Video yükleniyor...

Video Yüklenemedi

Bu video yüklenirken bir sorun oluştu. Bu geçici bir ağ sorunundan kaynaklanıyor olabilir veya video kullanılamıyor olabilir.

Ana Sayfaya Dön

Yay, llama2.c can now load and inference the Meta released models! :) E.g. here inferencing the smallest 7B model at ~3 tokens/s on 96 OMP threads on a cloud Linux box. Still just CPU, fp32, one single .c file of 500 lines: expecting ~300 tok/s tomorrow :)

Andrej Karpathy

3,057,405 subscribers

409,394 görüntüleme • 3 yıl önce •via X (Twitter)

Eğitim Bilim & Teknoloji

Anya Rossi• Live Now

Private livecam show

9 Yorum

Andrej Karpathy profil fotoğrafı

Andrej Karpathy3 yıl önce

If we can get 7B model to run at nice and interactive rates then we can go from "scratch-trained micromodels" to "LoRA finetuned 7B base model", all within the code of the minimal llama2.c repo (both training and inference). Can reach more capability and with less training data.

Matt Schwab profil fotoğrafı

Matt Schwab3 yıl önce

Translation from ChatGPT: Great news! The program 'llama2.c' has been updated, and now it can handle and process the Meta released models. For example, it can now process the smallest 7-billion model at around 3 tokens per second using 96 OMP threads on a cloud Linux computer. It's worth noting that this is all done using just the CPU, which is the main processor of the computer, and it's using a standard type of computing precision called fp32. What's even more impressive is that the entire program is contained in a single .c file with only 500 lines of code.

Loreto Parisi profil fotoğrafı

Loreto Parisi3 yıl önce

I know it's a minimalistic design, but it would be cool to add MPS via Accelerate and eventually MPI for distributed inference

ai geek (wishesh) ⚡️ profil fotoğrafı

ai geek (wishesh) ⚡️3 yıl önce

We are getting there. Let's go 🚀

Andreas Cambitsis profil fotoğrafı

Andreas Cambitsis3 yıl önce

How much RAM is available and being used?

Nikhil Verma profil fotoğrafı

Nikhil Verma3 yıl önce

You make AI approachable and accessibile for all. Thank you for that its incredible to see so much can be achieved with so little amount of code

Kratius profil fotoğrafı

Kratius3 yıl önce

I included -fopenmp while compiling because omp_num_threads were unable to use all the cores. On 15 core machine achieving 41 tok/s

arkephi profil fotoğrafı

arkephi3 yıl önce

I always knew efficiency was the answer

grimes apologist profil fotoğrafı

grimes apologist3 yıl önce

I love you and I would die for you

Benzer Videolar

Llama 2 (7B, 13B and 70B) is now available in Lit-GPT! 🚀 Try running inference on the chat model or finetune the foundation model ➡️ #Llama2 #MachineLearning #opensource

Llama 2 (7B, 13B and 70B) is now available in Lit-GPT! 🚀 Try running inference on the chat model or finetune the foundation model ➡️ #Llama2 #MachineLearning #opensource

Lightning AI ⚡️

17,845 görüntüleme • 3 yıl önce

LLM running on Windows 98 PC 26 year old hardware with Intel Pentium II CPU and 128MB RAM. Uses llama98.c, our custom pure C inference engine based on Andrej Karpathy llama2.c Code and DIY guide 👇

LLM running on Windows 98 PC 26 year old hardware with Intel Pentium II CPU and 128MB RAM. Uses llama98.c, our custom pure C inference engine based on Andrej Karpathy llama2.c Code and DIY guide 👇

EXO Labs

483,032 görüntüleme • 1 yıl önce

Qwen3.5-35B with only 3B active parameters This MoE model runs FASTER than most 7B dense models. Tested on 3 generations of NVIDIA: - 5090: 137 tok/s - 4090: 112 tok/s - 3090: 78 tok/s The surprise? The 4090 <> 5090 gap is only 22%. With a 3B active MoE, even old GPUs fly.

Qwen3.5-35B with only 3B active parameters This MoE model runs FASTER than most 7B dense models. Tested on 3 generations of NVIDIA: - 5090: 137 tok/s - 4090: 112 tok/s - 3090: 78 tok/s The surprise? The 4090 <> 5090 gap is only 22%. With a 3B active MoE, even old GPUs fly.

stevibe

69,501 görüntüleme • 4 ay önce

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 görüntüleme • 2 ay önce

PowerInfer - a high-speed inference engine for deploying LLMs locally. Just came across this super interesting project on speeding up inference. It's not MoE but it's a simple approach that exploits the high locality in LLM inference to design a GPU-CPU hybrid inference engine. Hot-activated neurons are preloaded onto the GPU for fast access, while cold-activated neurons (the majority) are computed on the CPU. This approach significantly reduces GPU memory demands and CPU-GPU data transfer. It achieves an average token generation rate of 13.20 tokens/s, with a peak of 29.08 tokens/s, across various LLMs on a single NVIDIA RTX 4090 GPU. It's on only 18% lower than that achieved by a top-tier server-grade A100 GPU. It also significantly outperforms llama.cpp by up to 11.69x while retaining model accuracy. There is a lot more innovation around inference that's coming fast. Really encouraged by the study on sparse computation to enhance the computational efficiency of LLMs. It's now possible to use PowerInfer with Llama 2 and Faclon 40B. Mistral-7B support is coming soon!

PowerInfer - a high-speed inference engine for deploying LLMs locally. Just came across this super interesting project on speeding up inference. It's not MoE but it's a simple approach that exploits the high locality in LLM inference to design a GPU-CPU hybrid inference engine. Hot-activated neurons are preloaded onto the GPU for fast access, while cold-activated neurons (the majority) are computed on the CPU. This approach significantly reduces GPU memory demands and CPU-GPU data transfer. It achieves an average token generation rate of 13.20 tokens/s, with a peak of 29.08 tokens/s, across various LLMs on a single NVIDIA RTX 4090 GPU. It's on only 18% lower than that achieved by a top-tier server-grade A100 GPU. It also significantly outperforms llama.cpp by up to 11.69x while retaining model accuracy. There is a lot more innovation around inference that's coming fast. Really encouraged by the study on sparse computation to enhance the computational efficiency of LLMs. It's now possible to use PowerInfer with Llama 2 and Faclon 40B. Mistral-7B support is coming soon!

elvis

261,622 görüntüleme • 2 yıl önce

The RTX 3090 is a 5-year-old GPU and it still runs a 27B model at 20 tok/s I tested Qwen3.5:27b across 3 generations of NVIDIA: 5090 → ~60 tok/s 4090 → ~40 tok/s 3090 → ~20 tok/s Perfectly linear scaling. Double the generation, double the speed.

The RTX 3090 is a 5-year-old GPU and it still runs a 27B model at 20 tok/s I tested Qwen3.5:27b across 3 generations of NVIDIA: 5090 → ~60 tok/s 4090 → ~40 tok/s 3090 → ~20 tok/s Perfectly linear scaling. Double the generation, double the speed.

stevibe

142,379 görüntüleme • 4 ay önce

Wrote a brutal C++23 CPU monitor for my KianV RISC-V Linux SoC on FPGA. It uses 8 GPIO LEDs to show live CPU load. Clean modern C++, signal-safe shutdown, no mercy , just pure hardware feedback.

Wrote a brutal C++23 CPU monitor for my KianV RISC-V Linux SoC on FPGA. It uses 8 GPIO LEDs to show live CPU load. Clean modern C++, signal-safe shutdown, no mercy , just pure hardware feedback.

logic destroyer

17,838 görüntüleme • 1 yıl önce

Curious about ollama kimi-k2.6:cloud speed? 3 test runs: > 77.9 tok/s, TTFT 979ms > 114.3 tok/s, TTFT 788ms > 86.3 tok/s, TTFT 1117ms For comparison, OpenRouter stats: > Parasail: 14 tok/s > Moonshot AI: 27 tok/s > NovitaAI: 27 tok/s > Cloudflare: 71 tok/s Obvious caveat: cloud speeds fluctuate with load. Just sharing numbers for the curious.

Curious about ollama kimi-k2.6:cloud speed? 3 test runs: > 77.9 tok/s, TTFT 979ms > 114.3 tok/s, TTFT 788ms > 86.3 tok/s, TTFT 1117ms For comparison, OpenRouter stats: > Parasail: 14 tok/s > Moonshot AI: 27 tok/s > NovitaAI: 27 tok/s > Cloudflare: 71 tok/s Obvious caveat: cloud speeds fluctuate with load. Just sharing numbers for the curious.

stevibe

35,129 görüntüleme • 2 ay önce

DFlash speculative decoding on Apple Silicon Qwen3.5-9B bf16 · M5 Max · greedy exact match ▸ 85 tok/s, 3.3× at 1024 tokens (runtime) ▸ ~70 tok/s, 2.6× in the video (terminal I/O overhead) ▸ 80 tok/s, 3.1× at 2048 tokens (runtime) Currently working on: → Long context (speedup degrades past 4K tokens, KV cache growth) → Int4 quantized models (27B class) Built on MLX, no CUDA, single machine. Draft generates 16 tokens in parallel, target verifies in one forward pass. Will open source when ready.

DFlash speculative decoding on Apple Silicon Qwen3.5-9B bf16 · M5 Max · greedy exact match ▸ 85 tok/s, 3.3× at 1024 tokens (runtime) ▸ ~70 tok/s, 2.6× in the video (terminal I/O overhead) ▸ 80 tok/s, 3.1× at 2048 tokens (runtime) Currently working on: → Long context (speedup degrades past 4K tokens, KV cache growth) → Int4 quantized models (27B class) Built on MLX, no CUDA, single machine. Draft generates 16 tokens in parallel, target verifies in one forward pass. Will open source when ready.

bstn 👁️

36,942 görüntüleme • 2 ay önce

Last week we released Meta Chameleon: a new mixed-modal research model from Meta FAIR. Get the models ➡️ The 7B & 34B safety tuned models we’ve released can take any combination of text and images as input and produce text outputs using a new early fusion approach. While some LLMs have separate image and text encoders or decoders, Chameleon is one of the first publicly released approaches using a single unified architecture. We’re releasing Chameleon models under a research license to help democratize access to foundational mixed-modal models & further research on early fusion. Approach & training details in the paper ➡️

Last week we released Meta Chameleon: a new mixed-modal research model from Meta FAIR. Get the models ➡️ The 7B & 34B safety tuned models we’ve released can take any combination of text and images as input and produce text outputs using a new early fusion approach. While some LLMs have separate image and text encoders or decoders, Chameleon is one of the first publicly released approaches using a single unified architecture. We’re releasing Chameleon models under a research license to help democratize access to foundational mixed-modal models & further research on early fusion. Approach & training details in the paper ➡️

AI at Meta

54,410 görüntüleme • 2 yıl önce

I want cloud computing to feel like a video game. Running Llama-2-7b is slow on base Replit ⠕? Boost 32x in 2 clicks without stopping the program and it goes from ~0 tok/s to 4 tok/s. Expand to infinity!

I want cloud computing to feel like a video game. Running Llama-2-7b is slow on base Replit ⠕? Boost 32x in 2 clicks without stopping the program and it goes from ~0 tok/s to 4 tok/s. Expand to infinity!

Amjad Masad

34,575 görüntüleme • 2 yıl önce

Cerebras Inference powers the world's top models - like Qwen3 Coder at 2000 tokens/s 🤯 And you can now access these models directly in Visual Studio Code with the extension and an API key (that you can get for free from

Cerebras Inference powers the world's top models - like Qwen3 Coder at 2000 tokens/s 🤯 And you can now access these models directly in Visual Studio Code with the extension and an API key (that you can get for free from

Visual Studio Code

57,360 görüntüleme • 9 ay önce

Meet Tencent HY-MT1.5 — our latest open-source translation models, featuring 1.8B and 7B models designed for seamless on-device and cloud deployment. 🌍 The 1.8B version is already trending at #1 on Hugging Face! 🥇It’s optimized for consumer hardware, delivering 0.18s latency (50 tokens) with just a 1GB memory footprint. Meanwhile, the 7B model achieves exceptional performance, surpassing many mid-sized models and rivaling the 90th percentile of Gemini 3.0 Pro. 🚀🧠 👉🏻 Try it now: 🔗 GitHub: 🤗 Hugging Face:

Meet Tencent HY-MT1.5 — our latest open-source translation models, featuring 1.8B and 7B models designed for seamless on-device and cloud deployment. 🌍 The 1.8B version is already trending at #1 on Hugging Face! 🥇It’s optimized for consumer hardware, delivering 0.18s latency (50 tokens) with just a 1GB memory footprint. Meanwhile, the 7B model achieves exceptional performance, surpassing many mid-sized models and rivaling the 90th percentile of Gemini 3.0 Pro. 🚀🧠 👉🏻 Try it now: 🔗 GitHub: 🤗 Hugging Face:

Tencent HY

17,105 görüntüleme • 6 ay önce

I have created this OS, based on Alpine Linux, & it is stable enough to use (i made the file manager, etc) It's all in C, and as you can tell on file view, the terminal/shell runs simultaneously with the windows on a grid, theyre not box characters they're vectors on the grid

I have created this OS, based on Alpine Linux, & it is stable enough to use (i made the file manager, etc) It's all in C, and as you can tell on file view, the terminal/shell runs simultaneously with the windows on a grid, theyre not box characters they're vectors on the grid

Shimazu.S

12,519 görüntüleme • 2 ay önce

Qwen3.5 runs quite well in mlx-lm. Awesome that we have a frontier-level hybrid model. The context gets longer but the inference speed and memory use barely change. Here's the Q4 generating a space invaders game on an M3 Ultra. Generated 4,120 tokens at 37.6 tok/s.

Qwen3.5 runs quite well in mlx-lm. Awesome that we have a frontier-level hybrid model. The context gets longer but the inference speed and memory use barely change. Here's the Q4 generating a space invaders game on an M3 Ultra. Generated 4,120 tokens at 37.6 tok/s.

Awni Hannun

48,673 görüntüleme • 4 ay önce

the tiebreaker is done. qwen 3.5 27B dense. single RTX 3090. one prompt. zero steering. zero human edits. 1,827 lines across 10 files. 13 minutes. full thinking mode. runs on first load. hermes 4.3 got the same prompt with 2x 3090s and 5x the context it needed. wrote 1,249 lines, left empty files, needed 3 interventions, game was broken on load. same architecture class. same quant. hermes got double the hardware. completely different result. dense wasn't the problem. hermes was. but here's what got me. this model thinks at 27 tok/s. every single token carries 27 billion parameters of reasoning. MoE hit 112 tok/s but only 3B active per token. the dense model is slower and it doesn't matter. watch 13 minutes of autonomous coding on a consumer GPU with zero intervention and tell me speed is what matters. a year ago this wasn't possible. now it runs on hardware you can buy used for $900. no API. no subscription. no cloud. just a 3090 doing what data centers did 18 months ago. full unedited session in the video. every token, every file, every thinking chain. 16 minutes. hit play.

the tiebreaker is done. qwen 3.5 27B dense. single RTX 3090. one prompt. zero steering. zero human edits. 1,827 lines across 10 files. 13 minutes. full thinking mode. runs on first load. hermes 4.3 got the same prompt with 2x 3090s and 5x the context it needed. wrote 1,249 lines, left empty files, needed 3 interventions, game was broken on load. same architecture class. same quant. hermes got double the hardware. completely different result. dense wasn't the problem. hermes was. but here's what got me. this model thinks at 27 tok/s. every single token carries 27 billion parameters of reasoning. MoE hit 112 tok/s but only 3B active per token. the dense model is slower and it doesn't matter. watch 13 minutes of autonomous coding on a consumer GPU with zero intervention and tell me speed is what matters. a year ago this wasn't possible. now it runs on hardware you can buy used for $900. no API. no subscription. no cloud. just a 3090 doing what data centers did 18 months ago. full unedited session in the video. every token, every file, every thinking chain. 16 minutes. hit play.

Sudo su

91,135 görüntüleme • 4 ay önce

This is the fastest I've seen Llama 3.3 running anywhere! Llama 3.3 70B running at 652 t/s is lightning fast. And if you want Llama 3.1, here are the speeds I was able to get: • Llama 3.1 8B: 1006 t/s • Llama 3.1 70B: 709 t/s • Llama 3.1 405B: 206 t/s (You can access all of these models for free! See the link below.) This speed is incredible, but the interesting part is what's happening behind the scenes: These models aren't running on a GPU! In this video, I'm using the SambaNova cloud to access these models. They built a custom chip (SN40L) optimized for AI workflows. A single SN40L chip can hold hundreds of models (trillions of parameters) in memory! The speed alone is a huge deal, but the big advantage is for agentic workflows running multiple specialized models. A GPU can only host a single model and switch (unload and load) to a different model if necessary. An SN40L, on the other hand, can host every model at once, making it much faster. Here is the video where you can see how fast these chips are:

This is the fastest I've seen Llama 3.3 running anywhere! Llama 3.3 70B running at 652 t/s is lightning fast. And if you want Llama 3.1, here are the speeds I was able to get: • Llama 3.1 8B: 1006 t/s • Llama 3.1 70B: 709 t/s • Llama 3.1 405B: 206 t/s (You can access all of these models for free! See the link below.) This speed is incredible, but the interesting part is what's happening behind the scenes: These models aren't running on a GPU! In this video, I'm using the SambaNova cloud to access these models. They built a custom chip (SN40L) optimized for AI workflows. A single SN40L chip can hold hundreds of models (trillions of parameters) in memory! The speed alone is a huge deal, but the big advantage is for agentic workflows running multiple specialized models. A GPU can only host a single model and switch (unload and load) to a different model if necessary. An SN40L, on the other hand, can host every model at once, making it much faster. Here is the video where you can see how fast these chips are:

Santiago

97,505 görüntüleme • 1 yıl önce

Most C Programmers don't have a clue as to how the CPU works. Here is a lecture from my course on the C Language that introduces how to visualise the CPU as a C (or general software) Programmer.... More Details are at: Being able to model the behaviour of CPU provides a solid base to reason about your code. The summary is - - The CPU is a state machine which maintains it's states in registers. - It has busses to memory for requesting data/instruction read/writes. - Interrupts to attend to events. I discuss this and more in my course on the C Language.

Most C Programmers don't have a clue as to how the CPU works. Here is a lecture from my course on the C Language that introduces how to visualise the CPU as a C (or general software) Programmer.... More Details are at: Being able to model the behaviour of CPU provides a solid base to reason about your code. The summary is - - The CPU is a state machine which maintains it's states in registers. - It has busses to memory for requesting data/instruction read/writes. - Interrupts to attend to events. I discuss this and more in my course on the C Language.

Piyush Itankar

28,019 görüntüleme • 1 yıl önce