deepseek-ai DeepSeek-V2 issues

deepseek-ai / DeepSeek-V2

DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model

MIT License

3.47k stars 143 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Inquiry about Key/Value Storage and Matrix Merging in DeepSeekerV2 Inference Code

#92 xlim1996 opened 6 days ago
0
doc: followup #89 add client demo for using SGLang

#91 Ying1123 closed 1 week ago
1
doc: followup #89 add client demo

#90 zhyncs closed 1 week ago
1
doc: recommend SGLang for DeepSeek V2 inference

#89 zhyncs closed 1 week ago
1
Function Calling比以前难触发了

#88 whoisfucker opened 3 weeks ago
5
Exploring the Combined Effects of YaRN and Adjusted rope_base Values in deepseek v2

#87 hannlp opened 4 weeks ago
0
docs: fix incorrect link in README.md

#86 itaowei opened 1 month ago
0
Question about the design of bos and eos token

#85 jojo23333 opened 1 month ago
0
线上api如何稳定的触发 tool_calls

#84 wssnail opened 1 month ago
1
ValueError: The model's max seq len (163840) is larger than the maximum number of tokens that can be stored in KV cache (13360). Try increasing gpu_memory_utilization or decreasing max_model_len when initializing the engine.

#83 ArtificialZeng opened 1 month ago
3
empty response from server

#82 879611427 opened 1 month ago
2
fix

#81 ArtificialZeng opened 1 month ago
0
HuggingFace中开源的代码似乎没有实现矩阵合并

#80 meteorlin opened 1 month ago
1
多轮在训练中是否需要特殊间隔符，用什么间隔符号？

#79 AceCHQ opened 1 month ago
0
DeepSeek-V2-Lite-Chat模型启动依赖问题

#78 Malowking opened 2 months ago
1
自配大模型服务器如何选择GPU，CPU和内存

#77 zhanghanting opened 2 months ago
0
Error executing method determine_num_available_blocks: vLLM multi node fails for both DeepSeek-Coder-V2-Instruct and DeepSeek-Coder-V2-Lite-Instruct

#76 liangfang opened 2 months ago
1
0628版本加载报错

#75 bestpredicts opened 2 months ago
0
为何我在A800上运行DeepSeek-V2-Lite-Chat (SFT)，竟然消耗60G的显存？！

#74 juhengzhe opened 2 months ago
3
about the active param counts of DeepSeek-V2-Lite

#73 imhmhm opened 2 months ago
0
Why max_model_len only 8192 when inferencing with vLLM for DeepSeek-V2-Chat?

#72 ybdesire opened 2 months ago
0
怎么用dspy里的方法来调用deepseek？

#71 buchikeke opened 2 months ago
2
What's the Prompt and Response length in the Paper?

#70 JadeRay opened 2 months ago
0
Add support llama.cpp

#69 techn0man1ac opened 2 months ago
0
希望做vs2022扩展

#68 woaidianqian closed 2 months ago
0
怎么在langchain里面使用deepseek计算embedding？

#67 ShuoAndy opened 3 months ago
3
网页端的默认参数

#66 JaheimLee closed 3 months ago
0
您好，可以查看源码吗？

#65 Darleen71 opened 3 months ago
0
同一个请求连续多次尝试都是相同错误

#64 gauss-clb opened 3 months ago
0
docs: update README for LMDeploy support

#63 zhyncs closed 2 months ago
1
Main

#62 bang78945 opened 3 months ago
0
如何优化deepseek用来做文本审查时的prompt定义

#61 xfghvgnfyjssjgte opened 3 months ago
0
如何让模型能够回答完问题自动停止

#60 hensiesp32 opened 3 months ago
0
关于DeepSeek-Coder-V2-Lite-Base的128k捞针测试结果

#59 chaochen99 opened 3 months ago
1
it swapped to chinese and i cant get it to change back to english

#58 james28909 opened 3 months ago
1
It won't answer questions about the events that transpired in Tiananmen Square from April 15, 1989, to June 4, 1989.

#57 richpav opened 3 months ago
4
128k的推理有例子吗？

#56 520jefferson opened 3 months ago
2
Will the Deepseek platform's API call be updated to support generating multiple texts (n>1)?

#55 zchuz opened 3 months ago
1
Chat API响应的role字段不要设为null

#54 jichulu opened 3 months ago
1
hi, could you provide a code like llama3?

#53 lambda7xx opened 4 months ago
2
Compatibility issues with the OpenAI Python client.

#52 dennymao opened 4 months ago
2
敏感词封禁问题

#51 gauss-clb opened 4 months ago
2
Knowledge cutoff date

#50 Shadow-Alex opened 4 months ago
0
模型部署困惑

#49 ylhou opened 4 months ago
2
Drop Token

#48 Richie-yan closed 4 months ago
2
你好，现在不支持，计划支持函数工具调用吗？

#47 cristianohello closed 4 months ago
1
has it function calling?

#46 cristianohello opened 4 months ago
1
has it function calling?

#45 cristianohello closed 4 months ago
1
docker for vllm. with deepseekv2 support merged

#44 supdizh opened 4 months ago
0
有没有计划将 deepseek-v2-lite 上传到 modelscope

#43 Tendo33 closed 4 months ago
0