Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

We’re thrilled to open-source TriAttention! 🚀 🦞 Deploy OpenClaw (32B LLM) on a single 24GB RTX 4090 locally 💻Full code open-source & vLLM-ready for one-click deployment ⚡️ 2.5× faster inference speed & 10.7× less KV cache memory usage TriAttention is a novel KV cache compression method built on rigorous... show more

Yukang Chen

1,589 subscribers

197,334 Aufrufe • vor 3 Monaten •via X (Twitter)

Bildung Wissenschaft & Technologie

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

I added KV caching and INT8 KV quantization to our transformer inference, improving throughput by 35x. All of this was done from scratch in Rust + CUDA, on top of a homemade ML framework. On a 4-token prompt with 252 generated tokens: - Original: 0.76 tok/s - KV cache fp32: 27.21 tok/s - KV cache int8 (quantized): 27.29 tok/s Try it out yourself here: In practice: - KV caching gave us about a 35x end-to-end speedup - INT8 KV cache kept roughly the same speed as fp32 but cut KV cache memory by 3.78x FP32 cache used 4.5 MB in this run while the INT8 cache used only 1.19 MB This simple change to inference created a huge impact on performance. To learn more about the KV cache and other optimizations like this, check out the blog at

I added KV caching and INT8 KV quantization to our transformer inference, improving throughput by 35x. All of this was done from scratch in Rust + CUDA, on top of a homemade ML framework. On a 4-token prompt with 252 generated tokens: - Original: 0.76 tok/s - KV cache fp32: 27.21 tok/s - KV cache int8 (quantized): 27.29 tok/s Try it out yourself here: In practice: - KV caching gave us about a 35x end-to-end speedup - INT8 KV cache kept roughly the same speed as fp32 but cut KV cache memory by 3.78x FP32 cache used 4.5 MB in this run while the INT8 cache used only 1.19 MB This simple change to inference created a huge impact on performance. To learn more about the KV cache and other optimizations like this, check out the blog at

Reese Chong

52,588 Aufrufe • vor 2 Monaten

Hot off the press 🔥 AEON vLLM Ultimate → v0.24.0. Link below ⤵️ The ultimate container for AI on NVIDIA DGX Spark — vLLM compiled from source for GB10/sm_121a. 🔓 UNLOCKED (impossible until this release): ▸ DFlash + FP8 KV cache — 2× KV capacity, draft acceptance intact, measured on-device ▸ FP8 KV on vision models — the KV-bandwidth lever is back ▸ DFlash on FlashInfer & Triton backends (was FlashAttention-only) 🛠 Carried ahead of upstream (still-open PRs, merged + validated): ▸ Triton NVFP4 KV cache → up to 3× KV capacity ▸ DFlash sliding-window drafters + prefix-cache stability for DFlash ▸ High-concurrency DFlash — clean through c=64 🧰 Fixes stock v0.24.0 doesn't have: ▸ Gemma-4 tied embeddings (ModelOpt) — stock crashes at load, ours doesn't ▸ Tool calling out of the box (stock --no-deps builds 500) ▸ Unified-memory hardened — no silent KV overcommit on the 121 GB pool 📊 One image, the whole fleet: Gemma-4-26B-A4B · Qwen3.6-27B · Qwen3.6-35B-A3B (hybrid GDN) 536 tok/s @ c16 · 472 tok/s @ c12 · 262K context · image+video · tools + reasoning parsers · full CUDA-graph capture 🐳 docker pull

Hot off the press 🔥 AEON vLLM Ultimate → v0.24.0. Link below ⤵️ The ultimate container for AI on NVIDIA DGX Spark — vLLM compiled from source for GB10/sm_121a. 🔓 UNLOCKED (impossible until this release): ▸ DFlash + FP8 KV cache — 2× KV capacity, draft acceptance intact, measured on-device ▸ FP8 KV on vision models — the KV-bandwidth lever is back ▸ DFlash on FlashInfer & Triton backends (was FlashAttention-only) 🛠 Carried ahead of upstream (still-open PRs, merged + validated): ▸ Triton NVFP4 KV cache → up to 3× KV capacity ▸ DFlash sliding-window drafters + prefix-cache stability for DFlash ▸ High-concurrency DFlash — clean through c=64 🧰 Fixes stock v0.24.0 doesn't have: ▸ Gemma-4 tied embeddings (ModelOpt) — stock crashes at load, ours doesn't ▸ Tool calling out of the box (stock --no-deps builds 500) ▸ Unified-memory hardened — no silent KV overcommit on the 121 GB pool 📊 One image, the whole fleet: Gemma-4-26B-A4B · Qwen3.6-27B · Qwen3.6-35B-A3B (hybrid GDN) 536 tok/s @ c16 · 472 tok/s @ c12 · 262K context · image+video · tools + reasoning parsers · full CUDA-graph capture 🐳 docker pull

ÆON FORGE ✨

24,160 Aufrufe • vor 7 Tagen

Someone built a free and better alternative to Claude that runs 100% locally. → works with any LLM (Claude, GPT, Gemini, vLLM) → beats it on deep research → has Cowork-like capabilities → 50+ connectors out of the box → deploy in literally one command 100% open source. MIT license. 28k stars.

Someone built a free and better alternative to Claude that runs 100% locally. → works with any LLM (Claude, GPT, Gemini, vLLM) → beats it on deep research → has Cowork-like capabilities → 50+ connectors out of the box → deploy in literally one command 100% open source. MIT license. 28k stars.

How To Prompt

39,202 Aufrufe • vor 2 Monaten

Turn any GitHub repository into LLM-ready text! Simply replace "hub" with "ingest" in a GitHub URL and receive a prompt-friendly text ingest for LLMs. Gitingest is 100% open-source and provides: - Directory structure - A brief summary of the project - The entire content as LLM-ready text Plus, it comes with a nice python package and you can run the UI locally! Stay tuned, I'm working on something really cool with this!✨ Link to the GitHub repo in the next tweet! _____ Find me → Akshay 🚀 ✔️ For more insights and tutorials on ML and AI Engineering!

Turn any GitHub repository into LLM-ready text! Simply replace "hub" with "ingest" in a GitHub URL and receive a prompt-friendly text ingest for LLMs. Gitingest is 100% open-source and provides: - Directory structure - A brief summary of the project - The entire content as LLM-ready text Plus, it comes with a nice python package and you can run the UI locally! Stay tuned, I'm working on something really cool with this!✨ Link to the GitHub repo in the next tweet! _____ Find me → Akshay 🚀 ✔️ For more insights and tutorials on ML and AI Engineering!

Akshay 🚀

191,335 Aufrufe • vor 1 Jahr

google's notebookLM charges $19.99/mo for pro. this open-source alternative gives you unlimited access for $0. zero subscription. runs locally. MIT licensed. features: → podcast generation → multi-modal ingestion → up to 6 languages → universal LLM selection github repo link in the comments.

google's notebookLM charges $19.99/mo for pro. this open-source alternative gives you unlimited access for $0. zero subscription. runs locally. MIT licensed. features: → podcast generation → multi-modal ingestion → up to 6 languages → universal LLM selection github repo link in the comments.

m0h

28,464 Aufrufe • vor 12 Tagen

A tricky LLM interview question: You're serving a reasoning model on vLLM, and it keeps running out of GPU memory on long traces. So you add KV cache compression and evict 90% of the cached tokens. VRAM usage stays as is and GPU still runs out of memory. Why? (answer below) Evicting 90% of the KV cache can free almost none of the memory it was using. This sounds counterintuitive, but it follows directly from how production servers store the cache today. The KV cache grows with every token a model generates. Each token appends its key and value vectors across every layer, and nothing is freed while generation continues. This is the dominant memory cost for reasoning models. If a 32K-token CoT caches ~32K tokens of KV vectors, a Qwen3-32B with 4-bit weights will run out-of-memory around 24K tokens on a 24GB GPU. One obvious solution is to keep the important tokens and drop the rest, since attention is sparse enough to allow it. But this does not solve the memory problem yet. The reason is paged attention, which is the memory manager behind vLLM and most production servers. Under the hood, it splits GPU memory into fixed physical blocks, each one holds the KV for about 16 tokens. This block returns to the allocator only when every slot inside it is empty. Since the eviction logic selects tokens by importance, and such tokens are scattered across blocks... ...so despite eviction, almost every block is left with at least some survivor tokens. For instance, if the logic evicts 14k of 16k tokens across 1,000 blocks, most likely every block will still have a token. This means the allocator frees almost nothing. Placing the new tokens into those freed slots is not ideal because it breaks the cache's layout. Say token 16,001 arrives, and it's placed in the slot the 40th token used to hold. The cache now reads position 38, then 16,001, then 41, so the cache is no longer in token order. Attention can still compute the right answer from that, but only if every slot now carries a separate note recording which position it actually holds. This introduces another bookkeeping cost that an in-order layout inherently avoids. So the cache is logically 90% smaller and still physically the same size. Many compression results miss this because they measure on pre-allocated contiguous tensors rather than a paged server. There's another problem. Eviction methods pick which tokens to keep by looking at the attention scores themselves (as expected). But fast attention kernels used in production, like FlashAttention, never save those scores. They compute attention in small pieces and throw the full score grid away as they go, which is also why they're fast. So the exact signal eviction methods need isn't available in memory. The workaround is to fall back to eager attention and build the full matrix, which gives up the speed FlashAttention was there to provide. NVIDIA published a method called TriAttention to solve both these problems. It never needs attention scores. Instead, it scores tokens from the geometry of the model's key and query vectors before RoPE is applied, where those vectors sit in stable clusters. For the memory problem, it runs a compaction pass every 128 decoded tokens. The surviving tokens slide forward to close the holes eviction creates, so whole blocks empty out and return to the allocator while the cache stays in token order. On long reasoning traces, the approach matches full-attention accuracy while decoding 2.5x faster and using 10.7x less KV memory. KV cache compression is a big infrastructure problem. The number that decides whether it works is the count of freed blocks, not the count of evicted tokens. You can find the NVIDIA write-up here: I wrote a first-principles breakdown of how the KV cache works. It walks through why the model stores keys and values at all, why the cache grows with every token, and a comparison of LLM generation speed with and without KV caching. Read it below.

A tricky LLM interview question: You're serving a reasoning model on vLLM, and it keeps running out of GPU memory on long traces. So you add KV cache compression and evict 90% of the cached tokens. VRAM usage stays as is and GPU still runs out of memory. Why? (answer below) Evicting 90% of the KV cache can free almost none of the memory it was using. This sounds counterintuitive, but it follows directly from how production servers store the cache today. The KV cache grows with every token a model generates. Each token appends its key and value vectors across every layer, and nothing is freed while generation continues. This is the dominant memory cost for reasoning models. If a 32K-token CoT caches ~32K tokens of KV vectors, a Qwen3-32B with 4-bit weights will run out-of-memory around 24K tokens on a 24GB GPU. One obvious solution is to keep the important tokens and drop the rest, since attention is sparse enough to allow it. But this does not solve the memory problem yet. The reason is paged attention, which is the memory manager behind vLLM and most production servers. Under the hood, it splits GPU memory into fixed physical blocks, each one holds the KV for about 16 tokens. This block returns to the allocator only when every slot inside it is empty. Since the eviction logic selects tokens by importance, and such tokens are scattered across blocks... ...so despite eviction, almost every block is left with at least some survivor tokens. For instance, if the logic evicts 14k of 16k tokens across 1,000 blocks, most likely every block will still have a token. This means the allocator frees almost nothing. Placing the new tokens into those freed slots is not ideal because it breaks the cache's layout. Say token 16,001 arrives, and it's placed in the slot the 40th token used to hold. The cache now reads position 38, then 16,001, then 41, so the cache is no longer in token order. Attention can still compute the right answer from that, but only if every slot now carries a separate note recording which position it actually holds. This introduces another bookkeeping cost that an in-order layout inherently avoids. So the cache is logically 90% smaller and still physically the same size. Many compression results miss this because they measure on pre-allocated contiguous tensors rather than a paged server. There's another problem. Eviction methods pick which tokens to keep by looking at the attention scores themselves (as expected). But fast attention kernels used in production, like FlashAttention, never save those scores. They compute attention in small pieces and throw the full score grid away as they go, which is also why they're fast. So the exact signal eviction methods need isn't available in memory. The workaround is to fall back to eager attention and build the full matrix, which gives up the speed FlashAttention was there to provide. NVIDIA published a method called TriAttention to solve both these problems. It never needs attention scores. Instead, it scores tokens from the geometry of the model's key and query vectors before RoPE is applied, where those vectors sit in stable clusters. For the memory problem, it runs a compaction pass every 128 decoded tokens. The surviving tokens slide forward to close the holes eviction creates, so whole blocks empty out and return to the allocator while the cache stays in token order. On long reasoning traces, the approach matches full-attention accuracy while decoding 2.5x faster and using 10.7x less KV memory. KV cache compression is a big infrastructure problem. The number that decides whether it works is the count of freed blocks, not the count of evicted tokens. You can find the NVIDIA write-up here: I wrote a first-principles breakdown of how the KV cache works. It walks through why the model stores keys and values at all, why the cache grows with every token, and a comparison of LLM generation speed with and without KV caching. Read it below.

Avi Chawla

265,992 Aufrufe • vor 12 Tagen

Turn any open-source LLM into reasoning powerhouse! Using reinforcement finetuning you can add reasoning abilities to any LLM, even without a labelled dataset. Step-by-step explanation with code:

Turn any open-source LLM into reasoning powerhouse! Using reinforcement finetuning you can add reasoning abilities to any LLM, even without a labelled dataset. Step-by-step explanation with code:

Akshay 🚀

50,423 Aufrufe • vor 1 Jahr

We’re thrilled to open-source LabClaw — the Skill Operating Layer for LabOS by Stanford-Princeton Team One command turns any OpenClaw agent into a full AI Co-Scientist. Demo: Dragon Shrimp Army reporting for duty 🦞🔬 #AIforScience #OpenClaw

We’re thrilled to open-source LabClaw — the Skill Operating Layer for LabOS by Stanford-Princeton Team One command turns any OpenClaw agent into a full AI Co-Scientist. Demo: Dragon Shrimp Army reporting for duty 🦞🔬 #AIforScience #OpenClaw

AI4Science Catalyst

445,829 Aufrufe • vor 4 Monaten

What compression looks like on vLLM. Same Gemma 4 31B. Red Hat AI's quantized version runs at nearly 2x tokens/sec, half the memory, 99%+ accuracy retained. Open source. Quantized with LLM Compressor. Links in comments. 🙏 Sawyer Bowerman for the 2-minute demo.

What compression looks like on vLLM. Same Gemma 4 31B. Red Hat AI's quantized version runs at nearly 2x tokens/sec, half the memory, 99%+ accuracy retained. Open source. Quantized with LLM Compressor. Links in comments. 🙏 Sawyer Bowerman for the 2-minute demo.

Red Hat AI

34,136 Aufrufe • vor 2 Monaten

Sentra just killed Google Research's TurboQuant. SpectralQuant — 5.95× KV cache compression on Mistral 7B at +7.5% perplexity overhead. TurboQuant at the same compression: +22%. 3× less degradation. 15-second calibration. One per-model, then drop-in for any HuggingFace LLM, ViT, ESM, AlphaFold Evoformer, or VideoMAE. Check out the findings and how the mechanism works below. ↓

Sentra just killed Google Research's TurboQuant. SpectralQuant — 5.95× KV cache compression on Mistral 7B at +7.5% perplexity overhead. TurboQuant at the same compression: +22%. 3× less degradation. 15-second calibration. One per-model, then drop-in for any HuggingFace LLM, ViT, ESM, AlphaFold Evoformer, or VideoMAE. Check out the findings and how the mechanism works below. ↓

Ashwin Gopinath

59,538 Aufrufe • vor 1 Monat

Introducing MiniCPM 4.1-8B: First Open-Source Reasoning LLM with Trainable Sparse Attention ✅ Strong Reasoning Capability: Surpasses similar-sized models on 15 tasks! ✅ Fast Generation: 3x decoding speedup for reasoning ✅ Efficient Architecture: Trainable sparse attention, frequency-ranked speculative decoding Download Models: Huggingface: Github: Technical Report: #AI #MiniCPM #LLM #OpenBMB #ArtificialIntelligence #MachineLearning

Introducing MiniCPM 4.1-8B: First Open-Source Reasoning LLM with Trainable Sparse Attention ✅ Strong Reasoning Capability: Surpasses similar-sized models on 15 tasks! ✅ Fast Generation: 3x decoding speedup for reasoning ✅ Efficient Architecture: Trainable sparse attention, frequency-ranked speculative decoding Download Models: Huggingface: Github: Technical Report: #AI #MiniCPM #LLM #OpenBMB #ArtificialIntelligence #MachineLearning

OpenBMB

19,236 Aufrufe • vor 10 Monaten

🚀 Excited to release LongLive 2.0! 🎬 An end-to-end infrastructure for long video generation, with FP4 and parallelism at the core of both training and inference. ⚡45.7 FPS generation speed on 5B model⚡ ✨ LongLive 2.0 supports real-video training, few-step distillation, multi-shot training/inference, sequence-parallel acceleration, NVFP4 KV cache, and async VAE decoding deployment. 🧩 To our knowledge, this is the first open-source 4-bit long video generation infra that covers both training and inference. 🙌 Welcome to check it out, try it, and share feedback! 🔗 Code: 📰 Paper: 🎥 Demo: #LongVideoGeneration #VideoGeneration #Realtime #AIInfra #EfficientAI #FP4 #Parallel #NVIDIA

🚀 Excited to release LongLive 2.0! 🎬 An end-to-end infrastructure for long video generation, with FP4 and parallelism at the core of both training and inference. ⚡45.7 FPS generation speed on 5B model⚡ ✨ LongLive 2.0 supports real-video training, few-step distillation, multi-shot training/inference, sequence-parallel acceleration, NVFP4 KV cache, and async VAE decoding deployment. 🧩 To our knowledge, this is the first open-source 4-bit long video generation infra that covers both training and inference. 🙌 Welcome to check it out, try it, and share feedback! 🔗 Code: 📰 Paper: 🎥 Demo: #LongVideoGeneration #VideoGeneration #Realtime #AIInfra #EfficientAI #FP4 #Parallel #NVIDIA

Yukang Chen

58,301 Aufrufe • vor 1 Monat

Just built an MCP for Ghidra. Now basically any LLM (Claude, Gemini, local...) can Reverse Engineer malware for you. With the right prompting, it automates a *ton* of tedious tasks. One-shot markups of entire binaries with just a click. Open source, on Github now.

Just built an MCP for Ghidra. Now basically any LLM (Claude, Gemini, local...) can Reverse Engineer malware for you. With the right prompting, it automates a ton of tedious tasks. One-shot markups of entire binaries with just a click. Open source, on Github now.

LaurieWired

284,828 Aufrufe • vor 1 Jahr

LLM inference speed with vs. without KV caching: (learn how and why it works below)

LLM inference speed with vs. without KV caching: (learn how and why it works below)

Avi Chawla

395,232 Aufrufe • vor 3 Monaten

LLM inference speed with vs. without KV caching: (learn how and why it works below)

LLM inference speed with vs. without KV caching: (learn how and why it works below)

Daily Dose of Data Science

59,218 Aufrufe • vor 2 Monaten

We shipped OpenClaw for Windows 🦞 – Free with your LLM api keys – Custom skills from ClawHub – File system access - Browser control - Local Models (soon) - Open source. Inspired by Peter Steinberger 🦞

We shipped OpenClaw for Windows 🦞 – Free with your LLM api keys – Custom skills from ClawHub – File system access - Browser control - Local Models (soon) - Open source. Inspired by Peter Steinberger 🦞

atomicbot.ai

91,906 Aufrufe • vor 3 Monaten

We’ve been working on exploring building a library of open source UI patterns for agents on mobile, starting with Live Activities for observability. It’s available on GitHub and compatible with 🦞OpenClaw.

We’ve been working on exploring building a library of open source UI patterns for agents on mobile, starting with Live Activities for observability. It’s available on GitHub and compatible with 🦞OpenClaw.

Aaron Abentheuer

189,580 Aufrufe • vor 4 Monaten

📢 It's the open source news heard 'round the world: The GitHub Copilot Chat extension for Visual Studio Code is now fully open source under the MIT license. 🎉 Check out the repo. ⬇️

📢 It's the open source news heard 'round the world: The GitHub Copilot Chat extension for Visual Studio Code is now fully open source under the MIT license. 🎉 Check out the repo. ⬇️

GitHub

27,773 Aufrufe • vor 11 Monaten

Seedance 2 just dropped as open-source.!.. ok, mostly... Helios - 14B model for minute-scale vid gen. - high-quality, fast -No KV-cache, no quantization, no TinyVAE, just pure architectural efficiency. - T2V/I2V/V2V ready.

Seedance 2 just dropped as open-source.!.. ok, mostly... Helios - 14B model for minute-scale vid gen. - high-quality, fast -No KV-cache, no quantization, no TinyVAE, just pure architectural efficiency. - T2V/I2V/V2V ready.

Wildminder

25,148 Aufrufe • vor 4 Monaten

Introducing DuoAttention: Our new framework slashes both memory and latency for long-context LLMs without sacrificing performance! By applying full KV cache only to critical heads, we achieve: ⚡ 2.55x memory reduction ⚡ 2.18x decoding speedup ⚡ 3.3M tokens on a single A100 GPU

Introducing DuoAttention: Our new framework slashes both memory and latency for long-context LLMs without sacrificing performance! By applying full KV cache only to critical heads, we achieve: ⚡ 2.55x memory reduction ⚡ 2.18x decoding speedup ⚡ 3.3M tokens on a single A100 GPU

Guangxuan Xiao

31,023 Aufrufe • vor 1 Jahr