Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / baichuan-inc/Baichuan-13B issues and pull requests
#207 - 为了方便技术交流,拉了一个多模态大模型技术交流群,有需要的大家可以加入
Issue -
State: open - Opened by feihuamantian about 1 month ago
#206 - npu部署
Issue -
State: open - Opened by httang1224 4 months ago
#205 - 如何加速模型推理速度?
Issue -
State: open - Opened by hunfwj 5 months ago
#203 - 想问一下百川2量化版本的算法是什么?
Issue -
State: open - Opened by huangxiancun 6 months ago
#202 - Baichuan13B vllm 效果很差
Issue -
State: open - Opened by moseshu 9 months ago
- 1 comment
#201 - feat: function calling
Issue -
State: open - Opened by wey-gu 9 months ago
#200 - 将训练好的模型进行放入到web.demo中报错
Issue -
State: open - Opened by ghh1125 9 months ago
#199 - 各位大佬,请问采用官网给出的fine-tune文件做微调大概需要多少显存,使用A6000(48G)显示内存溢出。
Issue -
State: closed - Opened by nevesaynever1 10 months ago
#198 - 各位大佬,微调baichuan2-13b后得到pth文件,该如何推理
Issue -
State: open - Opened by dongdongqiang2018 10 months ago
#197 - v100能部署Baichuan-13B-Base么?
Issue -
State: open - Opened by JasonFlyBeauty 10 months ago
#196 - 本地部署版本问题
Issue -
State: closed - Opened by JasonFlyBeauty 10 months ago
- 2 comments
#195 - baichuan-13b-chat批量生成示例
Issue -
State: open - Opened by MrInouye 10 months ago
#194 - baichuan2 mmlu结果复现的问题
Issue -
State: open - Opened by zhanghan1992 11 months ago
#193 - 请问下,大家都是租用GPU服务器来运行大模型吗
Issue -
State: closed - Opened by jiuwenyu 11 months ago
#192 - 这个模型不支持多gpu模式吗
Issue -
State: closed - Opened by 394988736 11 months ago
#190 - ValueError: Tokenizer class BaichuanTokenizer does not exist or is not currently imported.
Issue -
State: open - Opened by lonngxiang 12 months ago
- 2 comments
#189 - 如何离线部署?
Issue -
State: open - Opened by wangnaihao almost 1 year ago
- 1 comment
#188 - baichuan-13b-chat sft微调loss不下降
Issue -
State: open - Opened by xiaohuihwh about 1 year ago
- 1 comment
#187 - 对baichuan13b还没有开始微调,仅仅是对话就自言自语?总是泄露Human: Assistant:对话数据
Issue -
State: open - Opened by hahajinghuayuan about 1 year ago
- 1 comment
#186 - 调用api.py流式代码谁能分享一下啊
Issue -
State: open - Opened by xuyaokun about 1 year ago
- 2 comments
#185 - Alibi编码为什么和标准的Alibi编码不一致?
Issue -
State: open - Opened by wx971025 about 1 year ago
#184 - 在 prompt中增加长度限制,无法生效
Issue -
State: open - Opened by Jessie37464 about 1 year ago
#183 - fasttransformer inference
Issue -
State: open - Opened by HalcyonLiang about 1 year ago
#182 - 有时候卡住一直在输出,model.chat一直在等待响应很久,有没有办法快速结束这种超长响应的办法
Issue -
State: open - Opened by janglichao about 1 year ago
- 1 comment
#181 - 如何把采样的随机种子固定?
Issue -
State: open - Opened by wangzhijian-tal about 1 year ago
#180 - 请问如何单卡运行模型推理
Issue -
State: closed - Opened by tanglu86 about 1 year ago
- 1 comment
#179 - Add OpenCompass badge in README
Pull Request -
State: open - Opened by vansin about 1 year ago
#178 - webdemo在多用户并发时,输出结果会变得很慢?
Issue -
State: open - Opened by jamesruio about 1 year ago
- 1 comment
#177 - [Evaluation] 提供 Baichuan 模型在 OpenCompass 上的评测结果
Issue -
State: open - Opened by Leymore about 1 year ago
#176 - cli_demo.py 切换Baichuan-13B-Base 问答异常
Issue -
State: open - Opened by cgq0816 about 1 year ago
- 2 comments
#175 - 使用baichuan-13b训练reward model loss不下降
Issue -
State: open - Opened by zhangzuizui about 1 year ago
#174 - Deploy Failed
Issue -
State: closed - Opened by herdonyan about 1 year ago
- 1 comment
#173 - 请问自动下载模型 下载到哪里去了?
Issue -
State: open - Opened by lonely1215225 about 1 year ago
#172 - alibi的mask和论文不一致
Issue -
State: open - Opened by ReactiveCJ about 1 year ago
- 4 comments
#171 - 一台机器有2张RTX3060显卡,跑官方web_demo的时候报了异常
Issue -
State: open - Opened by youyajike about 1 year ago
- 1 comment
#170 - 百川怎么做预训练或增量预训练????
Issue -
State: open - Opened by ArtificialZeng about 1 year ago
#169 - 推理时使用BF16和FP32结果差距较大,这是否符合预期
Issue -
State: open - Opened by NicholasYoungAI about 1 year ago
#168 - Baichuan-13B-Chat对提出过的问题有没有记忆力?
Issue -
State: open - Opened by drpanhuaming about 1 year ago
- 1 comment
#167 - 每次提问页面都出现“好的”这两个多余的字
Issue -
State: open - Opened by drpanhuaming about 1 year ago
#166 - 并发请求大模型接口的时候,模型推理速度线性增加,有没有一个比较好的增加推理速度的方案?
Issue -
State: open - Opened by chaotec about 1 year ago
- 2 comments
#165 - 有人在macbook pro 16GB上跑过百川13B吗?
Issue -
State: open - Opened by mosthandsomeman about 1 year ago
- 1 comment
#164 - 百川的开源版本模型有计划增强一下function calling的能力吗?
Issue -
State: open - Opened by huajianmao about 1 year ago
- 2 comments
#163 - 模型推理速度变慢
Issue -
State: closed - Opened by tianbuwei about 1 year ago
- 2 comments
#162 - 词表里 \U0010fc06 和 <reserved_7> 这两类有什么妙用吗?
Issue -
State: closed - Opened by CanvaChen about 1 year ago
#161 - Baichuan tokenizer对诗词分词不准确
Issue -
State: closed - Opened by CanvaChen about 1 year ago
- 3 comments
#160 - 两张A10,可以进行baichuan-13b-chat的lora微调吗
Issue -
State: open - Opened by suihuoliuying about 1 year ago
#159 - web_demo.py 运行时报错 CUDA error: CUBLAS_STATUS_NOT_INITIALIZED when calling `cublasCreate(handle)`
Issue -
State: open - Opened by matyhtf about 1 year ago
- 1 comment
#158 - 对大模型进行int8和int4量化后,如何保存模型?
Issue -
State: open - Opened by wanglaiqi about 1 year ago
#157 - 能否适配text-generation-webui的lora微调?
Issue -
State: open - Opened by BUJIDAOVS about 1 year ago
#156 - 推理速度这么慢,符合预期吗?
Issue -
State: open - Opened by ggjge about 1 year ago
#155 - baichuan batch 推理demo
Issue -
State: closed - Opened by yanzihan1 about 1 year ago
#154 - 8bit量化加载后推理速度变慢
Issue -
State: open - Opened by NLPerxue about 1 year ago
- 5 comments
#153 - Update Wechat QRCode
Pull Request -
State: closed - Opened by baichuan-assistant about 1 year ago
#152 - Baichuan 13B如何通过API调用?
Issue -
State: open - Opened by tanglu86 about 1 year ago
- 1 comment
#151 - 全参数微调loss收敛问题
Issue -
State: open - Opened by ryandok about 1 year ago
- 3 comments
#150 - Update Wechat QRCode
Pull Request -
State: closed - Opened by baichuan-assistant about 1 year ago
#149 - Update readme to match the API of the fine-tuning tools
Pull Request -
State: open - Opened by hiyouga about 1 year ago
#148 - Baichuan 13B 全参微调时命令报错
Issue -
State: closed - Opened by tanglu86 about 1 year ago
- 2 comments
#147 - 百川的模型应该用什么模型作为reward model?
Issue -
State: open - Opened by xlinsz about 1 year ago
- 1 comment
#146 - Baichuan-13B-Chat可以批量调用吗?
Issue -
State: open - Opened by FutureForMe about 1 year ago
- 5 comments
#145 - 百川13B的token上限什么时候能提高到8k或者16k呢?
Issue -
State: open - Opened by Deraemon about 1 year ago
- 2 comments
#144 - 通过api调用的情况下,stream流式输出。如果想终止当前正在输出的内容,应该怎么做?
Issue -
State: open - Opened by bxjxxyy about 1 year ago
- 4 comments
#143 - 显存一直在增长,不会释放?
Issue -
State: open - Opened by minguo about 1 year ago
- 7 comments
#140 - 关于softmax在fp32和bf16下的精度误差问题
Issue -
State: open - Opened by shiqingzhangCSU about 1 year ago
- 6 comments
#137 - The model 'BaichuanForCausalLM' is not supported for text-generation.
Issue -
State: open - Opened by gzusgw about 1 year ago
- 3 comments
#135 - 怎么13b-chat模型的角色、性格等进行定义?
Issue -
State: open - Opened by twinkle-star0 about 1 year ago
- 1 comment
#134 - AttributeError: module 'torch.backends' has no attribute 'mps'
Issue -
State: open - Opened by jamesruio about 1 year ago
- 1 comment
#133 - 求教如何设置 解码方式=贪婪解码(greedy decoding)?
Issue -
State: open - Opened by sevenzard about 1 year ago
- 1 comment
#132 - 输入的prompt是中文,然而回答的内容却是英文
Issue -
State: open - Opened by ZJ-CAI about 1 year ago
- 2 comments
#130 - 萌新求解
Issue -
State: closed - Opened by SolarKnight1 about 1 year ago
- 1 comment
#129 - 英文推理报错
Issue -
State: open - Opened by hanswang73 about 1 year ago
- 12 comments
#126 - 两张4090在做推理的时候,一张卡的利用率很低,另外一张利用率比较高
Issue -
State: closed - Opened by anyuese about 1 year ago
- 1 comment
#125 - stream = True 后为什么没有随机性了。
Issue -
State: closed - Opened by IssacLin about 1 year ago
- 7 comments
#123 - 模型批量预测的问题
Issue -
State: open - Opened by TheHonestBob about 1 year ago
- 1 comment
#121 - load in 4bit和在线量化的区别?
Issue -
State: closed - Opened by NaCloudAI about 1 year ago
- 2 comments
#119 - 怎么切换成int4之后,推理速度变慢了
Issue -
State: closed - Opened by ZJ-CAI about 1 year ago
- 5 comments
#116 - int8没问题,int4部署报错
Issue -
State: open - Opened by qingquanli-tech about 1 year ago
- 3 comments
#115 - 同样的数据集同样的训练参数,在4张A100使用lora tuning baichuan-7b和baichuan-13b, baichuan-13b的显存占用比baichuan-7b小很多,请问这是正常现象吗
Issue -
State: open - Opened by fringe-k about 1 year ago
- 2 comments
#114 - 请教一下,指令微调baichuan13,训练后期出现重复\n问题
Issue -
State: open - Opened by HuipengXu about 1 year ago
- 2 comments
#112 - v100 32g 单卡模型加载完成,对话报错RuntimeError: "addmm_impl_cpu_" not implemented for 'Half'
Issue -
State: open - Opened by huangb5760 about 1 year ago
- 5 comments
#110 - int8的加载方式为什么是16位的?
Issue -
State: open - Opened by iaston about 1 year ago
- 5 comments
#106 - baichuan-13B-chat在Ceval、CMMLU、MMLU的复现指标均低于pr结果
Issue -
State: open - Opened by zhengx18 about 1 year ago
- 2 comments
#105 - How to change model parameters(temperature, top_k, top_p, etc.) in dynamic when making an inference call?
Issue -
State: open - Opened by terrificdm about 1 year ago
- 1 comment
#97 - 微信群二维码已过期
Issue -
State: closed - Opened by JayLiu7319 about 1 year ago
- 4 comments
#96 - KeyError: 'st.session_state has no key "messages". Did you forget to initialize it?
Issue -
State: closed - Opened by tommyjex about 1 year ago
- 3 comments
#91 - 提供一个自己写的http调用的api 吧
Issue -
State: open - Opened by lilongwei5054 about 1 year ago
- 5 comments
#88 - int8/int4 模型不存在
Issue -
State: closed - Opened by QuchenFu about 1 year ago
- 7 comments
#69 - Baichuan-13B-Chat的评估
Issue -
State: closed - Opened by xianghuisun about 1 year ago
- 8 comments
#67 - cli_demo.py 切换Baichuan-13B-Chat 问答异常
Issue -
State: closed - Opened by xuxiaosheng0606 about 1 year ago
- 13 comments
#45 - P40 int8推理过于慢
Issue -
State: closed - Opened by luanshaotong about 1 year ago
- 12 comments
#42 - AttributeError: 'BaiChuanForCausalLM' object has no attribute 'chat'
Issue -
State: closed - Opened by wanglaiqi about 1 year ago
- 6 comments
#35 - 有mac部署方法吗,能用GPU么,需要多少内存
Issue -
State: closed - Opened by liaoweiguo about 1 year ago
- 3 comments
#34 - Completions API in OpenAI format
Pull Request -
State: closed - Opened by huajianmao about 1 year ago
- 4 comments
#29 - 多卡运行demoRuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cuda:1!
Issue -
State: closed - Opened by Young1993 about 1 year ago
- 3 comments
#25 - 如果想用指令数据finetune Baichuan-13B-Chat,应该使用什么样的prompt模板?
Issue -
State: closed - Opened by egrcc about 1 year ago
- 21 comments
#22 - ImportError: This modeling file requires the following packages that were not found in your environment: configuration_baichuan. Run `pip install configuration_baichuan`
Issue -
State: closed - Opened by luohao123 about 1 year ago
- 5 comments
#6 - 为何24g做int8推理,也out of memory
Issue -
State: closed - Opened by zhangyunming about 1 year ago
- 2 comments