Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / alibaba/rtp-llm issues and pull requests
#120 - 2张16G的T4卡都跑不起来examples/test.py
Issue -
State: open - Opened by zhangtaibo 9 days ago
#119 - [cpu] add sampleGreedy implementation
Pull Request -
State: closed - Opened by wenhuanh 10 days ago
#118 - fix: open source build and deps on Arm
Pull Request -
State: open - Opened by TianyuLi0 11 days ago
#117 - perf: optimization of attention, softmax, layernorm
Pull Request -
State: closed - Opened by Reyfone 12 days ago
#116 - Add grouped query attention support
Pull Request -
State: closed - Opened by Reyfone 17 days ago
#115 - [Doc] 多卡并行文档修改建议
Issue -
State: open - Opened by flliny 23 days ago
#114 - RTP-LLM 模式下,llama3.1 FP16 效果不一样
Issue -
State: open - Opened by anigi98932 23 days ago
#113 - support to run example/test.py and integrate optimized gemm/attention operator
Pull Request -
State: closed - Opened by TianyuLi0 24 days ago
- 1 comment
#112 - support to run example/test.py on Arm
Pull Request -
State: closed - Opened by TianyuLi0 about 1 month ago
#111 - 双卡A6000推理,模型推理结束,一张卡GPU利用率为0,一张卡GPU利用率100%
Issue -
State: open - Opened by zf761 about 1 month ago
- 1 comment
#110 - 无法运行tests目录下的Python测试脚本,缺少libtest_ops.so
Issue -
State: open - Opened by leepoly about 1 month ago
- 1 comment
#109 - 双卡A6000推理,模型推理结束,一张卡GPU利用率为0,一张卡GPU利用率100%
Issue -
State: open - Opened by zf761 about 1 month ago
- 1 comment
#108 - fix: unit test and cpp model test
Pull Request -
State: closed - Opened by Reyfone about 1 month ago
#107 - Enable MHA parallel on Arm
Pull Request -
State: closed - Opened by Reyfone about 1 month ago
#106 - attention: add MHA parallel support
Pull Request -
State: closed - Opened by Reyfone about 1 month ago
- 1 comment
#105 - speculate sampling用medusa加载medusa官方模型报错
Issue -
State: open - Opened by wcsjtu about 1 month ago
- 6 comments
#104 - reranker token长度拦截异常
Issue -
State: closed - Opened by invisifire about 2 months ago
- 2 comments
#103 - add opt_125M
Pull Request -
State: open - Opened by Nanuion about 2 months ago
- 2 comments
#102 - 新增OPT模型,模型输出不符合预期
Issue -
State: closed - Opened by samaritan1998 about 2 months ago
#101 - [CPU] add implementation for GEMM and token embedding
Pull Request -
State: closed - Opened by wenhuanh about 2 months ago
#100 - 推理出现乱码(показать показать показать показать показать)(USE_NEW_DEVICE_IMPL=1)
Issue -
State: open - Opened by w066650 about 2 months ago
- 1 comment
#99 - [ROCm] refine quantization related code
Pull Request -
State: closed - Opened by feifei14119 2 months ago
- 2 comments
#98 - [ROCm] MoE version1
Pull Request -
State: closed - Opened by feifei14119 2 months ago
- 1 comment
#97 - [ROCm] Support Int4 and bf16 for rocm version
Pull Request -
State: closed - Opened by feifei14119 2 months ago
#96 - [ROCm] add quant op and port rccl
Pull Request -
State: closed - Opened by feifei14119 2 months ago
#95 - 新增OPT模型后跑不通,报CUDA错误
Issue -
State: closed - Opened by samaritan1998 2 months ago
- 5 comments
#94 - [ROCm] Includes docker container creation script file for rocm build
Pull Request -
State: closed - Opened by feifei14119 2 months ago
- 1 comment
#93 - [ROCm] Fix ROCm sampler OP test
Pull Request -
State: closed - Opened by feifei14119 2 months ago
#92 - [cpu-impl] Add for layernorm and rmsnorm
Pull Request -
State: closed - Opened by wenhuanh 3 months ago
#91 - HELP: No matching distribution found for torch==2.1.0+cu121 error while install maga_transformer with .whl in release 0.2.0
Issue -
State: open - Opened by HuXinjing 3 months ago
- 7 comments
#90 - fix: adapt to index based kv cache for Arm device
Pull Request -
State: closed - Opened by Reyfone 3 months ago
- 1 comment
#89 - `Illegal instruction` error when running version 0.2.0
Issue -
State: closed - Opened by frankang 3 months ago
- 2 comments
#88 - bazel build error
Issue -
State: closed - Opened by frankang 3 months ago
- 2 comments
#87 - [ROCm] Port basic gpt model to rocm. qwen2 end-to-end test pass
Pull Request -
State: closed - Opened by feifei14119 3 months ago
- 5 comments
#86 - 您好,I'd like to ask a question that might not be very professional. In the code, the weights are loaded through Python. Where are they passed to the C++(fasttransformer) part?
Issue -
State: closed - Opened by samaritan1998 3 months ago
- 1 comment
#85 - [DRAFT] not ready, please do NOT review
Pull Request -
State: closed - Opened by feifei14119 3 months ago
- 1 comment
#84 - support DeepSeek-V2-Lite-Chat
Issue -
State: open - Opened by jianglan89 3 months ago
- 1 comment
#83 - feat: add arm cpu device support
Pull Request -
State: closed - Opened by TianyuLi0 3 months ago
- 1 comment
#82 - 多机单卡/多卡,报错 gang_info self None
Issue -
State: closed - Opened by MasterJanus 3 months ago
#81 - [ROCm] Init rocm_impl device and add test op
Pull Request -
State: closed - Opened by feifei14119 3 months ago
- 4 comments
#80 - feat: add cpu attention api
Pull Request -
State: closed - Opened by wenhuanh 3 months ago
#79 - [ROCm] Initial enablement
Pull Request -
State: closed - Opened by draganmladjenovic 3 months ago
- 6 comments
#78 - git clone Error
Issue -
State: closed - Opened by hz0ne 3 months ago
- 3 comments
#77 - fix(src): fix bazel build special type cast and template match for cuda118
Pull Request -
State: closed - Opened by khan-yin 3 months ago
- 14 comments
#76 - 单机多卡如何制定卡号
Issue -
State: closed - Opened by 256785 3 months ago
- 1 comment
#75 - Glm4v运行问题
Issue -
State: closed - Opened by samaritan1998 3 months ago
- 3 comments
#74 - v0.2.0(cuda12)对比 v0.1.13(cuda11)表现下降
Issue -
State: open - Opened by invisifire 3 months ago
- 1 comment
#73 - glm4v 单卡Cuda out of memory
Issue -
State: closed - Opened by samaritan1998 3 months ago
- 1 comment
#72 - 请问 0.2.0 支持cuda 11环境么?
Issue -
State: closed - Opened by samaritan1998 3 months ago
#71 - qwen2 gptq tp=4 报错:AssertionError: error config
Issue -
State: open - Opened by xinge666 3 months ago
- 3 comments
#70 - 编译报错 v0.2.0 版error: cannot convert ‘<brace-enclosed initializer list>’ to ‘const fastertransformer::AllGatherParams&’
Issue -
State: closed - Opened by invisifire 3 months ago
- 8 comments
#69 - ChatGLM4-9B运行不起来
Issue -
State: closed - Opened by samaritan1998 3 months ago
- 1 comment
#68 - v0.1.13 load qwen2 gptq失败
Issue -
State: closed - Opened by xinge666 4 months ago
- 2 comments
#67 - 请问multimodel_mixin.py中BaseMultiModalWeightInfo的_get_vit_info函数中self.vit_weights?
Issue -
State: closed - Opened by samaritan1998 4 months ago
- 8 comments
#66 - openai.InternalServerError: Error code: 500 - {'error_code': 514, 'message': 'ErrorMsg: failed to malloc 134 blocks, only 28 blocks left
Issue -
State: closed - Opened by wanglaiqi 4 months ago
- 1 comment
#65 - Does it support Qwen2、ChatGLM4-9B?
Issue -
State: closed - Opened by ZCDu 4 months ago
- 4 comments
#64 - 多卡部署空闲但导致的其他模型速度降低很多
Issue -
State: closed - Opened by invisifire 4 months ago
- 2 comments
#63 - Qwen Chat CUDA OutOfMemory
Issue -
State: open - Opened by xorange 4 months ago
- 2 comments
#62 - [Feature Request] llama3
Issue -
State: closed - Opened by samaritan1998 4 months ago
- 1 comment
#61 - [Feature Request] Add support for CogVLM2
Issue -
State: closed - Opened by samaritan1998 5 months ago
- 5 comments
#60 - Qwen-vl-chat的结果和transformer的结果不一样,有点奇怪地像续写出来的
Issue -
State: closed - Opened by chiquitita-101 5 months ago
- 3 comments
#59 - 请问支持流式吗?
Issue -
State: closed - Opened by lcvcl 5 months ago
- 1 comment
#58 - + Add ffn layer cpu impl
Pull Request -
State: closed - Opened by wenhuanh 5 months ago
- 1 comment
#57 - Buffer overflow at CudaAttentionOpTest::selfAttentionOpTest
Issue -
State: closed - Opened by skmkt 5 months ago
- 1 comment
#56 - rtp-llm example test issue
Issue -
State: closed - Opened by haic0 5 months ago
- 1 comment
#55 - Remove print statements
Issue -
State: closed - Opened by mrchi 5 months ago
- 1 comment
#54 - update multi-gpu.md
Pull Request -
State: closed - Opened by gujingit 6 months ago
#53 - Feature request: encoder-decoder model support
Issue -
State: closed - Opened by samaritan1998 6 months ago
- 1 comment
#52 - Build error in cuda:tensor_utils: ‘getTypeSize’ is not a member of ‘fastertransformer::Tensor
Issue -
State: closed - Opened by skmkt 6 months ago
- 1 comment
#51 - 多卡推理
Issue -
State: closed - Opened by Vincent131499 6 months ago
- 8 comments
#50 - failed to run : RuntimeError: torch.cat()
Issue -
State: closed - Opened by davideuler 6 months ago
- 1 comment
#49 - 请问src/fastertransformer/models/multi_gpu_gpt/ParallelAttentionWrapper.cc的ContextAttention是什么概念呢?和SelfAttention有什么区别呢
Issue -
State: closed - Opened by samaritan1998 6 months ago
- 2 comments
#48 - qwen1.5-14b-chat部署awq
Issue -
State: closed - Opened by Vincent131499 6 months ago
- 3 comments
#47 - awq
Issue -
State: closed - Opened by Vincent131499 6 months ago
- 2 comments
#46 - ValueError: max() arg is an empty sequence
Issue -
State: closed - Opened by boxiaowave 6 months ago
- 4 comments
#45 - 请问kmonitor metrics怎么开启打印呢?想测试一下每个阶段的耗时
Issue -
State: closed - Opened by samaritan1998 6 months ago
- 1 comment
#44 - bazel构建成功,但是测试报错
Issue -
State: closed - Opened by samaritan1998 6 months ago
- 4 comments
#43 - 2 GPUs with TP=2 run Lora inference, one GPU
Issue -
State: closed - Opened by cwlseu 6 months ago
- 1 comment
#42 - bazel构建失败
Issue -
State: closed - Opened by samaritan1998 6 months ago
- 10 comments
#41 - Error in DeployDocker.md
Issue -
State: closed - Opened by vegetable-yx 6 months ago
- 1 comment
#40 - Poor performance at batchsize=1 on V100
Issue -
State: closed - Opened by cwlseu 6 months ago
- 12 comments
#39 - 按照官方示例 https://github.com/alibaba/rtp-llm/blob/main/docs/Multimodal-Tutorial.md 报错 maga_transformer.config.exceptions.FtRuntimeException: raw request format cannot accept dict prompt
Issue -
State: closed - Opened by samaritan1998 6 months ago
- 1 comment
#38 - bazel cu11x 编译失败
Issue -
State: closed - Opened by cwlseu 6 months ago
- 1 comment
#37 - 怎么使用qwen medusa推理加速
Issue -
State: closed - Opened by BucherLi 6 months ago
- 2 comments
#36 - 0.1.8 release cuda12.1 whl包不完整
Issue -
State: closed - Opened by is 6 months ago
- 3 comments
#35 - 最新whl包无法启动server
Issue -
State: closed - Opened by frankang 7 months ago
- 5 comments
#34 - follow readme then error
Issue -
State: closed - Opened by okwinds 7 months ago
- 2 comments
#33 - bazel编译失败
Issue -
State: closed - Opened by yuhui-xie 7 months ago
- 1 comment
#32 - Problem:多模态的部分是如何处理的?
Issue -
State: closed - Opened by t90tank 7 months ago
- 1 comment
#31 - Is there a plan to support Eagle?
Issue -
State: closed - Opened by cdliang11 7 months ago
- 1 comment
#30 - docs: fix link
Pull Request -
State: closed - Opened by cdliang11 7 months ago
- 1 comment
#29 - BUG: MISSING QUOTATION MARKS AND LINE BREAKS
Issue -
State: closed - Opened by invisifire 7 months ago
- 2 comments
#28 - build error:ERROR: An error occurred during the fetch of repository 'pip_gpu_cuda12_torch
Issue -
State: closed - Opened by FightingMan 7 months ago
- 2 comments
#27 - random_seed未生效
Issue -
State: closed - Opened by frankang 7 months ago
- 1 comment
#26 - [bug ?] mega_transformer/models/llava.py中encode_images方法
Issue -
State: closed - Opened by samaritan1998 7 months ago
- 1 comment
#25 - #RTP-LLM Developer Event# 春季限定活动,捉bug送美味咖啡☕️
Issue -
State: open - Opened by tt0718 7 months ago
- 2 comments
#24 - KeyError: 'MODEL_TYPE'
Issue -
State: closed - Opened by York-RDWang 7 months ago
- 1 comment
#23 - support Yi-Vl
Issue -
State: open - Opened by Lzhang-hub 7 months ago
- 5 comments
#22 - NameError: name 'Middleware'is not defined, Did you mean: 'CoRsMiddleware'?
Issue -
State: closed - Opened by zzhdbw 7 months ago
- 1 comment
#21 - doc: update cuda12 dep file path
Pull Request -
State: closed - Opened by gujingit 7 months ago
- 1 comment