Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / LlamaFamily/Llama-Chinese issues and pull requests
#365 - 使用ollama本地部署后的llama3模型生成xml格式的脚本时突然就停止输出,是超出输出限制了吗,但是有的时候又能完整输出,只是偏后面的结果不是很准确
Issue -
State: open - Opened by 75M16 5 days ago
#364 - 执行finetune.sh遇到exits with return code = -9是怎么错误呢?
Issue -
State: closed - Opened by PapaMadeleine2022 13 days ago
- 1 comment
#363 - 我用你们提供的llama3 chinese版本推理速度超慢
Issue -
State: open - Opened by fujingnan about 2 months ago
#362 - 社区在线体验推理与直接部署模型推理结果不一致
Issue -
State: open - Opened by liulan93 about 2 months ago
#361 - 执行脚本会出现 json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)
Issue -
State: closed - Opened by dshwei about 2 months ago
- 1 comment
#360 - 如何利用非问答知识进行垂直领域微调
Issue -
State: open - Opened by wdi-nancy 3 months ago
#359 - Mac 上怎么用这里写的lora来微调,新手
Issue -
State: open - Opened by Sinnsekainokami 3 months ago
- 1 comment
#358 - llama2-7B为什么在开启load_in_8bit后推理速度反而会降低至1/4呢?
Issue -
State: open - Opened by wangloveyan 3 months ago
#357 - 执行pip install -r requirements.txt报错
Issue -
State: open - Opened by RuizhePang 3 months ago
#356 - Llama2-Chinese-13b-Chat-4bit量化模型无法运行
Issue -
State: open - Opened by Kevin0412 3 months ago
#355 - 安装后运行出错
Issue -
State: open - Opened by rightWx100 4 months ago
#342 - Ollama 运行 Llama3-Chinese-8B-Instruct 回答乱生成
Issue -
State: open - Opened by hubianluanma 6 months ago
- 8 comments
#338 - 本地部署llama 3 70B,微调与推理,需要的硬件配置
Issue -
State: open - Opened by P11GH 6 months ago
- 6 comments
#330 - llama3 8B中文微调模型Llama3-Chinese-8B-Instruct 存在对生成的答案一直重复的问题
Issue -
State: open - Opened by Ryan-0805 7 months ago
- 9 comments
#320 - 执行python quick_start.py始终报错
Issue -
State: closed - Opened by moffefei 7 months ago
- 4 comments
#312 - pretrain.sh 中预训练样例数据未提供
Issue -
State: open - Opened by liuyongjie985 7 months ago
- 3 comments
#201 - 'Dataset' object has no attribute 'take'
Issue -
State: open - Opened by LiuChen19960902 about 1 year ago
- 1 comment
#101 - FasterTransformer for Llama2-70B and int8 weight-only quant are not support now.
Issue -
State: open - Opened by CN-COTER over 1 year ago
- 3 comments
#100 - 如何int4量化,并保存模型
Issue -
State: closed - Opened by xxm1668 over 1 year ago
- 2 comments
#99 - 微调支持多轮问答语料吗?
Issue -
State: closed - Opened by aaronliu7 over 1 year ago
- 2 comments
#98 - Generate方法生成的对话效果很差,有chat方法的函数吗
Issue -
State: open - Opened by dayL-W over 1 year ago
- 2 comments
#97 - 请教大佬:使用vllm时,怎样可以把history传入到里面实现上下文
Issue -
State: closed - Opened by White-Friday over 1 year ago
- 1 comment
#96 - 微调的时候这个准确率代表什么意思?
Issue -
State: open - Opened by hhh-create over 1 year ago
- 2 comments
#95 - 请问finetune_clm_lora可以实现全参数微调吗?
Issue -
State: closed - Opened by chengzhen123 over 1 year ago
- 1 comment
#94 - Llama2-Chinese-13b-Chat 使用vllm 测试在A10上出现CUDA out of memory , 官方13b-chat 可以运行
Issue -
State: closed - Opened by stevensu1977 over 1 year ago
- 2 comments
#93 - 完整的SFT数据是否准备开源?以及数据的大小和分布的全面介绍?
Issue -
State: open - Opened by juyongjiang over 1 year ago
- 1 comment
#92 - “从预训练开始对Llama2模型进行中文能力的持续迭代升级”意思是也进行了中文的Continual Pre-training?
Issue -
State: open - Opened by juyongjiang over 1 year ago
- 4 comments
#91 - 最大的输出长度只有2048吗
Issue -
State: closed - Opened by zihao-wang-owo over 1 year ago
- 1 comment
#90 - 为什么将微调脚本改成8bits会报错呢
Issue -
State: closed - Opened by hhh-create over 1 year ago
- 3 comments
#89 - 怎样实现stream流式输出?
Issue -
State: closed - Opened by xiaobaiha over 1 year ago
- 1 comment
#88 - 微调脚本可以在chinese版本上微调吗?
Issue -
State: closed - Opened by hhh-create over 1 year ago
- 1 comment
#87 - 4bit模型跑出来都是乱码
Issue -
State: open - Opened by l1006986533 over 1 year ago
- 1 comment
#86 - 部署之后运行RuntimeError
Issue -
State: open - Opened by AlanWangyiran over 1 year ago
- 3 comments
#85 - 可以出一个教程,在阿里云机器学习PAI 上面直接部署 Llama2-Chinese-7b-Chat 模型吗?
Issue -
State: closed - Opened by dividez over 1 year ago
- 1 comment
#84 - 我以为能直接搭建出```在线体验:llama.family```
Issue -
State: closed - Opened by bincooo over 1 year ago
- 1 comment
#83 - AttributeError: 'dict' object has no attribute 'load_in_8bit'
Issue -
State: closed - Opened by qishenghu over 1 year ago
#82 - 微调时候一直datasets报错是因为什么呀
Issue -
State: closed - Opened by Richard-Wth over 1 year ago
- 1 comment
#81 - 请问下大规模中文数据预训练的llama 2模型地址在哪可以找到?
Issue -
State: closed - Opened by huanghuidmml over 1 year ago
- 1 comment
#80 - TypeError: expected str, bytes or os.PathLike object, not NoneType
Issue -
State: closed - Opened by bincooo over 1 year ago
- 1 comment
#79 - 13b 词表扩了多少,支持多大上下文长度
Issue -
State: open - Opened by zhemingyang over 1 year ago
- 3 comments
#78 - 现在的微调脚本是4bit量化的吗?
Issue -
State: open - Opened by bg1avd over 1 year ago
- 1 comment
#77 - 执行docker发布命令出错!
Issue -
State: closed - Opened by Jin-Huiming over 1 year ago
- 2 comments
#76 - deepspeed是哪个版本
Issue -
State: closed - Opened by chensongcan over 1 year ago
- 1 comment
#75 - AutoGPTQForCausalLM.from_quantized 加载官方4bit量化模型报错:NameError: name 'autogptq_cuda_256' is not defined
Issue -
State: open - Opened by gpww over 1 year ago
- 1 comment
#74 - 微调所需显存大小
Issue -
State: closed - Opened by lizhipengpeng over 1 year ago
- 5 comments
#73 - 在线连接测试和本地部署Llama2-Chinese-13b-Chat效果差距很大,是哪里出问题了吗?
Issue -
State: open - Opened by IKeepMoving over 1 year ago
- 3 comments
#72 - 模型使用问题
Issue -
State: closed - Opened by wzg-zhuo over 1 year ago
- 3 comments
#71 - 评测
Issue -
State: closed - Opened by mynewstart over 1 year ago
- 1 comment
#70 - 能否将上下文长度扩展到32K大小
Issue -
State: open - Opened by KonngExplorer over 1 year ago
- 1 comment
#69 - 微调完以后怎么调用呢?
Issue -
State: closed - Opened by ymmbb8882ymmbb over 1 year ago
- 1 comment
#68 - 有更详细的信息吗
Issue -
State: closed - Opened by ymmbb8882ymmbb over 1 year ago
#67 - 预训练数据选择
Issue -
State: open - Opened by Anery over 1 year ago
- 1 comment
#66 - 微调中途报错Exception: Current loss scale already at minimum - cannot decrease scale anymore. Exiting run.
Issue -
State: open - Opened by cat-sun over 1 year ago
- 1 comment
#65 - 构建docker容器最后一步安装pip依赖时报错`No such file or directory: '/usr/local/cuda/bin/nvcc'`
Issue -
State: closed - Opened by amazingfate over 1 year ago
- 1 comment
#64 - When using GPTQ for quantification, which dataset is used?
Issue -
State: open - Opened by aresa7796 over 1 year ago
#63 - module 'bitsandbytes' has no attribute 'nn'
Issue -
State: open - Opened by Canyao45 over 1 year ago
#62 - 预训练的代码会开源吗
Issue -
State: closed - Opened by aaronliu7 over 1 year ago
- 2 comments
#61 - 请问可以提供一下huggingface上Llama2-Chinese-13b-Chat模型的sha1吗
Issue -
State: open - Opened by Mewral over 1 year ago
#60 - Llama2-Chinese 13B以及其他模型在CEVAL表现怎麽样
Issue -
State: open - Opened by cll-mtk over 1 year ago
#59 - 请问推理时如何batch样本预测
Issue -
State: open - Opened by qweasdzxc110 over 1 year ago
#58 - 如何用Langchain把英语改成汉语?
Issue -
State: open - Opened by gkgkska over 1 year ago
#57 - 使用 4bit 參數報錯
Issue -
State: open - Opened by harryliou over 1 year ago
#56 - 安装依赖报错,在爬的情况下
Issue -
State: open - Opened by JiangYeming over 1 year ago
- 2 comments
#55 - 数据使用分布情况
Issue -
State: open - Opened by echo-valor over 1 year ago
#54 - cpu运行,加速
Issue -
State: closed - Opened by wzg-zhuo over 1 year ago
#53 - 数学问题,胡说八道。
Issue -
State: closed - Opened by yunOxO over 1 year ago
- 2 comments
#52 - 情感分析
Issue -
State: closed - Opened by zhangsirya over 1 year ago
- 1 comment
#51 - 在哪里合并LoRa模型?
Issue -
State: closed - Opened by gkgkska over 1 year ago
- 1 comment
#50 - finetune_clm_lora.py中的代码是否可以直接用于T5架构的模型微调?
Issue -
State: open - Opened by HangXue-lab over 1 year ago
#49 - 模型自言自语
Issue -
State: closed - Opened by JunityZhan over 1 year ago
#48 - bitsandbytes库始终提示You might need to add them to your LD_LIBRARY_PATH.
Issue -
State: open - Opened by syusama over 1 year ago
- 5 comments
#47 - 如果我想训练llama2cn写小说,应该如何组织训练集?
Issue -
State: open - Opened by wizardforcel over 1 year ago
#46 - 感谢开发者开源,想问一下 huggingface 上两个模型 https://huggingface.co/FlagAlpha/Llama2-Chinese-13b-Chat, https://huggingface.co/FlagAlpha/Llama2-Chinese-7b-Chat finetune 使用的中文数据集是哪些呢?
Issue -
State: closed - Opened by jijeng over 1 year ago
- 1 comment
#45 - 开源的13B模型和网站上体验的demo感觉完全不一样
Issue -
State: closed - Opened by GXYM over 1 year ago
- 3 comments
#44 - Update Dockerfile
Pull Request -
State: closed - Opened by stormdragongardin over 1 year ago
#43 - Update Dockerfile
Pull Request -
State: closed - Opened by stormdragongardin over 1 year ago
#42 - cuda 11.7 RuntimeError: Error building extension 'cpu_adam'
Issue -
State: closed - Opened by ymmbb8882ymmbb over 1 year ago
- 4 comments
#41 - 请问怎么将模型与langchain-ChatGLM整合
Issue -
State: closed - Opened by string-new over 1 year ago
- 1 comment
#40 - 请问微调前是否做了中文词表扩展?
Issue -
State: closed - Opened by LilFoxProgramer over 1 year ago
- 1 comment
#39 - Llama-2-13b-chat-hf模型好像有点问题
Issue -
State: open - Opened by zhaoyin214 over 1 year ago
#38 - FlagAlpha/Llama2-Chinese-13b-Chat-4bit模型需要什么配置可以运行
Issue -
State: closed - Opened by magictext over 1 year ago
- 1 comment
#37 - 请问finetune的时候如何设置多卡训练,默认是使用单卡
Issue -
State: closed - Opened by BUAAtao over 1 year ago
- 1 comment
#36 - 请问llama.family什么时候可以更新提供在线测试呢?目前访问超时
Issue -
State: closed - Opened by Be-a-Warior over 1 year ago
- 1 comment
#35 - 调整为英文符号
Pull Request -
State: closed - Opened by csfrez over 1 year ago
#34 - 有做中文词表扩充吗
Issue -
State: closed - Opened by louieyliu over 1 year ago
- 2 comments
#33 - docs: update README.md
Pull Request -
State: closed - Opened by ycjcl868 over 1 year ago
#32 - 语法错误修改
Pull Request -
State: closed - Opened by Wenw97 over 1 year ago
#30 - 模型效果与chatglm2对比如何?
Issue -
State: closed - Opened by ZSitong over 1 year ago
- 2 comments
#29 - 请问为什么我部署的LLaMa2-chinese-13b-chat的输出结果和您这边在线提供的LLama-family上差距很大?
Issue -
State: closed - Opened by huangjianzuishuai over 1 year ago
- 2 comments
#28 - 中文LLAMA测试结果
Issue -
State: closed - Opened by Victoriaheiheihei over 1 year ago
- 1 comment
#27 - 安装 Deepspeed时,Unable to pre-compile async_io
Issue -
State: open - Opened by mmhzlrj over 1 year ago
- 2 comments
#26 - $ pip install -r requirements.txt 报错
Issue -
State: closed - Opened by mmhzlrj over 1 year ago
- 1 comment
#25 - 训练所需资源?7b-chat,lora
Issue -
State: closed - Opened by elegant-bot over 1 year ago
- 2 comments
#24 - 注释deepspeeds,添加dockerfile,docker-compose启动文件,修改启动脚本可以在局域网内访问
Pull Request -
State: closed - Opened by stormdragongardin over 1 year ago
#23 - 怎样quantize?
Issue -
State: closed - Opened by luckfu over 1 year ago
- 1 comment
#22 - 新增docker文件夹,增加dockerfile文件,增加Docker compose启动文件Docker compose
Pull Request -
State: closed - Opened by stormdragongardin over 1 year ago
#21 - OSError: No such device (os error 19)
Issue -
State: closed - Opened by yajunDai over 1 year ago
- 2 comments
#20 - 能否提供sha256
Issue -
State: closed - Opened by AnxietyYoungPoet over 1 year ago
- 1 comment
#19 - ERROR: Invalid requirement: 'torch torchvision torchaudio' (from line 13 of requirements.txt)
Issue -
State: closed - Opened by yangjianchun over 1 year ago
- 2 comments