ByteMLPerf commits | Ecosyste.ms: OpenCollective

[Micro] fix output json path typo error (#55)

github.com/bytedance/ByteMLPerf - 73544e71d466cd9239d93c5ddc63f25c33d1becf authored 10 months ago by HanTengfei99 <[email protected]>

[Micro Perf]change build_tensor logic and unique dtype (#54)

* add layer norm op

* change unique dtype and build tensor

* change unique dtype and build...

github.com/bytedance/ByteMLPerf - e274b33d52a921d3ebbe95ef032b492c57ee6858 authored 10 months ago by HanTengfei99 <[email protected]>

[ByteMicroPerf]add layer norm op (#53)

github.com/bytedance/ByteMLPerf - 510c7570f1915bc0b495b1b74267f80ba9d246cd authored 10 months ago by HanTengfei99 <[email protected]>

(byte_micro_perf):add a variety of input shapes (#52)

github.com/bytedance/ByteMLPerf - 9ba1861e0565cfbc60c45e644790cab5ca4e08a9 authored 10 months ago by HanTengfei99 <[email protected]>

[Micro Perf]Apply python black&isort lint (#51)

github.com/bytedance/ByteMLPerf - 491e0e81597da8bfb75102ff340bed018de5951c authored 10 months ago by HanTengfei99 <[email protected]>

[Micro Benchmark]add microbenchmark to evaluate some DL ops on GPU (#50)

github.com/bytedance/ByteMLPerf - d51c4e2d3cbee1f9645bd88fc59b15bc49bfca51 authored 10 months ago by HanTengfei99 <[email protected]>

[General Perf] Update vae-decoder-onnx-fp32.json

github.com/bytedance/ByteMLPerf - 031a1ba21439548db52ec088eee6c6a6ad2cec67 authored 10 months ago by JianzheXiao <[email protected]>

[Core] Update Readme in ByteMLPerf (#46)

* update ReadMe

* Update README.md

* Update README.zh_CN.md

* Upd...

github.com/bytedance/ByteMLPerf - 21506b9c0fa54024b8f426e7dae9b8274a83f897 authored 10 months ago by JianzheXiao <[email protected]>

[LLM Perf] Bug Fix and new features supported

Feature:
1. 抽象卡间通信类，实现GPU从NCCL修改为IPC，删除非RANK0卡周期sleep逻辑，进行堵塞获取请求，减少延迟
2. 修改packet退出判断条件，支持fini...

github.com/bytedance/ByteMLPerf - 52b9b05b67d8ce08dd8abc7da04fb0fb5843fed6 authored 10 months ago by N <[email protected]>

[LLM Perf] Fix several bugs: (#45)

1. rename llama-2 to llama2
2. fix llama2 and chatglm2 new_token_ids with generate_ids
3. fix ...

github.com/bytedance/ByteMLPerf - 7c0f1209a14b8f6bb91ff5a89a22f5cf6413a435 authored 10 months ago by N <[email protected]>

[Core] Update Vendor Zoo

github.com/bytedance/ByteMLPerf - 0ee93017022509f29cdd2da20e4008f7ef8501aa authored 11 months ago by jianzhe.xiao <[email protected]>

update vendor zoo

github.com/bytedance/ByteMLPerf - e8e7ff7b0d422759779161ba7cb5ebab30c092d5 authored 11 months ago by jianzhe.xiao <[email protected]>

[General Perf] Update deberta-torch-fp32

github.com/bytedance/ByteMLPerf - f5c6dbaa92f4299027a39c466879479288a249de authored 12 months ago by JianzheXiao <[email protected]>

[LLM-Task] dataset: remove useless datasets (#42)

Signed-off-by: MikuGhoul <[email protected]>

github.com/bytedance/ByteMLPerf - 216c72f90e0ef6902ce252713855d81e9df654a3 authored 12 months ago by N <[email protected]>

[Core] Rename byte_mlperf to general task (#40)

* update general perf

* update

* add ByteMLPerf Inference

* up...

github.com/bytedance/ByteMLPerf - 7ed25996453278682f7835b154b32014d9f3a9da authored 12 months ago by JianzheXiao <[email protected]>

[Core]Update Vendor Zoo Path

github.com/bytedance/ByteMLPerf - 942c43d2ca9d7f900ea5cb39ff6410d47eb6a99b authored about 1 year ago by jianzhe.xiao <[email protected]>

[Core]Update Vendor Zoo Path

github.com/bytedance/ByteMLPerf - 564489dcc80b60493ee596e4238ef5c5fbd44318 authored about 1 year ago by jianzhe.xiao <[email protected]>

[Core]Update ByteMLPerf Doc

github.com/bytedance/ByteMLPerf - 61af5d3098578c185b9a2c79ca98c27df6719818 authored about 1 year ago by jianzhe.xiao <[email protected]>

[Moffett]Update Hardware Images

github.com/bytedance/ByteMLPerf - d595fedf32690da9be9364e9a820413cec94eeef authored about 1 year ago by AngryPowman <[email protected]>

[Core] Move Vendor Zoo

github.com/bytedance/ByteMLPerf - 37c0921c96d547645e7132f8ecdcb01eb0bc7fb6 authored about 1 year ago by jianzhe.xiao <[email protected]>

[Core] Add Vendor Zoo

github.com/bytedance/ByteMLPerf - 3a5147414deb33732e58c0f1b9e3d6f76e3d8fca authored about 1 year ago by jianzhe.xiao <[email protected]>

add Vendor Zoo

github.com/bytedance/ByteMLPerf - 7ea8eee7adcfa0305cbf700a147069423deff624 authored about 1 year ago by jianzhe.xiao <[email protected]>

[Core]Update README.md

github.com/bytedance/ByteMLPerf - 473695f6e89d0ecb52ae72caea1adf719a3ac83e authored about 1 year ago by JianzheXiao <[email protected]>

[Core] Update README.md

github.com/bytedance/ByteMLPerf - 982ac21369ca42d8e1c4b9fdcaf228af641f3220 authored about 1 year ago by JianzheXiao <[email protected]>

[HPU] Support HPU Backend in ByteMLPerf

* support gpt2

* update

* add hpu runtime

* enable hpu bf16

* fix bug: add import to...

github.com/bytedance/ByteMLPerf - ad0a9e4a4307af3721ee088929c0a33fde38e070 authored about 1 year ago by huijuanz <[email protected]>

[Core] Update gpt2-torch-fp32.json

github.com/bytedance/ByteMLPerf - 3d537442c11cff72d5864f67d28adf1c3ea23cda authored about 1 year ago by JianzheXiao <[email protected]>

[Core] Update prepare_model_and_dataset.sh

github.com/bytedance/ByteMLPerf - ef77ba584a4a210a046c579b807d28e4dfe4e7ed authored about 1 year ago by JianzheXiao <[email protected]>

[Graphcore] PopRT 1.4 support & corresponded performance update (#32)

* [update] 1. GraphCore IPU supports poprt 1.4; reports updated base on it. 2. bugs fixes

* [...

github.com/bytedance/ByteMLPerf - 43d48ddc449de4631cb089eb6c91114d57bc39ab authored over 1 year ago by keg0704 <[email protected]>

[General] fixed minor bugs

github.com/bytedance/ByteMLPerf - 693d29fba34ccc5aa7091900f0b684efc873e801 authored over 1 year ago by keg0704 <[email protected]>

[Core] Update ReadMe for ByteIR

github.com/bytedance/ByteMLPerf - c38a060b94455f1e112b1ed8d999ea09a6a09569 authored over 1 year ago by jianzhe.xiao <[email protected]>

update

github.com/bytedance/ByteMLPerf - 1154988c60c00e29f2628e325745975f9afb2a31 authored over 1 year ago by jianzhe.xiao <[email protected]>

update

github.com/bytedance/ByteMLPerf - a7bdb33a672cb893e5401d51ee772e485bfd7919 authored over 1 year ago by jianzhe.xiao <[email protected]>

Update README.md

github.com/bytedance/ByteMLPerf - 7e62edbcd34c0529880bbe4121e313fc2fba434c authored over 1 year ago by JianzheXiao <[email protected]>

[Core] Add support for ByteIR Mhlo Model

github.com/bytedance/ByteMLPerf - ea8337abd6205cdd801d503bff6154c8b16218d5 authored over 1 year ago by root <[email protected]>

[SPU] Fix reports Performance typo

github.com/bytedance/ByteMLPerf - ad37d8ebbe8066555071d1f84ae6d141321ce42e authored over 1 year ago by jianzhe.xiao <[email protected]>

[SPU] Update ByteMLPerf with vendor Moffett-AI (S30) and backend SPU. (#15)

* feat: add spu backend

* doc: update readme

* Update README.md

*...

github.com/bytedance/ByteMLPerf - 717b0ecb4fb71f1550be2bb25dce1f2f1ce803c9 authored over 1 year ago by Moffett AI <[email protected]>

[GraphCore] Ipu add fp8 models (#30)

* [new] 1. add ipu fp8 support and results for bert-torch-fp32/swin-large-torch-fp32/resnet50-to...

github.com/bytedance/ByteMLPerf - b9e78e82788830de01ce89359219bfed332a7842 authored over 1 year ago by keg0704 <[email protected]>

update unet model store

github.com/bytedance/ByteMLPerf - 10c260141d3e128e2207b3269e42934b793bc566 authored over 1 year ago by jianzhe.xiao <[email protected]>

[CORE]Support muti precision compilation

github.com/bytedance/ByteMLPerf - 87050302029ff2ff1bdb796b55b9d131d4eb5559 authored over 1 year ago by jianzhe.xiao <[email protected]>

[GraphCore] IPU support pack deberta

* [new] 1. ipu support pack deberta

* [update] 1. update reports_summary.png

* [update] 1....

github.com/bytedance/ByteMLPerf - f958248177ee805394ac9c923e73ccf53ce2de19 authored over 1 year ago by keg0704 <[email protected]>

Merge branch 'main' of https://github.com/bytedance/ByteMLPerf into main

github.com/bytedance/ByteMLPerf - 0095d838be233feeffc2bde71f83bc157dee9a4f authored over 1 year ago by JianzheXiao <[email protected]>

[Core] Add host info in the result report

github.com/bytedance/ByteMLPerf - ceebf322d67b77a2aa3147f4701c515f756a3e3a authored over 1 year ago by JianzheXiao <[email protected]>

[Core] Update README.zh_CN.md

github.com/bytedance/ByteMLPerf - ed1b66e3078e4151515f41bf9a59eda9397f103e authored over 1 year ago by JianzheXiao <[email protected]>

[Core] Update README.md

github.com/bytedance/ByteMLPerf - 8a43d68a4cb84c1468bbd5b2026ae6e9dc1e4839 authored over 1 year ago by JianzheXiao <[email protected]>

[Core]Add LlaMa2 7B to Model Store

github.com/bytedance/ByteMLPerf - 4bf190c294841ce42f98df46b56a2f58bd7c9310 authored over 1 year ago by JianzheXiao <[email protected]>

[Core]Update Result Config / Add ChatGLM2-6b to dataset

github.com/bytedance/ByteMLPerf - 6b6f84d11ff5c07cd8643dce853bc5cc012a4984 authored over 1 year ago by JianzheXiao <[email protected]>

Merge pull request #27 from jackzipu/update_supported_model_list

Update the supported model list

github.com/bytedance/ByteMLPerf - 431e40831dda8b35813152bc41b95e8307841ec7 authored over 1 year ago by Minghui-BD <[email protected]>

Update the supported model list to refresh the data and add the new models

github.com/bytedance/ByteMLPerf - 9787b0ff79e41e7cef5ff27060e891ed12da90f6 authored over 1 year ago by Jack Zhang <[email protected]>

Update the reports_summary.png (#26)

github.com/bytedance/ByteMLPerf - 8c52327ebbcb57c4c95e49fc2e58269e2e5f7a8e authored over 1 year ago by Jack Zhang <[email protected]>

[GraphCore] Ipu native pack support (#24)

* [update] 1. re-pushed ipu native-pack support to avoid too many rebase commits

* [update] 1...

github.com/bytedance/ByteMLPerf - 8098e48944b32ffa8854c379fb7eae2d231ff6e3 authored over 1 year ago by keg0704 <[email protected]>

[GraphCore] IPU adds clip-onnx-fp32 model support

IPU adds clip-onnx-fp32 model support

github.com/bytedance/ByteMLPerf - a3c2e3cc4c428a8aa27e255ec5e7546fd9ab15dc authored over 1 year ago by JianzheXiao <[email protected]>

IPU adds clip-onnx-fp32 model support

github.com/bytedance/ByteMLPerf - 83c2543367fb0375a089878f86f88499663a30fb authored over 1 year ago by Jack Zhang <[email protected]>

Merge pull request #23 from jackzipu/update_readme

[GraphCore] Update readme files

github.com/bytedance/ByteMLPerf - 7d2cc19c7bef0d09d8390d882fdd3ca4fd99d57b authored over 1 year ago by JianzheXiao <[email protected]>

Merge pull request #21 from jackzipu/ipu_support_swin_large

[new] 1. ipu added support for swin large

github.com/bytedance/ByteMLPerf - 244eb6d11f22f6708350795129ea001ecfc45376 authored over 1 year ago by JianzheXiao <[email protected]>

Merge branch 'main' into ipu_support_swin_large

github.com/bytedance/ByteMLPerf - 5a9a49d8f52ade9d9f35d7cec1ede908fd14ffb4 authored over 1 year ago by Ke Gong <[email protected]>

[fix] 1. use model_zoo's input names for torch to onnx conversion

github.com/bytedance/ByteMLPerf - d4b1fc5821e777c61e357748143ec3a564c5a4f7 authored over 1 year ago by Ke Gong <[email protected]>

[update] 1. update bert/albert support for torch 1.13.1

github.com/bytedance/ByteMLPerf - db2028688b239f6aa3e50d37f4c23a32d0f1baa2 authored over 1 year ago by Ke Gong <[email protected]>

Fix a typo cifra100->cifar100 for videobert dataset

github.com/bytedance/ByteMLPerf - 99bc18dccffd071653effcb514bd72db03c52d43 authored over 1 year ago by Jack Zhang <[email protected]>

Update readme files

github.com/bytedance/ByteMLPerf - 6d82344c27179b56a1362b38da66dc32e7395353 authored over 1 year ago by Jack Zhang <[email protected]>

[update] 1. fixes based on comments

github.com/bytedance/ByteMLPerf - 8ff72ee108dcc644217c2a313bdbd7292cc286ea authored over 1 year ago by Ke Gong <[email protected]>

upgrade torch version to 1.13.1

github.com/bytedance/ByteMLPerf - ca48dc336a358070d60b42c54a3b6cbb12a44b42 authored over 1 year ago by JianzheXiao <[email protected]>

Merge pull request #22 from jackzipu/release_engine

Release the engine before create a new one

github.com/bytedance/ByteMLPerf - 5c566c67b0c5f706d2533339d916765bae8c2873 authored over 1 year ago by JianzheXiao <[email protected]>

Release the engine before create a new one

github.com/bytedance/ByteMLPerf - 702877c889ffeaf8d16b39e52b9e43593550d471 authored over 1 year ago by Jack Zhang <[email protected]>

[fix] 1. add pandas back to cpu's requiements.txt

github.com/bytedance/ByteMLPerf - 1d90f2eb234776e6b8b6ccc5931fb7b4095d6def authored over 1 year ago by Ke Gong <[email protected]>

[new] 1. ipu added support for swin large

github.com/bytedance/ByteMLPerf - 13cd1725d6a185193ca896a74cb83cc261a0eddb authored over 1 year ago by Ke Gong <[email protected]>

Merge pull request #19 from jackzipu/video_bert_update

Add smallest batch size of workload to make accuracy check happy

github.com/bytedance/ByteMLPerf - 238281cb5dc07dd5ae16c5e2e10dc2c77b4e7f37 authored over 1 year ago by Minghui-BD <[email protected]>

Add smallest batch size of workload to make accuracy check happy

github.com/bytedance/ByteMLPerf - 69fbd4485beeac28aa0bcf53db21293015bfb5b9 authored over 1 year ago by Jack Zhang <[email protected]>

Merge pull request #18 from jackzipu/add_batch_sizes

Add pandas to CPU requirement.txt which needed by wide_deep

github.com/bytedance/ByteMLPerf - 6a3037523dddf98fdf40f2ce87e5612e199dc5fd authored over 1 year ago by Minghui-BD <[email protected]>

Merge pull request #16 from jackzipu/video_bert_model

Add video bert model

github.com/bytedance/ByteMLPerf - 52cbe80c1d78ef207b5fb8b650243a6acb3a9943 authored over 1 year ago by Minghui-BD <[email protected]>

Add pandas to CPU requirement.txt which needed by wide_deep

github.com/bytedance/ByteMLPerf - e5098379f3af2c8b1b6a2503de1276db30a653e5 authored over 1 year ago by Jack Zhang <[email protected]>

Add performance result

github.com/bytedance/ByteMLPerf - ed02907763002da89cfbe4fa0e68384283e550e1 authored over 1 year ago by Jack Zhang <[email protected]>

Update the timeout_ns to adapt the new PopRT API

github.com/bytedance/ByteMLPerf - 2c5a29cff310e4773f0b86d1017a3f8d2afc28ce authored over 1 year ago by Jack Zhang <[email protected]>

Add videobert model

github.com/bytedance/ByteMLPerf - d42690d751308361a7bdb18709b83a65086ce177 authored over 1 year ago by Jack Zhang <[email protected]>

Update postprocess for swin large

github.com/bytedance/ByteMLPerf - 06853be80c318c6807ae81290224e4d9fd6aa826 authored over 1 year ago by JianzheXiao <[email protected]>

Remove pdb in IPU Backend

github.com/bytedance/ByteMLPerf - cd216fcb6e412b8415e21fc126b78478680ace08 authored over 1 year ago by JianzheXiao <[email protected]>

Merge pull request #11 from jackzipu/update_ipu_info

Update the IPU info in the README.md

github.com/bytedance/ByteMLPerf - 9de173bb37c99b9178242ad35405ff4b95d0ab9f authored over 1 year ago by Minghui-BD <[email protected]>

Update the compiled batch size and compiled executable path

github.com/bytedance/ByteMLPerf - 78a07c3642312aacae153538f8195d222bfa3f86 authored over 1 year ago by Jack Zhang <[email protected]>

Update torch_to_onnx

github.com/bytedance/ByteMLPerf - cd5bd1da8def7e0838943c70cb94f9fbc845055d authored over 1 year ago by JianzheXiao <[email protected]>

Add Deberta Model

github.com/bytedance/ByteMLPerf - d69cb90ac5e15954ecb94c3faede9f28cea60d76 authored over 1 year ago by jianzhe.xiao <[email protected]>

Add Python Verions Check

github.com/bytedance/ByteMLPerf - 1122015f1fbe7f14ab6ac5a92e53d0b6c5a663bd authored over 1 year ago by jianzhe.xiao <[email protected]>

Update the IPU info in the README.md

github.com/bytedance/ByteMLPerf - 9e8a5facdd1324d0760308188e5c034dc77edf65 authored over 1 year ago by Jack Zhang <[email protected]>

update pip with --quite

github.com/bytedance/ByteMLPerf - 3b691370224e69d0f76bf93bda466d605eb881eb authored over 1 year ago by yuminghui.exp <[email protected]>

use python3

pip instead of pip

github.com/bytedance/ByteMLPerf - c40ecae238a5593b810b1e9d61c4103179ba0a59 authored over 1 year ago by yuminghui.exp <[email protected]>

correct usage: --tasks -> --task

github.com/bytedance/ByteMLPerf - e8ad6351f96bca289cff264fe69858740becccb2 authored over 1 year ago by Minghui-BD <[email protected]>

add benchmark summary

github.com/bytedance/ByteMLPerf - 5c3ce8b99b9f71cc8033dc425f05e576bbb3f7e1 authored over 1 year ago by yuminghui.exp <[email protected]>

Merge pull request #9 from jackzipu/commit_branch

add IPU backend

github.com/bytedance/ByteMLPerf - b2040e2ccb7de934a0415839b425c027a477380c authored over 1 year ago by Minghui-BD <[email protected]>

Add pack and runtime_optiions to interact info

github.com/bytedance/ByteMLPerf - 846d41a043f4a077c86de1993e03988997a06b57 authored over 1 year ago by Jack Zhang <[email protected]>

Update README information

github.com/bytedance/ByteMLPerf - 18dfa924dc78cf156726ac6316e6efb4cd2875ea authored over 1 year ago by Jack Zhang <[email protected]>

Remove the download script

github.com/bytedance/ByteMLPerf - 8f00ec40008871fe46d421118b9dbd663f3f5186 authored over 1 year ago by Jack Zhang <[email protected]>

1. Fix the review comments about

a. Product introduction
b. Make the precision with the precision set for conversion and
com...

github.com/bytedance/ByteMLPerf - 3b8b5339047d9550c8745188f1eb9c10adfb3fd0 authored over 1 year ago by Jack Zhang <[email protected]>

Pack solution for bert and albert models

github.com/bytedance/ByteMLPerf - bb949a13c3aeef6128af69e2ab1baf1dfde6ed77 authored over 1 year ago by Jack Zhang <[email protected]>

restore unnecessary changes

github.com/bytedance/ByteMLPerf - 81c9526e1440ecdd37aefa32ea9f08133a6d4801 authored over 1 year ago by Jack Zhang <[email protected]>

Initial version of IPU backend

github.com/bytedance/ByteMLPerf - 8e3039acac4d2755b450d3a2b972a28353b48bf4 authored over 1 year ago by Jack Zhang <[email protected]>

Add models supported list in the README.md and performance reports

github.com/bytedance/ByteMLPerf - e4173580afb3f30475fb79c2c5b33bf1e9328892 authored over 1 year ago by Jack Zhang <[email protected]>

Update README.zh_CN.md

correct typo

github.com/bytedance/ByteMLPerf - a9b035b15104bee646df29a3b6f5afd3a1e56a32 authored over 1 year ago by Minghui-BD <[email protected]>

Update README.md

correct typo

github.com/bytedance/ByteMLPerf - d8163e596ad00de599ca6c1fe0eed01680090aab authored over 1 year ago by Minghui-BD <[email protected]>

Update README.zh_CN.md

github.com/bytedance/ByteMLPerf - 75f86a93813d58721749db174988f0161f239956 authored over 1 year ago by Minghui-BD <[email protected]>

Update README.md

github.com/bytedance/ByteMLPerf - 3fe78c5574b5517c770b080118f7b7953974a1bd authored over 1 year ago by Minghui-BD <[email protected]>

Merge pull request #10 from stc-qiupeng/main

Add STC backend code and model reports

github.com/bytedance/ByteMLPerf - 8559491e1ff02b8785a1239055c4a28ea6ca0cf6 authored over 1 year ago by Minghui-BD <[email protected]>

Merge pull request #5 from stc-qiupeng/stc

update README

github.com/bytedance/ByteMLPerf - e99af4f466cfbbf496557c29f59b1330b6da9ee2 authored over 1 year ago by stc-qiupeng <[email protected]>