1 Pull request merged by 1 person

• docs: Add system requirements for DeepSeek-Infer demo

  #341 merged Jan 26, 2025

36 Pull requests opened by 32 people

• Update README.md

  #359 opened Jan 27, 2025

• Update README.md

  #360 opened Jan 27, 2025

• Add table of contents to README for better navigation

  #364 opened Jan 27, 2025

• Update model.py

  #367 opened Jan 27, 2025

• Create xxx.py

  #371 opened Jan 27, 2025

• Create xxx.py

  #372 opened Jan 27, 2025

• Fixes #374: Suppress Torch Error in MoE Module by Configuring `torch._dynamo`

  #375 opened Jan 27, 2025

• Fix typos and ensure consistency in documentation

  #379 opened Jan 27, 2025

• Update README.md

  #383 opened Jan 27, 2025

• Masking: avoid modifying tensor in-place to improve performance

  #386 opened Jan 28, 2025

• [README_WEIGHTS.md]. Update link and fix grammar

  #388 opened Jan 28, 2025

• chore: Add tqdm & python to requirements.txt. Format and documents.

  #390 opened Jan 28, 2025

• Refactored convert.py Code

  #391 opened Jan 28, 2025

• On macOS we can use Ollama and Kerlig

  #399 opened Jan 28, 2025

• Clarify assertion errors

  #408 opened Jan 28, 2025

• Update requirements.txt

  #410 opened Jan 28, 2025

• Update README.md

  #419 opened Jan 28, 2025

• add link to readme

  #422 opened Jan 28, 2025

• Update generate.py: Add parallel processing for token generation

  #426 opened Jan 28, 2025

• exceptions_generate_models.py

  #428 opened Jan 28, 2025

• docs: remove redundant asterisks in note

  #432 opened Jan 28, 2025

• OCD level minor fix for consistent capitalization of term MTP

  #434 opened Jan 28, 2025

• Add Troubleshooting Section to README

  #437 opened Jan 28, 2025

• Add syntax highlighting to requirements code block

  #440 opened Jan 28, 2025

• Cleanup README

  #441 opened Jan 28, 2025

• Refactored/codebase By defining different classes for different operations and much more

  #444 opened Jan 29, 2025

• Added various error handlers and Issue templates.

  #447 opened Jan 29, 2025

• Added redirect links to github repositories of Deepseek-R1 and Deepseek-V2

  #448 opened Jan 29, 2025

• Added optional GPU Memory Logging

  #459 opened Jan 29, 2025

• feat:feat: Added logging, parallel processing, and CPU processing option for FP8 to BF16 conversion

  #461 opened Jan 29, 2025

• Fix the Readme.md issue #456

  #467 opened Jan 29, 2025

• feat: add apple silicon support

  #469 opened Jan 30, 2025

• Improve Weight File Documentation for Clarity and Readability

  #481 opened Jan 30, 2025

• Update generate.py

  #488 opened Jan 30, 2025

• sugiero un cambio en el archivo readme

  #496 opened Jan 31, 2025

• Optimization to Model Script

  #499 opened Jan 31, 2025

22 Issues closed by 15 people

• Bug Report: Stored XSS Vulnerability in DeepSeek Chat

  #470 closed Jan 30, 2025

• [BUG] it ain't responding to some questions such as "list Indian states"

  #475 closed Jan 30, 2025

• deepseek.com is not working

  #473 closed Jan 30, 2025

• [BUG] Bug Report: Critical Account Takeover Vulnerability

  #462 closed Jan 29, 2025

• Appreciation for DeepSeek AI

  #421 closed Jan 29, 2025

• [BUG] Asked him about Tiananmen massacre

  #397 closed Jan 29, 2025

• [BUG] V3 Function calling 还不能用吗?

  #349 closed Jan 29, 2025

• Documented Analysis: Bias and Behavior of DeepSeek AI on Sensitive Topics

  #406 closed Jan 28, 2025

• Can't lOgin

  #398 closed Jan 28, 2025

• [BUG] Can't login

  #369 closed Jan 27, 2025

• [BUG] 账单统计错误！

  #317 closed Jan 27, 2025

• [BUG]调用 DeepSeek API 时返回与 OpenAI 相关的内容

  #348 closed Jan 26, 2025

• [BUG] Fail to import is_torch_greater_or_equal_than_1_13 since transformers v4.48.0 for all deepseek models

  #290 closed Jan 26, 2025

• 1

  #345 closed Jan 26, 2025

• [BUG] Image content is missing.

  #344 closed Jan 26, 2025

• Can this model be run on a single GPU? NVIDIA A10G with 24GB VRAM

  #332 closed Jan 26, 2025

• [BUG] request failed with Image failed in base64 through OpenAI Python SDK

  #337 closed Jan 26, 2025

• OpenAI ChatGPT discussion integration

  #338 closed Jan 26, 2025

• [BUG] 🔴 Critical Security Bug in Payment System

  #334 closed Jan 25, 2025

• 所以deepseek-v3就是套壳CHATGPT?

  #328 closed Jan 24, 2025

• confused answer in chat

  #20 closed Jan 24, 2025

• [BUG] convert.py cannot convert DeepSeek-R1-Distill-Qwen-1.5B

  #330 closed Jan 24, 2025

91 Issues opened by 90 people

• Copyleft

  #500 opened Jan 31, 2025

• Reference Fine-Tuning Code

  #498 opened Jan 31, 2025

• [BUG] Error occured in Torch files

  #497 opened Jan 31, 2025

• How to Use DeepSeek AI: A Comprehensive Guide for Beginners

  #495 opened Jan 31, 2025

• [BUG]使用Roo Code调用api失败

  #494 opened Jan 31, 2025

• Redirecting Feature

  #493 opened Jan 31, 2025

• MoE only load activated expert(s) to GPU while rest non-used experts are not loaded (to CPU/GPU) for DeekSeek-R1 or V3 Inference on consumer GPU

  #492 opened Jan 30, 2025

• Fork and Remake Chinese AI to Remove CCP Ties

  #491 opened Jan 30, 2025

• [FEATURE REQUEST] Images Insertion & documents insertion

  #490 opened Jan 30, 2025

• training hyper-parameters for ablation studies

  #489 opened Jan 30, 2025

• [BUG]

  #487 opened Jan 30, 2025

• Cloudflare Turnstile BUG

  #485 opened Jan 30, 2025

• 什么时候接入Python？

  #484 opened Jan 30, 2025

• [BUG] Not deleting all chats

  #483 opened Jan 30, 2025

• [BUG] Exceeding the text length causes the output to be in Chinese (Simplified)

  #482 opened Jan 30, 2025

• 能把介绍翻译成中文吗

  #480 opened Jan 30, 2025

• Question about NVLink bandwidth mentioned in DeepSeek_V3.pdf

  #479 opened Jan 30, 2025

• pip3 install -r requirements.txt ERROR: Could not find a version that satisfies the requirement torch==2.4.1 (from versions: 2.6.0) ERROR: No matching distribution found for torch==2.4.1

  #478 opened Jan 30, 2025

• [BUG] Why DeepSeek Isn't reponding to some questions like "list Indian States"

  #477 opened Jan 30, 2025

• [BUG]

  #476 opened Jan 30, 2025

• [BUG]

  #474 opened Jan 30, 2025

• [BUG] Server error in deepseek

  #472 opened Jan 30, 2025

• 新手如何学习ai 有python基础

  #471 opened Jan 30, 2025

• [NOOB QUESTION] How should one go about digesting and learning this codebase?

  #468 opened Jan 29, 2025

• Holy C implementation/Cuda offline/Hyrid

  #466 opened Jan 29, 2025

• [web app] support arbitrary email domains

  #464 opened Jan 29, 2025

• Triton Installation, Compatibility & Documentation Improvements

  #463 opened Jan 29, 2025

• 关于应对恶意攻击并加强DeepSeek安全防护的建议（边缘安全加速、WAF、DDoS防护与CDN优化）

  #460 opened Jan 29, 2025

• [NOOB] The code is only this?

  #457 opened Jan 29, 2025

• [BUG] Spelling mistakes / grammatical errors in Readme file

  #456 opened Jan 29, 2025

• [BUG] convert.py does not work for the DeepSeek-R1-Distill-Qwen-7B model

  #455 opened Jan 29, 2025

• The server is busy. Please try again later.

  #454 opened Jan 29, 2025

• Can we make him more human more O/pentagram?

  #453 opened Jan 29, 2025

• Recipe to run DeepSeek online inference on a SLURM Cluster

  #452 opened Jan 29, 2025

• The response to an API anomaly.

  #450 opened Jan 29, 2025

• [BUG]:Often showing server is busy

  #449 opened Jan 29, 2025

• Deep Seek Casually responds in russian from an english prompt:

  #446 opened Jan 29, 2025

• [BUG]

  #445 opened Jan 29, 2025

• Terraform resource

  #443 opened Jan 29, 2025

• [BUG] Russian text response

  #442 opened Jan 29, 2025

• RL code

  #439 opened Jan 28, 2025

• [BUG]Can't login with google

  #438 opened Jan 28, 2025

• [BUG] Can retrieve the information beyond scope

  #436 opened Jan 28, 2025

• [BUG] easy to bypass the censoring

  #435 opened Jan 28, 2025

• [BUG]

  #433 opened Jan 28, 2025

• [web app] add support for passmail.net email domain

  #427 opened Jan 28, 2025

• Rename the Chat Title with some relevant name.

  #425 opened Jan 28, 2025

• Response Disappearing After Generation

  #423 opened Jan 28, 2025

• Multi-Lingual Support in DeepSeek-V3?

  #420 opened Jan 28, 2025

• Consistency, can Deepseek pass?一致性，deepseek能及格吗？

  #418 opened Jan 28, 2025

• optimize the generate function

  #417 opened Jan 28, 2025

• code optimization

  #416 opened Jan 28, 2025

• [BUG] Unable to Upload Image

  #415 opened Jan 28, 2025

• Documented Analysis: Bias and Behavior of DeepSeek AI on Sensitive Topics

  #414 opened Jan 28, 2025

• [BUG] Nav bar title not updating

  #413 opened Jan 28, 2025

• Amazing AI

  #412 opened Jan 28, 2025

• [BUG] Biased question being halfed awnsered

  #411 opened Jan 28, 2025

• [BUG] Unable to login with Japan Yahoo Email

  #409 opened Jan 28, 2025

• [BUG] Deepseek keep returning error Unexpected end of JSON input

  #407 opened Jan 28, 2025

• Deepseek API result getting latency

  #405 opened Jan 28, 2025

• Question: How to start on Linux - after run nothing happen...

  #404 opened Jan 28, 2025

• Lite Version V3 weights

  #403 opened Jan 28, 2025

• [BUG] Urgent: RTL Text Alignment Issue in Persian-English Mixed Content

  #402 opened Jan 28, 2025

• [Deploy issue] multi-node(2) tensor parallelism in docker on SGLang engine would fail. What's the minimum hardware requirements?

  #401 opened Jan 28, 2025

• Fine-tuning requirements

  #400 opened Jan 28, 2025

• [BUG] Tiananmen Square 1989

  #396 opened Jan 28, 2025

• 能否为缓存命中机制增加开关

  #393 opened Jan 28, 2025

• Amazing work!

  #389 opened Jan 28, 2025

• Confusion over underscore `_` used in special tokens

  #385 opened Jan 28, 2025

• [BUG] deepseek claiming that it's better than chatgpt at maths but it's in fact worse

  #384 opened Jan 27, 2025

• LICENCE-MODEL formatting not ideal

  #380 opened Jan 27, 2025

• Desktop Windows/Linux app

  #378 opened Jan 27, 2025

• torch.fx.experimental.symbolic_shapes.GuardOnDataDependentSymNode

  #374 opened Jan 27, 2025

• [BUG] LaTeX输出可能有误

  #373 opened Jan 27, 2025

• [BUG]关于ai使用的数据混乱

  #370 opened Jan 27, 2025

• [BUG] Can't signup

  #368 opened Jan 27, 2025

• Unable to register, policy risk control?

  #365 opened Jan 27, 2025

• Cannot create docment and makes it downloadable

  #363 opened Jan 27, 2025

• 本地模型都要加审查。

  #362 opened Jan 27, 2025

• [BUG] Can't acess

  #361 opened Jan 27, 2025

• Allow connecting multiple login methods

  #358 opened Jan 27, 2025

• intel arc a770独显可以本地部署吗

  #357 opened Jan 27, 2025

• MTP support in demo inference code

  #352 opened Jan 26, 2025

• Feature Request: Profile Customization and Integration with Google/Microsoft Accounts

  #351 opened Jan 26, 2025

• [BUG] Login or signup on Arc Browser loading forever

  #350 opened Jan 26, 2025

• Add other payment providers other than Paypal that support all countries.

  #347 opened Jan 26, 2025

• [BUG] Photo link problem

  #346 opened Jan 26, 2025

• Request: Ammending the end licence to include planet/environment focused restrictions

  #343 opened Jan 26, 2025

• How to pass Image-Based Math/Geometry Problems to Model

  #339 opened Jan 25, 2025

• [BUG] Image recognition content is missing

  #336 opened Jan 25, 2025

• generator_model = AutoModelForCausalLM.from_pretrained('deepseek-ai/DeepSeek-R1', trust_remote_code=True) throws error in RAG model/产生错误

  #335 opened Jan 24, 2025

20 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Question about FP8 Tensor Core Mantissa Precision

  #197 commented on Jan 24, 2025 • 0 new comments

• 我想知道一下 技术报告里面激活37B是怎么算出来的？

  #231 commented on Jan 25, 2025 • 0 new comments

• [BUG] No option to login in via google

  #322 commented on Jan 27, 2025 • 0 new comments

• train过程模型代码是没有上传吗？

  #323 commented on Jan 27, 2025 • 0 new comments

• [BUG]使用TensorRT-llm 的Deepseek分支 部署4bit weight only的deepseekV3回答乱码

  #272 commented on Jan 28, 2025 • 0 new comments

• [BUG]多轮对话中，如果用户输入内容有多次重复，模型回复会出现之前回复内容一样的情况

  #243 commented on Jan 28, 2025 • 0 new comments

• v3 repetitive function call ?

  #15 commented on Jan 28, 2025 • 0 new comments

• how to finetune this model?

  #10 commented on Jan 28, 2025 • 0 new comments

• Hi, how many A100's will fine tune the DeepSeek-V3?

  #109 commented on Jan 28, 2025 • 0 new comments

• [BUG] DeepSeek V3 Does Not Support Structured Output in LangChain with `ChatOpenAI()`

  #302 commented on Jan 28, 2025 • 0 new comments

• Fix Critical Bug in Right-to-Left Language Support and Add Persian, Arabic, and Hebrew Languages

  #326 commented on Jan 28, 2025 • 0 new comments

• Where can I find the source code?

  #171 commented on Jan 29, 2025 • 0 new comments

• [BUG] RTL Arabic Direction

  #294 commented on Jan 29, 2025 • 0 new comments

• [BUG]no code on mail at signup

  #288 commented on Jan 29, 2025 • 0 new comments

• 网页版深度思考默认语言需要默认中文

  #221 commented on Jan 30, 2025 • 0 new comments

• Integrating Anthropic's MCP

  #309 commented on Jan 30, 2025 • 0 new comments

• 有没有int4量化版，int4量化版推理需要多少什么显卡配置

  #244 commented on Jan 30, 2025 • 0 new comments

• Bug: Broken Google OAuth login- RECAPTCHA_VERIFY_FAILED on platform.deepseek.com

  #170 commented on Jan 30, 2025 • 0 new comments

• Function Calling失效

  #7 commented on Jan 30, 2025 • 0 new comments

• Cline插件调用卡在Api request

  #265 commented on Jan 31, 2025 • 0 new comments