 1 year ago - 2 comments
Labels: bug

#736 - Using faster transformers to infer the bloom model, the accuracy rate is 0

Issue - State: open - Opened by hurun over 1 year ago - 2 comments
Labels: bug

#734 - TP=2， Loss of accuracy

Issue - State: open - Opened by coderchem over 1 year ago - 2 comments

#734 - TP=2， Loss of accuracy

Issue - State: open - Opened by coderchem over 1 year ago - 2 comments

#730 - Compatibility issue with CUDA 12.2

Issue - State: open - Opened by MinghaoYan over 1 year ago - 6 comments
Labels: bug

#729 - llama support inference？

Issue - State: open - Opened by double-vin over 1 year ago - 2 comments

#729 - llama support inference？

Issue - State: open - Opened by double-vin over 1 year ago - 2 comments

#728 - [Question] Is it possible to use my own pretrained weights for ViT QAT

Issue - State: open - Opened by proevgenii over 1 year ago - 3 comments
Labels: bug

#728 - [Question] Is it possible to use my own pretrained weights for ViT QAT

Issue - State: open - Opened by proevgenii over 1 year ago - 3 comments
Labels: bug

#727 - Are MQA and GQA in development?

Issue - State: open - Opened by ljayx over 1 year ago - 8 comments

#727 - Are MQA and GQA in development?

Issue - State: open - Opened by ljayx over 1 year ago - 8 comments

#720 - docker/Dockerfile.torch occurs errors

Issue - State: closed - Opened by b3y0nd over 1 year ago - 4 comments
Labels: bug

#720 - docker/Dockerfile.torch occurs errors

Issue - State: closed - Opened by b3y0nd over 1 year ago - 4 comments
Labels: bug