Skip to content
View shivendrra's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report shivendrra

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
shivendrra/README.md

About me

Hello, I'm Shivendra. I like to code and make videos. Check out some of my previous work at Vakya

Main Projects:

1- SmallLanguageModel [completed, closed]: Making a LLM from scratch all the way from generating raw training data to tokenizing it, creating a model & then training it.

2- Enigma-1.5b [completed, closed]: Transformer model trained on raw DNA data to predict the next letter of the DNA.

3- Axon [completed, updating]: Numpy from scratch in python without any external library. Also with a tiny scalar level aut0grad axon.micro

4- Axgrad [in progress]: Pytorch from scratch, mostly python based but soon will support c/c++ based backend.

5- Axon.drop [in progress]: Scalar level autograd engine written in c/c++ projected to use as tensor & accessed via python

6- WebGraze [completed, updating]: A Python-based library for webscraping & generating/downloading data from various sources on the internet for training ml models.

7- Synapse [completed, to be updated]: A free platform for streaming music & audio/podcasts, based on Youtube V3 API.

8- Micrograd.C [completed, no updates]: Micrograd by Karpathy written in C & C++.

9- AIVA-4x500M [in progress, backlogged]: A multimodal ai system inspired by AVA from Ex-Machina, but currently more like OpenAi's 4o, (I started working on it prior to the model launch).

10- Shredword [in progress, private]: A tokenizer library just like TikToken by OpenAi, using C based code at the backend for fast execution speeds with Python wrapper.

11- Enigma2 [in progress, backlogged]: Second version of Enigma with a different approach this time with the model & tokenization process.

Socials

Instagram | Twitter | LinkedIn | Youtube

📊 GitHub Stats:

Kevin's github stats

shivendrra

Pinned Loading

  1. SmallLanguageModel SmallLanguageModel Public

    a LLM cookbook, for building your own from scratch, all the way from gathering data to training a model

    Jupyter Notebook 128 16

  2. enigma-1.5b enigma-1.5b Public

    a dna sequence generation/classification using transformers

    Jupyter Notebook 1

  3. axon axon Public

    numpy written in python from scratch

    Python

  4. synapse synapse Public

    a music app that lets you stream music or any video available on youtube for free without ads

    JavaScript 1

  5. micrograd.c micrograd.c Public

    kaprathy's micrograd implemented in c & cpp

    C

  6. web-graze web-graze Public

    scrape raw data from various sources of the internet, like wikipedia, internet archieve, britannica, youtube, unsplash, etc

    Python 3