Aritra Roy Gosthipaty @arig23498.bsky.social · 03/03/20255/ Each of the tech mentioned above has its own pros and cons. The processor that you are running in your system (a phone, a laptop, etc) will have a weighted sum of all the above. It baffles me to think about all of this. 🤗 000
Aritra Roy Gosthipaty @arig23498.bsky.social · 03/03/20254/N Multi Threading in Single Core In a core, we can have multiple register blocks (context blocks) to access different instructions. This way if a process is stalled, the processor quickly jumps to another. 100
Aritra Roy Gosthipaty @arig23498.bsky.social · 03/03/20253/N SIMD paradigm In a single core if we have duplicate ALUs we can operate of a bunch of data in a single clock tick. The catch? Each operation should be the same. Single instruction Multiple Data 100
Aritra Roy Gosthipaty @arig23498.bsky.social · 03/03/20252/N Multi Core Processor: A single processor consists of a control unit, arithmetic unit and some registers. How about we duplicate this block into multiple blocks? This is the multi-core architecture. As a programmer you would need to explicitly mention which code runs where. 100
Aritra Roy Gosthipaty @arig23498.bsky.social · 03/03/20251/N Superscalar processors: Your program is a list of instructions. This list almost always has independent instructions. A superscalar processor would identify them and execute seperately in the same clock tick. 100
Aritra Roy Gosthipaty @arig23498.bsky.social · 03/03/2025Some pointers on parallel computing: A small thread 🧵👇 100
Reposted by Aritra Roy GosthipatyMichael Tschannen @mtschannen.bsky.social · 22/02/2025HF model collection for transformers: huggingface.co/collections/... HF model collection for OpenCLIP and timm: huggingface.co/collections/... And of course big_vision checkpoints: github.com/google-resea...huggingface.coSigLIP2 - a google CollectionWe’re on a journey to advance and democratize artificial intelligence through open source and open science. 021
Reposted by Aritra Roy GosthipatyMichael Tschannen @mtschannen.bsky.social · 22/02/2025Paper: arxiv.org/abs/2502.14786 HF blog post from @arig23498.bsky.social et al. with a gentle intro to the training recipe and a demo: huggingface.co/blog/siglip2 Thread with results overview from Xiaohua (only on X, sorry - these are all in the paper): x.com/XiaohuaZhai/...arxiv.orgSigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense FeaturesWe introduce SigLIP 2, a family of new multilingual vision-language encoders that build on the success of the original SigLIP. In this second iteration, we extend the original image-text training obje... 111
Reposted by Aritra Roy GosthipatyMichael Tschannen @mtschannen.bsky.social · 22/02/2025📢2⃣ Yesterday we released SigLIP 2! TL;DR: Improved high-level semantics, localization, dense features, and multilingual capabilities via drop-in replacement for v1. Bonus: Variants supporting native aspect and variable sequence length. A thread with interesting resources👇 1131
Reposted by Aritra Roy GosthipatySung Kim @sungkim.bsky.social · 29/01/2025Build a Qwen 2.5 VL API endpoint with Hugging Face spaces and Docker! by @arig23498.bsky.social Build a proof-of-concept API, hosting Qwen2.5-VL-7B-Instruct on Hugging Face Spaces using Docker. huggingface.co/blog/ariG234...huggingface.co🚀 Build a Qwen 2.5 VL API endpoint with Hugging Face spaces and Docker!A Blog post by Aritra Roy Gosthipaty on Hugging Face 051
Aritra Roy Gosthipaty @arig23498.bsky.social · 23/12/2024huggingface.co/blog/logits-...huggingface.coControlling Language Model Generation with NVIDIA's LogitsProcessorZooWe’re on a journey to advance and democratize artificial intelligence through open source and open science. 072
Aritra Roy Gosthipaty @arig23498.bsky.social · 03/12/2024I forgot to mention that you can use the same code to access any `warm` model on the Hub. Here is a list of all the `warm` models: huggingface.co/models?infer... Happy vibe checking 😇 [N/N]huggingface.coModels - Hugging FaceWe’re on a journey to advance and democratize artificial intelligence through open source and open science. 020
Aritra Roy Gosthipaty @arig23498.bsky.social · 03/12/2024I have created a simple and quick notebook to access this inference api and use `huggingface_hub` to access the model through it. huggingface.co/datasets/ari... [4/N]huggingface.coqwq-inference-api.ipynb · ariG23498/quick-notebooks at mainWe’re on a journey to advance and democratize artificial intelligence through open source and open science. 130
Aritra Roy Gosthipaty @arig23498.bsky.social · 03/12/2024But today it was my lucky day. I noticed that the model was already loaded on the Serverless Inference API and was ready to be used. No more spinning up my GPUs and stress testing them (happy GPU noises) [3/N] 100
Aritra Roy Gosthipaty @arig23498.bsky.social · 03/12/2024My usual workflow is to visit the Hugging Face Hub model card (here that was hf[dot]co[dot]Qwen/QwQ-32B-Preview) and copy the working code sample. I am sure this is how most of you work with a new model as well (if not, I would love to hear from you) [2/N] 100
Aritra Roy Gosthipaty @arig23498.bsky.social · 03/12/2024The Qwen team is doing so much for the community by keeping research open and constructive. They listen to the community and put efforts in building competitive models. I was intrigued by their latest `Qwen/QwQ-32B-Preview` model and wanted to play with it. [1/N] 1101
Reposted by Aritra Roy GosthipatySergio Paniego @sergiopaniego.bsky.social · 29/11/2024I've been exploring the latest Llama 3.2 releases and working on a couple of projects you may find interesting: 1️⃣ Understanding tool calling with Llama 3.2 🔧 2️⃣ Using Text Generation Inference (TGI) with Llama models 🦙 (links in the next post) 1123
Aritra Roy Gosthipaty @arig23498.bsky.social · 26/11/2024I like the evaluation part. Is there some evals you particularly like? 100
Aritra Roy Gosthipaty @arig23498.bsky.social · 26/11/2024What is THE pain point in training Vision Language Models according to you? I will go first, the data pipeline. 310
Aritra Roy Gosthipaty @arig23498.bsky.social · 23/11/2024Re-caption your webdataset with Qwen2-VL github.com/sayakpaul/si...github.comAdding support for Qwen model by ariG23498 · Pull Request #3 · sayakpaul/simple-image-recaptioningA working colab notebook 0140
Aritra Roy Gosthipaty @arig23498.bsky.social · 20/11/2024huggingface.co/blog/layerskiphuggingface.coFaster Text Generation with Self-Speculative DecodingWe’re on a journey to advance and democratize artificial intelligence through open source and open science. 083
Aritra Roy Gosthipaty @arig23498.bsky.social · 15/11/2024To the video generation enthusiats, Mochi 1 Preview is now supported in `diffusers` 060
Reposted by Aritra Roy GosthipatyLain @not-so-lain.bsky.social · 13/11/2024awesome, thanks a lot for sharing 🙌 011
Aritra Roy Gosthipaty @arig23498.bsky.social · 13/11/2024`bitsandbytes` makes it really easy to quantize models Note: MB should be GB in the diagram. 171
Aritra Roy Gosthipaty @arig23498.bsky.social · 12/11/2024Read about the Qwen2.5-Coder Series huggingface.co/blog/ariG234... 010
Aritra Roy Gosthipaty @arig23498.bsky.social · 08/11/2024Training ranking models for better retrieval from stores is GOD level thinking. 000
Aritra Roy Gosthipaty @arig23498.bsky.social · 07/11/2024I am diving head first into Vision Language Models. Comment below the papers that I definitely should read. 020
Aritra Roy Gosthipaty @arig23498.bsky.social · 06/11/2024Welcome the @huggingface.bsky.social integration in PyCharm. From instant model cards to navigating the local cache, working with Hugging Face models becomes a lot easier with PyCharm. Bonus: Claim a 3 month PyCharm subscription using PyCharm4HF Blog Post: huggingface.co/blog/pycharm...huggingface.coHugging Face + PyCharmWe’re on a journey to advance and democratize artificial intelligence through open source and open science. 000
Aritra Roy Gosthipaty @arig23498.bsky.social · 06/11/2024github.com/ml-gde/jflux Try out the FLUX model in JAX. It also works on TPUs if that is your thing. For people who want to work on it, there are open issues as well. Happy coding!github.comGitHub - ml-gde/jflux: JAX Implementation of Black Forest Labs' Flux.1 family of modelsJAX Implementation of Black Forest Labs' Flux.1 family of models - ml-gde/jflux 0151
