Gowtham Ramesh @gowtham_ramesh1

New in-depth blog post - "Inside vLLM: Anatomy of a High-Throughput LLM Inference System". Probably the most in depth explanation of how LLM inference engines and vLLM in particular work! Took me a while to get this level of understanding of the codebase and then to write up…

Wow! Prof Mitesh Khapra's name is in Times 100/AI in the same list as Fei-Fei Li, @sama, @elonmusk, @JeffDean and so on. I bless the day I was invited to work alongside him in AI4Bharat! Truly a visionary! time.com/collections/ti… Reply 🐐 to show respect!

Introducing the Environments Hub RL environments are the key bottleneck to the next wave of AI progress, but big labs are locking them down We built a community platform for crowdsourcing open environments, so anyone can contribute to open-source AGI

Instella-Math is the first fully open math reasoning model that is trained on AMD GPUs. As part of AMD’s commitment to open innovation, we’re sharing the full model weights, training setup, codebase, and datasets to foster collaboration and transparency across the AI dev…

Emad Barsoum @EmadBarsoumPi

4 weeks ago

Instella-Math is the first fully open math reasoning model that is trained on AMD GPUs. As part of AMD’s commitment to open innovation, we’re sharing the full model weights, training setup, codebase, and datasets to foster collaboration and transparency across the AI dev… https://t.co/T5rDFic8gN

2 8 54 7K 3

@AMD first reasoning model!!! Expanding on Instella family of models, Instella-Math is based on Instella-3B-Instruct model with reasoning capabilities added with multi-stage training pipeline: 2 stages of SFT and 3 stages of RL. Trained on @AMD MI300X and fully open (training…

Meet Instella-Math ✨ — the first language model trained with long chain-of-thought reinforcement learning, entirely on AMD GPUs. 🫡 RL training on ROCm is no longer a problem. rocm.blogs.amd.com/artificial-int…

✨𝗜𝗻𝘁𝗿𝗼𝗱𝘂𝗰𝗶𝗻𝗴 𝗜𝗻𝘀𝘁𝗲𝗹𝗹𝗮-𝗠𝗮𝘁𝗵: a fully open reasoning-focused language model trained with long CoT reinforcement learning entirely on AMD Instinct™ MI300X GPUs. 🤗The base model (Instella-3B-Instruct), reasoning SFT and RL training are fully open.…

🎉Introducing 𝗦𝗔𝗡𝗗-𝗠𝗮𝘁𝗵: 𝗨𝘀𝗶𝗻𝗴 𝗟𝗟𝗠𝘀 𝘁𝗼 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗲 𝗡𝗼𝘃𝗲𝗹, 𝗗𝗶𝗳𝗳𝗶𝗰𝘂𝗹𝘁 𝗮𝗻𝗱 𝗨𝘀𝗲𝗳𝘂𝗹 𝗠𝗮𝘁𝗵𝗲𝗺𝗮𝘁𝗶𝗰𝘀 𝗤𝘂𝗲𝘀𝘁𝗶𝗼𝗻𝘀 𝗮𝗻𝗱 𝗔𝗻𝘀𝘄𝗲𝗿𝘀 SAND-Math 𝗶𝘀 𝗮 𝘀𝘆𝗻𝘁𝗵𝗲𝘁𝗶𝗰 𝗱𝗮𝘁𝗮 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 𝗽𝗶𝗽𝗲𝗹𝗶𝗻𝗲…

🎉Introducing TTT-Bench: A new benchmark for evaluating reasoning ability with Simple and Novel Tic-Tac-Toe-style Games🧩 TTT-Bench, is designed to evaluate basic strategic, spatial, and logical reasoning abilities in LRMs through a suite novel two-player Tic-Tac-Toe-style games…

Introducing Instella-Long model that we trained from scratch, on MI300X, with up to 128K sequence length, fully open: checkpoints, code and dataset. And with competitive benchmark outperforming open-weights models on the long-context benchmark. @AIatAMD @AMD…

🚨Introducing Instella-3B-Long-Instruct✨ - 128K context length 📏 - Trained on 64 Instinct MI300X GPUs ⚡️ - Fully open model trained from scratch that supports long-context. 🤗 - Release model weights, detailed training configurations, datasets, and code. Blog📰:…

Jialian Wu @Wu_Jialian

3 months ago

🚨Introducing Instella-3B-Long-Instruct✨ - 128K context length 📏 - Trained on 64 Instinct MI300X GPUs ⚡️ - Fully open model trained from scratch that supports long-context. 🤗 - Release model weights, detailed training configurations, datasets, and code. Blog📰:…

1 1 5 910 0

We released Instella-Long, a fully open 3B long-context language model supporting 128K context length, trained on AMD MI300X GPUs. Training data, training code, and model weights are released! rocm.blogs.amd.com/artificial-int… #LLM #AMD #Longcontext #AI

Today we introduce Sarvam-M, a 24B open-weights hybrid model built on top of Mistral Small. Sarvam-M achieves a new benchmark across a range of Indian languages, math, and programming tasks, for a model of its size. Here is a detailed technical blog on how we customize…

Releasing INTELLECT-2: We’re open-sourcing the first 32B parameter model trained via globally distributed reinforcement learning: • Detailed Technical Report • INTELLECT-2 model checkpoint primeintellect.ai/blog/intellect…

It's official 🇮🇳 We're proud to announce that Sarvam has been selected by the Government of India under the IndiaAI Mission to build India's sovereign Large Language Model. Building India's sovereign model from the ground up is a crucial step toward Atmanirbhar Bharat. The…

I think CS336 has one of the best LLM problem sets of any AI/LM class thanks to our incredible TAs (@nelsonfliu,@GabrielPoesia,@marcelroed,@neilbband,@rckpudi). We're making it so you can do it all at home, and it's one of the best ways to learn LLMs deeply.

Stanford NLP Group @stanfordnlp

5 months ago

I think CS336 has one of the best LLM problem sets of any AI/LM class thanks to our incredible TAs (@nelsonfliu,@GabrielPoesia,@marcelroed,@neilbband,@rckpudi). We're making it so you can do it all at home, and it's one of the best ways to learn LLMs deeply.

10 162 1K 198K 1K

Reinforcement Learning from Human Feedback via VeRL on AMD MI300X and ROCm, single and multiple nodes training. @thu_yushengsu @AMD rocm.blogs.amd.com/artificial-int…

AMD has announced the open-sourcing of Instella, a fully open 3-billion-parameter LMs trained on AMD Instinct MI300X GPUs. These models aim to promote collaboration in the AI community by providing access to model weights, configurations, and more. See more from @phoronix ⬇️

Phoronix @phoronix

6 months ago

AMD has announced the open-sourcing of Instella, a fully open 3-billion-parameter LMs trained on AMD Instinct MI300X GPUs. These models aim to promote collaboration in the AI community by providing access to model weights, configurations, and more. See more from @phoronix ⬇️

1 18 105 34K 12

@ai4bharat started with a few students experimenting with deep learning projects at IIT Madras. What began with building small models, trying out computer vision, and translating text turned into a focused effort on Indian languages, and eventually the creation of AI4Bharat.…

🚀AMD Instinct™ GPUs go-brrrrr 💡Key releases and blogs: 🤖Model Releases: Introducing Instella✨: New State-of-the-art Fully Open 3B Language Models 🔸A family of fully open state-of-the-art 3-billion-parameter language models (LMs) trained from scratch on AMD Instinct™…

keyboard_arrow_left Previous keyboard_arrow_right Next