All tags
Topic: "tokenization"
Gemini 2.0 Flash GA, with new Flash Lite, 2.0 Pro, and Flash Thinking
gemini-2.0-flash gemini-2.0-flash-lite gemini-2.0-pro-experimental gemini-1.5-pro deepseek-r1 gpt-2 llama-3-1 google-deepmind hugging-face anthropic multimodality context-windows cost-efficiency pretraining fine-tuning reinforcement-learning transformer tokenization embeddings mixture-of-experts andrej-karpathy jayalammar maartengr andrewyng nearcyan
Google DeepMind officially launched Gemini 2.0 models including Flash, Flash-Lite, and Pro Experimental, with Gemini 2.0 Flash outperforming Gemini 1.5 Pro while being 12x cheaper and supporting multimodal input and a 1 million token context window. Andrej Karpathy released a 3h31m video deep dive into large language models, covering pretraining, fine-tuning, and reinforcement learning with examples like GPT-2 and Llama 3.1. A free course on Transformer architecture was introduced by Jay Alammar, Maarten Gr, and Andrew Ng, focusing on tokenizers, embeddings, and mixture-of-expert models. DeepSeek-R1 reached 1.2 million downloads on Hugging Face with a detailed 36-page technical report. Anthropic increased rewards to $10K and $20K for their jailbreak challenge, while BlueRaven extension was updated to hide Twitter metrics for unbiased engagement.
Meta BLT: Tokenizer-free, Byte-level LLM
byte-latent-transformer llama-3 phi-4 gpt-4o command-r7b meta-ai-fair llamaindex microsoft deepseek-ai openai cohere anthropic tokenization transformer-architecture model-efficiency benchmarking multimodality vision reinforcement-learning model-scaling jailbreaking model-optimization
Meta AI introduces the Byte Latent Transformer (BLT), a tokenizer-free architecture that dynamically forms byte patches for efficient compute allocation, outperforming Llama 3 on benchmarks including the CUTE benchmark. The model was trained on approximately 1 trillion tokens and features a three-block transformer design with local and global components. This approach challenges traditional tokenization and may enable new multimodal capabilities such as direct file interaction without retrieval-augmented generation. Additionally, Microsoft announced the Phi-4 14B parameter model achieving state-of-the-art results on STEM and reasoning benchmarks, surpassing GPT-4o. DeepSeek AI launched new vision-language models based on their MoE architecture with sizes ranging from 1.0B to 27B parameters. OpenAI released a new Projects feature for ChatGPT, and Cohere introduced their smallest and fastest Command R7B model. Anthropic published research on "Best-of-N Jailbreaking" vulnerabilities across text, vision, and audio models. Industry discussion highlights a trend of decreasing frontier LLM sizes, with GPT-4 at approximately 1.8 trillion parameters compared to newer models.
Llama 3.2: On-device 1B/3B, and Multimodal 11B/90B (with AI2 Molmo kicker)
llama-3-2 llama-3-1 claude-3-haiku gpt-4o-mini molmo-72b molmo-7b gemma-2 phi-3-5 llama-3-2-vision llama-3-2-3b llama-3-2-20b meta-ai-fair ai2 qualcomm mediatek arm ollama together-ai fireworks-ai weights-biases cohere weaviate multimodality vision context-windows quantization model-release tokenization model-performance model-optimization rag model-training instruction-following mira-murati daniel-han
Meta released Llama 3.2 with new multimodal versions including 3B and 20B vision adapters on a frozen Llama 3.1, showing competitive performance against Claude Haiku and GPT-4o-mini. AI2 launched multimodal Molmo 72B and 7B models outperforming Llama 3.2 in vision tasks. Meta also introduced new 128k-context 1B and 3B models competing with Gemma 2 and Phi 3.5, with collaborations hinted with Qualcomm, Mediatek, and Arm for on-device AI. The release includes a 9 trillion token count for Llama 1B and 3B. Partner launches include Ollama, Together AI offering free 11B model access, and Fireworks AI. Additionally, a new RAG++ course from Weights & Biases, Cohere, and Weaviate offers systematic evaluation and deployment guidance for retrieval-augmented generation systems based on extensive production experience.
Grok 2! and ChatGPT-4o-latest confuses everybody
gpt-4o grok-2 claude-3.5-sonnet flux-1 stable-diffusion-3 gemini-advanced openai x-ai black-forest-labs google-deepmind benchmarking model-performance tokenization security-vulnerabilities multi-agent-systems research-automation text-to-image conversational-ai model-integration ylecun rohanpaul_ai karpathy
OpenAI quietly released a new GPT-4o model in ChatGPT, distinct from the API version, reclaiming the #1 spot on Lmsys arena benchmarks across multiple categories including math, coding, and instruction-following. Meanwhile, X.ai launched Grok 2, outperforming Claude 3.5 Sonnet and previous GPT-4o versions, with plans for enterprise API release. Grok 2 integrates Black Forest Labs' Flux.1, an open-source text-to-image model surpassing Stable Diffusion 3. Google DeepMind announced Gemini Advanced with enhanced conversational features and Pixel device integration. AI researcher ylecun highlighted LLM limitations in learning and creativity, while rohanpaul_ai discussed an AI Scientist system generating publishable ML research at low cost. karpathy warned of security risks in LLM tokenizers akin to SQL injection.
Chameleon: Meta's (unreleased) GPT4o-like Omnimodal Model
chameleon gpt-4o gemini-1.5-flash claude-3 meta-ai-fair openai google-deepmind anthropic reddit multimodality early-fusion benchmarking model-training tokenization streaming tool-use vision coding hallucination-detection model-performance armen-aghajanyan sama alexandr-wang abacaj alexalbert__
Meta AI FAIR introduced Chameleon, a new multimodal model family with 7B and 34B parameter versions trained on 10T tokens of interleaved text and image data enabling "early fusion" multimodality that can natively output any modality. While reasoning benchmarks are modest, its "omnimodality" approach competes well with pre-GPT4o multimodal models. OpenAI launched GPT-4o, a model excelling in benchmarks like MMLU and coding tasks, with strong multimodal capabilities but some regression in ELO scores and hallucination issues. Google DeepMind announced Gemini 1.5 Flash, a small model with 1M context window and flash performance, highlighting convergence trends between OpenAI and Google models. Anthropic updated Claude 3 with streaming support, forced tool use, and vision tool integration for multimodal knowledge extraction. OpenAI also partnered with Reddit, raising industry attention.
Google I/O in 60 seconds
gemini-1.5-pro gemini-flash gemini-ultra gemini-pro gemini-nano gemma-2 llama-3-70b paligemma imagen-3 veo google google-deepmind youtube tokenization model-performance fine-tuning vision multimodality model-release model-training model-optimization ai-integration image-generation watermarking hardware-optimization voice video-understanding
Google announced updates to the Gemini model family, including Gemini 1.5 Pro with 2 million token support, and the new Gemini Flash model optimized for speed with 1 million token capacity. The Gemini suite now includes Ultra, Pro, Flash, and Nano models, with Gemini Nano integrated into Chrome 126. Additional Gemini features include Gemini Gems (custom GPTs), Gemini Live for voice conversations, and Project Astra, a live video understanding assistant. The Gemma model family was updated with Gemma 2 at 27B parameters, offering near-llama-3-70b performance at half the size, plus PaliGemma, a vision-language open model inspired by PaLI-3. Other launches include DeepMind's Veo, Imagen 3 for photorealistic image generation, and a Music AI Sandbox collaboration with YouTube. SynthID watermarking now extends to text, images, audio, and video. The Trillium TPUv6 codename was revealed. Google also integrated AI across its product suite including Workspace, Email, Docs, Sheets, Photos, Search, and Lens. "The world awaits Apple's answer."
GPT-4o: the new SOTA-EVERYTHING Frontier model (GPT4O version)
gpt-4o gpt-4-turbo openai lmsys multion adept multimodality vision speech-recognition tokenization real-time-processing coding model-performance model-optimization desktop-agents sama gdb
OpenAI has released GPT-4o, a new multimodal model capable of reasoning across text, audio, and video in real time with low latency (~300ms). It features voice and vision capabilities, improved non-English language performance with an expanded 200k vocabulary tokenizer, and is available to all ChatGPT users including free plans. GPT-4o is half the price and twice as fast as GPT-4-turbo with 5x rate limits. The model supports real-time voice and video input/output and shows strong coding capabilities. The release includes a new desktop app that can read screen and clipboard history, challenging existing desktop agent startups. The announcement was accompanied by demos including image generation and 3D object handling, with OpenAI achieving state-of-the-art performance in ASR and vision tasks. The update was widely discussed on social media, with comparisons to GPT-4T highlighting GPT-4o's speed and versatility. "GPT-4o is smart, fast, natively multimodal, and a step towards more natural human-computer interaction" and "extremely versatile and fun to play with".
Llama-3-70b is GPT-4-level Open Model
llama-3-70b llama-3-8b llama-3 llama-2-70b mistral-7b grok-3 stable-diffusion-3 vasa-1 meta-ai-fair groq nvidia amazon microsoft benchmarking model-performance fine-tuning function-calling arithmetic image-generation video-generation energy-usage gpu-demand political-bias ai-safety scaling context-windows tokenization elon-musk
Meta has released Llama 3, their most capable open large language model with 8B and 70B parameter versions supporting 8K context length and outperforming previous models including Llama 2 and Mistral 7B. Groq serves the Llama 3 70B model at 500-800 tokens/second, making it the fastest GPT-4-level token source. Discussions highlight AI scaling challenges with Elon Musk stating that training Grok 3 will require 100,000 Nvidia H100 GPUs, and AWS planning to acquire 20,000 B200 GPUs for a 27 trillion parameter model. Microsoft unveiled VASA-1 for lifelike talking face generation, while Stable Diffusion 3 and its extensions received mixed impressions. Concerns about AI energy usage and political bias in AI were also discussed.
Meta Llama 3 (8B, 70B)
llama-3-8b llama-3-70b llama-3-400b stable-diffusion-3 mixtral-8x22b-instruct-v0.1 vasa-1 meta-ai-fair stability-ai boston-dynamics microsoft mistral-ai hugging-face transformer tokenization model-training benchmarking robotics natural-language-processing real-time-processing synthetic-data dataset-cleaning behavior-trees ai-safety model-accuracy api model-release humor helen-toner
Meta partially released Llama 3 models including 8B and 70B variants, with a 400B variant still in training, touted as the first GPT-4 level open-source model. Stability AI launched Stable Diffusion 3 API with model weights coming soon, showing competitive realism against Midjourney V6. Boston Dynamics unveiled an electric humanoid robot Atlas, and Microsoft introduced the VASA-1 model generating lifelike talking faces at 40fps on RTX 4090. Mistral AI, a European OpenAI rival, is seeking $5B funding with its Mixtral-8x22B-Instruct-v0.1 model achieving 100% accuracy on 64K context benchmarks. AI safety discussions include calls from former OpenAI board member Helen Toner for audits of top AI companies, and the Mormon Church released AI usage principles. New AI development tools include Ctrl-Adapter for diffusion models, Distilabel 1.0.0 for synthetic dataset pipelines, Data Bonsai for data cleaning with LLMs, and Dendron for building LLM agents with behavior trees. Memes highlight AI development humor and cultural references. The release of Llama 3 models features improved reasoning, a 128K token vocabulary, 8K token sequences, and grouped query attention.
DBRX: Best open model (just not most efficient)
dbrx grok mixtral llama-2 mpt-7b gpt-4 databricks hugging-face mistral-ai mosaicml openai mixture-of-experts model-efficiency tokenization model-training code-generation model-architecture open-source-models benchmarking fine-tuning
Databricks Mosaic has released a new open-source model called DBRX that outperforms Grok, Mixtral, and Llama2 on evaluations while being about 2x more efficient than Llama2 and Grok. The model was trained on 12 trillion tokens using 3,000 H100 GPUs over 2 months, with an estimated compute cost of $10 million. It uses OpenAI's 100k tiktoken tokenizer and shows strong zero-shot code generation performance, even beating GPT-4 on the Humaneval benchmark. DBRX also upstreamed work to MegaBlocks open source. Despite its scale and efficiency, DBRX's performance on MMLU is only slightly better than Mixtral, raising questions about its scaling efficiency. The focus of DBRX is on enabling users to train models efficiently, with MoE training being about 2x more FLOP-efficient than dense models, achieving similar quality with nearly 4x less compute than previous MPT models. This release is part of the ongoing competition for open-source AI leadership, including models like Dolly, MPT, and Mistral. "If it activates 36B params, the model's perf should be equivalent to a 72B dense model or even 80B," says Qwen's tech lead.
Mistral Large disappoints
mistral-large mistral-small mixtral-8x7b gpt-4-turbo dreamgen-opus-v1 mistral-ai openai hugging-face benchmarking model-merging fine-tuning reinforcement-learning model-training tokenization model-optimization ai-assisted-decompilation performance cost-efficiency deception roleplay deep-speed dpo timotheeee1 cogbuji plasmator jsarnecki maldevide spottyluck mrjackspade
Mistral announced Mistral Large, a new language model achieving 81.2% accuracy on MMLU, trailing GPT-4 Turbo by about 5 percentage points on benchmarks. The community reception has been mixed, with skepticism about open sourcing and claims that Mistral Small outperforms the open Mixtral 8x7B. Discussions in the TheBloke Discord highlighted performance and cost-efficiency comparisons between Mistral Large and GPT-4 Turbo, technical challenges with DeepSpeed and DPOTrainer for training, advances in AI deception for roleplay characters using DreamGen Opus V1, and complexities in model merging using linear interpolation and PEFT methods. Enthusiasm for AI-assisted decompilation was also expressed, emphasizing the use of open-source projects for training data.
Karpathy emerges from stealth?
mistral-7b mixtral-8x7b zephyr-7b gpt-4 llama-2 intel mistral-ai audiogen thebloke tokenization quantization model-optimization fine-tuning model-merging computational-efficiency memory-optimization retrieval-augmented-generation multi-model-learning meta-reasoning dataset-sharing open-source ethical-ai community-collaboration andrej-karpathy
Andrej Karpathy released a comprehensive 2-hour tutorial on tokenization, detailing techniques up to GPT-4's tokenizer and noting the complexity of Llama 2 tokenization with SentencePiece. Discussions in AI Discord communities covered model optimization and efficiency, focusing on quantization of models like Mistral 7B and Zephyr-7B to reduce memory usage for consumer GPUs, including Intel's new weight-only quantization algorithm. Efforts to improve computational efficiency included selective augmentation reducing costs by 57.76% and memory token usage versus kNN for Transformers. Challenges in hardware compatibility and software issues were shared, alongside fine-tuning techniques such as LoRA and model merging. Innovative applications of LLMs in retrieval-augmented generation (RAG), multi-model learning, and meta-reasoning were explored. The community emphasized dataset sharing, open-source releases like SDXL VAE encoded datasets and Audiogen AI codecs, and ethical AI use with censorship and guardrails. Collaboration and resource sharing remain strong in these AI communities.
The Core Skills of AI Engineering
miqumaid olmo aphrodite awq exl2 mistral-medium internlm ssd-1b lora qlora loftq ai2 hugging-face ai-engineering quantization fine-tuning open-source model-deployment data-quality tokenization prompt-adherence distillation ai-security batching hardware role-playing eugene-yan
AI Discords for 2/2/2024 analyzed 21 guilds, 312 channels, and 4782 messages saving an estimated 382 minutes of reading time. Discussions included Eugene Yan initiating a deep dive into AI engineering challenges, highlighting overlaps between software engineering and data science skills. The TheBloke Discord featured talks on MiquMaid, OLMo (an open-source 65B LLM by AI2 under Apache 2.0), Aphrodite model batching, AWQ quantization, and LoRA fine-tuning techniques like QLoRA and LoftQ. The LAION Discord discussed SSD-1B distillation issues, data quality optimization with captioning datasets like BLIP, COCO, and LLaVA, and tokenization strategies for prompt adherence in image generation. Other topics included AI security with watermarking, superconductors and carbon nanotubes for hardware, and deployment of LLMs via Hugging Face tools.
1/13-14/2024: Don't sleep on #prompt-engineering
The OpenAI Discord community engaged in diverse discussions including prompt engineering techniques like contrastive Chain of Thought and step back prompting, and explored model merging and mixture-of-experts (MoE) concepts. Philosophical debates on AI consciousness and the ethics of AI-generated voices highlighted concerns about AI sentience and copyright issues. Technical clarifications were made on hyperdimensional vector space models used in modern AI embeddings. Users also discussed customizing GPT with personality profiles and prompt personalization to overcome token limits, and proposed a universal translator feature for multilingual Discord interactions. Key contributors included longtime regular MadameArchitect and community members such as @darthgustav and @metaldrgn.
12/28/2023: Smol Talk updates
tinyllama-1.1b mixtral tinygpt-v nous-research tyrannosaurus latex benchmarking knowledge-graphs model-finetuning tokenization decentralized-computation philosophy-of-ai multimodality vision open-source-models gary-marcus
Nous Research AI Discord discussions covered topics such as AI placement charts, ChatGPT's issues with Latex math format compatibility with Obsidian, and performance metrics of the TinyLlama 1.1B model on various benchmarks. Users shared resources including the math-centric corpus MathPile, knowledge graph building methods, and open-source large language model repositories. Technical discussions included decentralized computation feasibility for models like Mixtral, philosophical debates on AI sentience, and strategies for model finetuning and token counting. The community also discussed the Obsidian model, vision model training, and the release of the multimodal TinyGPT-V model by Tyrannosaurus. "ChatGPT not generating Latex math format compatible with Obsidian" and "optimistic about human-level AI within our lifetime" were notable quotes.