Activity
Stars and mentions · Last 30 days, Oct 5 – Nov 3
Top repositories
FareedKhan-dev’s repositories ranked by new forks, as on trending.
Contributes to
Repositories outside FareedKhan-dev’s account where FareedKhan-dev opened or reviewed pull requests in the past 365 days.
Who noticed FareedKhan-dev
Everyone who starred or forked FareedKhan-dev’s repositories since the archive began in 2011.
Notable accounts
Ordered by the stars their own repositories have received.
Biggest fans
Accounts that starred the most of these repositories.
Breakout weeks
Each repository’s best week of new stars, with the release published around it.
Achievements
Awards earned by the repositories FareedKhan-dev owns. Select an award to embed its badge.
No awards yet
Repositories earn awards by reaching the top three in a completed trending day or week.
A repository earns an award for finishing first, second or third in a completed trending day or week, ranked by new stars, new forks or pull request authors. Awards follow a repository through renames and transfers and count toward its current owner. Weeks run Monday to Sunday, UTC.
Mentions
Será que é verdade? FareedKhan-dev/kimi-k3-in-c
kimi-k3-in-c runs Moonshot's full 2.8T Kimi K3 MoE on one CPU by streaming experts from a 1.56 TB disk checkpoint. Peak RSS measured: 8.24 GB. Laptop preset: about 26-33 s/token. The RAM claim is real. Usable chat speed is not. FareedKhan-dev/kimi-k3-in-c
Kimi-k3-in-C: A 2.78T-parameter model. No GPU. 8 GB of RAM
You Can Now Run A 2.78-Trillion Parameter Kimi K3 Model With Just 8GB of RAM! | Kimi-k3-in-c is an open-source engine that runs the full Kimi K3 model on a single CPU. What you get: - The full 2.78T Kimi K3, not a cut-down version - Runs in 8.24GB of RAM with no GPU - The whole engine is a 176KB C file, with no PyTorch or CUDA - The same output as a 128GB machine, only slower - Chat mode, and it builds on Linux, macOS and Windows --- ####How to get it: 1. Clone FareedKhan-dev/kimi-k3-in-c 2. Run make, then ./scripts/k3-doctor.sh to check your machine 3. Download the checkpoint and pick a preset that fits your RAM It's slow and needs a big SSD, but a trillion parameter model on 8GB of RAM is wild. --- ####Pic Explaining What Tokens Per Second You Can Expect: freeimage.host
这个项目 FareedKhan-dev/kimi-k3-in-c 可以吧Kimi k3 完整的2.8T参数跑在8G内存的设备上🥲 8G内存运行 Kimi k3，这词有点小众了🥲 但是速度大概是 0.03 tok/s 左右，差不多23s-30s 出一个Token🤡 能接受这速度的兄弟们可以试试🤓
Top 3 GitHub Repos For AI Developers Today #Shorts Top 3 GitHub Repos For AI Developers Today Repos in this video: herdrdev/herdr vercel-labs/portless FareedKhan-dev/train-llm-from-scratch New GitHub finds every day. #Shorts #github #opensource #devtools #coding #ai
想真正看懂 LLM 训练链路，最好亲手把代码跑一遍。 Train LLM From Scratch 用纯 PyTorch 从零实现 Transformer，从 Token、注意力、多头一路做到预训练、SFT、Reward、DPO/PPO/GRPO，不依赖 Transformers、TRL、PEFT。每个模块都有说明和预期输出，适合边跑边理解。 🔗 FareedKhan-dev/train-llm-from-scratch
2.78 万亿超大参数的模型，居然能在一台 8GB 内存的普通单核 CPU 上跑起来。 关注模型底层优化的工程师建议存下这个案例。 仓库地址：FareedKhan-dev/kimi-k3-in-c