All tags
Topic: "diffusion-models"
ChatGPT Codex, OpenAI's first cloud SWE agent
codex-1 openai-o3 codex-mini gemma-3 blip3-o qwen-2.5 marigold-iid deepseek-v3 lightlab gemini-2.0 lumina-next openai runway salesforce qwen deepseek google google-deepmind j1 software-engineering parallel-processing multimodality diffusion-models depth-estimation scaling-laws reinforcement-learning fine-tuning model-performance multi-turn-conversation reasoning audio-processing sama kevinweil omarsar0 iscienceluvr akhaliq osanseviero c_valenzuelab mervenoyann arankomatsuzaki jasonwei demishassabis philschmid swyx teortaxestex jaseweston
OpenAI launched Codex, a cloud-based software engineering agent powered by codex-1 (an optimized version of OpenAI o3) available in research preview for Pro, Enterprise, and Team ChatGPT users, featuring parallel task execution like refactoring and bug fixing. The Codex CLI was enhanced with quick sign-in and a new low-latency model, codex-mini. Gemma 3 is highlighted as the best open model runnable on a single GPU. Runway released the Gen-4 References API for style transfer in generation. Salesforce introduced BLIP3-o, a unified multimodal model family using diffusion transformers for CLIP image features. The Qwen 2.5 models (1.5B and 3B versions) were integrated into the PocketPal app with various chat templates. Marigold IID, a new state-of-the-art open-source depth estimation model, was released.
In research, DeepSeek shared insights on scaling and hardware for DeepSeek-V3. Google unveiled LightLab, a diffusion-based light source control in images. Google DeepMind's AlphaEvolve uses Gemini 2.0 to discover new math and reduce costs without reinforcement learning. Omni-R1 studied audio's role in fine-tuning audio LLMs. Qwen proposed a parallel scaling law inspired by classifier-free guidance. Salesforce released Lumina-Next on the Qwen base, outperforming Janus-Pro. A study found LLM performance degrades in multi-turn conversations due to unreliability. J1 is incentivizing LLM-as-a-Judge thinking via reinforcement learning. A new Qwen study correlates question and strategy similarity to predict reasoning strategies.
not much happened today
qwen3-14b qwen3-32b qwen3-235b phi-4-reasoning o3-mini command-a gemini-2.5-pro o4-mini olm-o2-1b o3 alibaba together-ai scaling01 microsoft deepseek cohere google epoch-ai-research inception-labs openai allenai quantization fine-tuning reinforcement-learning benchmarking video-generation diffusion-models model-performance model-evaluation model-release text-generation cline _philschmid iscienceluvr alexalbert__ _lewtun teortaxestex sarahookr reach_vb
Qwen model family released quantized versions of Qwen3 models including 14B, 32B, and 235B parameters, with promising coding capabilities in Qwen3-235B. Microsoft launched Phi-4-reasoning, a 14B parameter model distilled from OpenAI's o3-mini, emphasizing supervised fine-tuning and reinforcement learning, outperforming larger models in some benchmarks. Cohere's Command A leads SQL performance on Bird Bench. Google introduced the TRAJAN eval for video generation temporal consistency and updated the Gemini OpenAI compatibility layer. Inception Labs launched a diffusion LLM API claiming 5x speed improvements over autoregressive models. Community rankings show OpenAI's o3 model debuting strongly in web app-building tasks. Other releases include AllenAI's OLMo2 1B and additional Phi 4 variants. "Qwen3-235B shows promise for coding" and "Phi-4-reasoning tech report emphasizes SFT gains" highlight key advancements.
not much happened today
nemotron-h nvidia-eagle-2.5 gpt-4o qwen2.5-vl-72b gemini-2.5-flash gemini-2.0-pro gemini-exp-1206 gemma-3 qwen2.5-32b deepseek-r1-zero-32b uni3c seedream-3.0 adobe-dragon kimina-prover qwen2.5-72b bitnet-b1.58-2b4t nvidia deepseek hugging-face alibaba bytedance adobe transformers model-optimization multimodality long-context reinforcement-learning torch-compile image-generation diffusion-models distributional-rewards model-efficiency model-training native-quantization sampling-techniques philschmid arankomatsuzaki osanseviero iScienceLuvr akhaliq
Nemotron-H model family introduces hybrid Mamba-Transformer models with up to 3x faster inference and variants including 8B, 56B, and a compressed 47B model. Nvidia Eagle 2.5 is a frontier VLM for long-context multimodal learning, matching GPT-4o and Qwen2.5-VL-72B on long-video understanding. Gemini 2.5 Flash shows improved dynamic thinking and cost-performance, outperforming previous Gemini versions. Gemma 3 now supports torch.compile for about 60% faster inference on consumer GPUs. SRPO using Qwen2.5-32B surpasses DeepSeek-R1-Zero-32B on benchmarks with reinforcement learning only. Alibaba's Uni3C unifies 3D-enhanced camera and human motion controls for video generation. Seedream 3.0 by ByteDance is a bilingual image generation model with high-resolution outputs up to 2K. Adobe DRAGON optimizes diffusion generative models with distributional rewards. Kimina-Prover Preview is an LLM trained with reinforcement learning from Qwen2.5-72B, achieving 80.7% pass@8192 on miniF2F. BitNet b1.58 2B4T is a native 1-bit LLM with 2B parameters trained on 4 trillion tokens, matching full-precision LLM performance with better efficiency. Antidistillation sampling counters unwanted model distillation by modifying reasoning traces from frontier models.
DeepSeek's Open Source Stack
qwen-qwq-32b start character-3 gemini gemini-2.0 mercury-coder gpt-4.5 jamba-mini-1.6 gemini-2.0-flash gpt-4o-mini mistral-small-3 mistral-ocr deepseek pyspur hugging-face togethercompute hedra-labs google-deepmind deeplearningai openai ai21-labs mistral-ai fine-tuning benchmarking multimodality code-generation diffusion-models model-performance model-optimization ocr embedding-models context-windows runtime-limits _akhaliq lmarena_ai reach_vb danielhanchen _philschmid aidan_mclau vikhyatk jerryjliu0
DeepSeek's Open Source Week was summarized by PySpur, highlighting multiple interesting releases. The Qwen QwQ-32B model was fine-tuned into START, excelling in PhD-level science QA and math benchmarks. Character-3, an omnimodal AI video generation model by Hedra Labs and Together AI, enables realistic animated content creation. Google DeepMind introduced the Gemini embedding model with an 8k context window, ranking #1 on MMTEB, alongside the Gemini 2.0 Code Executor supporting Python libraries and auto-fix features. Inception Labs' Mercury Coder is a diffusion-based code generation model offering faster token processing. OpenAI released GPT-4.5, their largest model yet but with less reasoning ability than some competitors. AI21 Labs launched Jamba Mini 1.6, noted for superior output speed compared to Gemini 2.0 Flash, GPT-4o mini, and Mistral Small 3. A new dataset of 1.9M scanned pages was released for OCR benchmarking, with Mistral OCR showing competitive but not top-tier document parsing performance compared to LLM/LVM-powered methods. "Cracked engineers are all you need."
LLaDA: Large Language Diffusion Models
llada-8b llama-3-8b step-video-t2v-30b step-audio-chat-132b llama-2-7b stepfun-ai scale-ai cambridge llamaindex diffusion-models text-generation multimodality video-generation voice-processing benchmarking instruction-following model-scaling gpu-usage long-context multi-turn-dialogue arankomatsuzaki _akhaliq omarsar0 iscienceluvr gallabytes maximelabonne reach_vb
LLaDA (Large Language Diffusion Model) 8B is a breakthrough diffusion-based language model that rivals LLaMA 3 8B while training on 7x fewer tokens (2 trillion tokens) and using 0.13 million H800 GPU hours. It introduces a novel text generation approach by predicting uniformly masked tokens in a diffusion process, enabling multi-turn dialogue and instruction-following. Alongside, StepFun AI released two major models: Step-Video-T2V 30B, a text-to-video model generating up to 204 frames with high coherence and motion quality, and Step-Audio-Chat 132B, a voice-to-voice model. Additionally, challenging multimodal benchmarks like Scale AI's EnigmaEval and Cambridge's ZeroBench highlight current frontier models scoring zero, emphasizing the difficulty of these tasks. The community also noted the return of diffusion models in language modeling, a previously speculative architecture now scaled successfully.
not much happened today
deepseek-v3 llama-3-1-405b gpt-4o gpt-5 minimax-01 claude-3-haiku cosmos-nemotron-34b openai deep-learning-ai meta-ai-fair google-deepmind saama langchain nvidia mixture-of-experts coding math scaling visual-tokenizers diffusion-models inference-time-scaling retrieval-augmented-generation ai-export-restrictions security-vulnerabilities prompt-injection gpu-optimization fine-tuning personalized-medicine clinical-trials ai-agents persistent-memory akhaliq
DeepSeek-V3, a 671 billion parameter mixture-of-experts model, surpasses Llama 3.1 405B and GPT-4o in coding and math benchmarks. OpenAI announced the upcoming release of GPT-5 on April 27, 2023. MiniMax-01 Coder mode in ai-gradio enables building a chess game in one shot. Meta research highlights trade-offs in scaling visual tokenizers. Google DeepMind improves diffusion model quality via inference-time scaling. The RA-DIT method fine-tunes LLMs and retrievers for better RAG responses. The U.S. proposes a three-tier export restriction system on AI chips and models, excluding countries like China and Russia. Security vulnerabilities in AI chatbots involving CSRF and prompt injection were revealed. Concerns about superintelligence and weapons-grade AI models were expressed. ai-gradio updates include NVIDIA NIM compatibility and new models like cosmos-nemotron-34b. LangChain integrates with Claude-3-haiku for AI agents with persistent memory. Triton Warp specialization optimizes GPU usage for matrix multiplication. Meta's fine-tuned Llama models, OpenBioLLM-8B and OpenBioLLM-70B, target personalized medicine and clinical trials.
Titans: Learning to Memorize at Test Time
minimax-01 gpt-4o claude-3.5-sonnet internlm3-8b-instruct transformer2 google meta-ai-fair openai anthropic langchain long-context mixture-of-experts self-adaptive-models prompt-injection agent-authentication diffusion-models zero-trust-architecture continuous-adaptation vision agentic-systems omarsar0 hwchase17 abacaj hardmaru rez0__ bindureddy akhaliq saranormous
Google released a new paper on "Neural Memory" integrating persistent memory directly into transformer architectures at test time, showing promising long-context utilization. MiniMax-01 by @omarsar0 features a 4 million token context window with 456B parameters and 32 experts, outperforming GPT-4o and Claude-3.5-Sonnet. InternLM3-8B-Instruct is an open-source model trained on 4 trillion tokens with state-of-the-art results. Transformer² introduces self-adaptive LLMs that dynamically adjust weights for continuous adaptation. Advances in AI security highlight the need for agent authentication, prompt injection defenses, and zero-trust architectures. Tools like Micro Diffusion enable budget-friendly diffusion model training, while LeagueGraph and Agent Recipes support open-source social media agents.
Moondream 2025.1.9: Structured Text, Enhanced OCR, Gaze Detection in a 2B Model
o1 vdr-2b-multi-v1 llava-mini openai llamaindex langchainai qdrant genmoai vision model-efficiency structured-output gaze-detection reasoning model-distillation multimodality embedding-models gan diffusion-models self-attention training-optimizations development-frameworks api cross-language-deployment semantic-search agentic-document-processing developer-experience philschmid saranormous jxmnop reach_vb iscienceluvr multimodalart arohan adcock_brett awnihannun russelljkaplan ajayj_
Moondream has released a new version that advances VRAM efficiency and adds structured output and gaze detection, marking a new frontier in vision model practicality. Discussions on Twitter highlighted advancements in reasoning models like OpenAI's o1, model distillation techniques, and new multimodal embedding models such as vdr-2b-multi-v1 and LLaVA-Mini, which significantly reduce computational costs. Research on GANs and decentralized diffusion models showed improved stability and performance. Development tools like MLX and vLLM received updates for better portability and developer experience, while frameworks like LangChain and Qdrant enable intelligent data workflows. Company updates include new roles and team expansions at GenmoAI. "Efficiency tricks are all you need."
Stripe lets Agents spend money with StripeAgentToolkit
gpt-4o gemini-exp-1114 stripe openai anthropic meta-ai-fair ai-computer-interfaces agentic-ai model-overfitting benchmarks scaling-laws agi chain-of-thought image-captioning dialogue-systems memory-efficient-fine-tuning diffusion-models mixture-of-experts adaptive-decoding creativity-optimization factuality-optimization pair-programming document-parsing retrieval-augmented-generation abacaj francois-fleuret lmarena_ai goodside jxmnop jaseweston stevenheidel
Stripe has pioneered an AI SDK specifically designed for agents that handle payments, integrating with models like gpt-4o to enable financial transactions and token-based charging. The AI developer tooling trend emphasizes better "AI-Computer Interfaces" for improved agent reliability, with tools like E2B and the
llms.txt documentation trend gaining traction, notably adopted by Anthropic. In AI model news, Gemini-Exp-1114 topped the Vision Leaderboard and improved in Math Arena, while discussions continue around model overfitting and the limits of scaling laws for AGI. OpenAI released a ChatGPT desktop app for macOS with integrations for VS Code, Xcode, and Terminal, enhancing developer workflows and pair programming. Anthropic introduced a prompt improver using chain-of-thought reasoning, and Meta AI shared top research from EMNLP2024 on image captioning, dialogue systems, and memory-efficient fine-tuning. Highlights from ICLR 2025 include diffusion-based illumination harmonization, open mixture-of-experts language models, and hyperbolic vision-language models. A new adaptive decoding method optimizes creativity and factuality per token. Tools like LlamaParse and RAGformation were also introduced for document parsing and retrieval-augmented generation. s{imple|table|calable} Consistency Models
llama-3-70b llama-3-405b llama-3-1 stable-diffusion-3.5 gpt-4 stability-ai tesla cerebras cohere langchain model-distillation diffusion-models continuous-time-consistency-models image-generation ai-hardware inference-speed multilingual-models yang-song
Model distillation significantly accelerates diffusion models, enabling near real-time image generation with only 1-4 sampling steps, as seen in BlinkShot and Flux Schnell. Research led by Yang Song introduced simplified continuous-time consistency models (sCMs), achieving under 10% FID difference in just 2 steps and scaling up to 1.5B parameters for higher quality. On AI hardware, Tesla is deploying a 50k H100 cluster potentially capable of completing GPT-4 training in under three weeks, while Cerebras Systems set a new inference speed record on Llama 3.1 70B with their wafer-scale AI chips. Stability AI released Stable Diffusion 3.5 and its Turbo variant, and Cohere launched new multilingual models supporting 23 languages with state-of-the-art performance. LangChain also announced ecosystem updates.
a calm before the storm
o1 o1-mini qwen2.5 gpt-4 llama-2-70b llama-7b anthropic openai alibaba microsoft blackrock groq aramco disney eth-zurich pudu-robotics slack long-context kv-cache-quantization diffusion-models reinforcement-learning robotics ai-integration multilinguality model-benchmarking model-performance model-optimization adcock_brett philschmid rohanpaul_ai jvnixon kateclarktweets sama
Anthropic is raising funds at a valuation up to $40 billion ahead of anticipated major releases. OpenAI launched new reasoning models o1 and o1-mini, with increased rate limits and a multilingual MMLU benchmark. Alibaba released the open-source Qwen2.5 model supporting 29+ languages, showing competitive performance to gpt-4 at lower cost. Microsoft and Blackrock plan to invest $30 billion in AI data centers, with Groq partnering with Aramco to build the world's largest AI inference center. Robotics advances include Disney Research and ETH Zurich's diffusion-based motion generation for robots and Pudu Robotics' semi-humanoid robot. Slack and Microsoft introduced AI-powered agents integrated into their platforms. Research highlights include long-context scaling for llama-2-70b using Dual Chunk Attention and KV cache quantization enabling 1 million token context on llama-7b models.
not much happened today
llama-3 o1 deepseek-2.5 gpt-4 claude-3.5-sonnet 3dtopia-xl cogvideox anthropic meta-ai-fair openai deepseek-ai llamaindex langchainai retrieval-augmented-generation prompt-caching multimodality multi-agent-systems reasoning diffusion-models image-to-video prompting enterprise-ai agentic-ai long-context model-evaluation caching model-cost-efficiency
Anthropic introduced a RAG technique called Contextual Retrieval that reduces retrieval failure rates by 67% using prompt caching. Meta is teasing multimodal Llama 3 ahead of Meta Connect. OpenAI is hiring for a multi-agent research team focusing on improved AI reasoning with their o1 models, which have sparked mixed reactions. DeepSeek 2.5 is noted as a cost-effective alternative to GPT-4 and Claude 3.5 sonnet. New models like 3DTopia-XL for 3D asset generation and CogVideoX for image-to-video conversion were highlighted. Techniques to boost reasoning by re-reading questions and combining retrieval with prompt caching were shared. Industry insights emphasize the necessity of AI adoption in enterprises and the disruption of traditional ML businesses. Tools like LangChainAI's LangGraph Templates and LlamaIndex's LlamaParse Premium enhance agentic applications and multimodal content extraction. Discussions on LLM evals and caching highlight production challenges and improvements. "Companies not allowing developers to use AI are unlikely to succeed" was a key sentiment.
a quiet weekend
o1 datagemma aloha demostart firefly-ai-video-model pixtral-12b gamegen-o openai google-deepmind adobe mistral-ai tencent supermaven 11x cohere anthropic latent-space-university stanford microsoft mila notre-dame reinforcement-learning chain-of-thought reasoning robotics diffusion-models multimodality video-generation model-training reflection-tuning mathematical-reasoning model-benchmarking fine-tuning george-hotz terence-tao adcock_brett rohanpaul_ai bindureddy fchollet philschmid
OpenAI released the new o1 model, leveraging reinforcement learning and chain-of-thought prompting to excel in reasoning benchmarks, achieving an IQ-like score of 120. Google DeepMind introduced DataGemma to reduce hallucinations by connecting LLMs with real-world data, and unveiled ALOHA and DemoStart for robot dexterity using diffusion methods. Adobe previewed its Firefly AI Video Model with text-to-video and generative extend features. Mistral launched the multimodal Pixtral 12B model, and Tencent presented the GameGen-O open-world video game generation model. Several research papers from Stanford, OpenAI, Microsoft, Mila, and Notre Dame focus on advanced reasoning, self-verification, and reflection tuning techniques. Experts like Terence Tao and George Hotz have shared mixed but optimistic views on o1's capabilities. Seed funding rounds include Supermaven ($12M) and 11x ($24M).
Summer of Code AI: $1.6b raised, 1 usable product
ltm-2 llama-3-1-405b gemini-advanced cognition poolside codeium magic google-deepmind nvidia google-cloud long-context model-efficiency custom-hardware cuda training-stack gpu-scaling neural-world-models diffusion-models quantization nat-friedman ben-chess rohan-paul
Code + AI is emphasized as a key modality in AI engineering, highlighting productivity and verifiability benefits. Recent major funding rounds include Cognition AI raising $175M, Poolside raising $400M, Codeium AI raising $150M, and Magic raising $320M. Magic announced their LTM-2 model with a 100 million token context window, boasting efficiency improvements over Llama 3.1 405B by about 1000x cheaper in sequence-dimension algorithm and drastically lower memory requirements. Magic's stack is built from scratch with custom CUDA and no open-source foundations, partnered with Google Cloud and powered by NVIDIA H100 and GB200 GPUs, aiming to scale to tens of thousands of GPUs. Google DeepMind revealed updates to Gemini Advanced with customizable expert "Gems." Neural Game Engines like GameNGen can run DOOM in a diffusion model trained on 0.9B frames. The content also references LLM quantization research by Rohan Paul.
not much happened today
gpt-4o claude-3.5-sonnet phi-3.5-mini phi-3.5-moe phi-3.5-vision llama-3-1-405b qwen2-math-72b openai anthropic microsoft meta-ai-fair hugging-face langchain box fine-tuning benchmarking model-comparison model-performance diffusion-models reinforcement-learning zero-shot-learning math model-efficiency ai-regulation ai-safety ai-engineering prompt-engineering swyx ylecun
OpenAI launched GPT-4o finetuning with a case study on Cosine. Anthropic released Claude 3.5 Sonnet with 8k token output. Microsoft Phi team introduced Phi-3.5 in three variants: Mini (3.8B), MoE (16x3.8B), and Vision (4.2B), noted for sample efficiency. Meta released Llama 3.1 405B, deployable on Google Cloud Vertex AI, offering GPT-4 level capabilities. Qwen2-Math-72B achieved state-of-the-art math benchmark performance with a Gradio demo. Discussions included model comparisons like ViT vs CNN and Mamba architecture. Tools updates featured DSPy roadmap, Flux Schnell improving diffusion speed on M1 Max, and LangChain community events. Research highlights zero-shot DUP prompting for math reasoning and fine-tuning best practices. AI ethics covered California's AI Safety Bill SB 1047 and regulatory concerns from Yann LeCun. Commentary on AI engineer roles by Swyx. "Chat with PDF" feature now available for Box Enterprise Plus users.
Lilian Weng on Video Diffusion
wizardlm-2 llama-3 reka-core devin opus sora openai adobe reka-ai diffusion-models video-generation training-free-adaptation multimodality intuition creativity analogy-recognition self-improving-ai model-recognition agi-timelines model-performance startup-competition lilian-weng sam-altman geoffrey-hinton yann-lecun
OpenAI expands with a launch in Japan, introduces a Batch API, and partners with Adobe to bring the Sora video model to Premiere Pro. Reka AI releases the Reka Core multimodal language model. WizardLM-2 is released showing impressive performance, and Llama 3 news is anticipated soon. Geoffrey Hinton highlights AI models exhibiting intuition, creativity, and analogy recognition beyond humans. The Devin AI model notably contributes to its own codebase. Opus demonstrates the ability to recognize its own generated outputs. Sam Altman warns startups about being steamrolled by OpenAI if they don't adapt quickly. Yann LeCun discusses AGI timelines, emphasizing it is inevitable but not imminent or solely from LLMs. Lilian Weng's blog on diffusion models for video generation highlights training-free adaptation as a breakthrough technique.
Gemini Pro and GPT4T Vision go GA on the same day by complete coincidence
gemini-1.5-pro gpt-4-turbo llama-3 orca-2.5-7b functionary-v2.4 cosxl google openai meta-ai-fair hugging-face cohere million-token-context-window audio-processing file-api text-embedding function-calling reasoning direct-nash-optimization contrastive-learning code-interpreter diffusion-models neural-odes inference-speed multilingual-dataset image-editing no-code-development
At Google Cloud Next, Gemini 1.5 Pro was released with a million-token context window, available in 180+ countries, featuring 9.5 hours of audio understanding, a new File API for nearly unlimited free uploads, and the Gecko-1b-256/768 embedding model. GPT-4 Turbo with Vision became generally available in the API with a major update improving reasoning capabilities. Meta Platforms plans to launch smaller versions of Llama 3 next week. The Orca 2.5 7B model using Direct Nash Optimization outperforms older GPT-4 versions in AlpacaEval. New releases include Functionary-V2.4 with enhanced function calling and code interpretation, and CosXL models for image editing. Research highlights include continuous U-Nets for diffusion models achieving up to 80% faster inference and a massive multilingual dataset with ~5.6 trillion word tokens. Creative applications include a no-code touch screen game made with Gemini 1.5 and AI-generated novel trailers.
Stable Diffusion 3 — Rombach & Esser did it again!
stable-diffusion-3 claude-3 orca dolphincoder-starcoder2-15b stability-ai anthropic microsoft latitude perplexity-ai llamaindex tripo-ai diffusion-models multimodality benchmarking human-evaluation text-generation image-generation 3d-modeling fine-tuning roleplay coding dataset-release soumith-chintala bill-peebles swyx kevinafischer jeremyphoward akhaliq karinanguyen_ aravsrinivas
Over 2500 new community members joined following Soumith Chintala's shoutout, highlighting growing interest in SOTA LLM-based summarization. The major highlight is the detailed paper release of Stable Diffusion 3 (SD3), showcasing advanced text-in-image control and complex prompt handling, with the model outperforming other SOTA image generation models in human-evaluated benchmarks. The SD3 model is based on an enhanced Diffusion Transformer architecture called MMDiT. Meanwhile, Anthropic released Claude 3 models, noted for human-like responses and emotional depth, scoring 79.88% on HumanEval but costing over twice as much as GPT-4. Microsoft launched new Orca-based models and datasets, and Latitude released DolphinCoder-StarCoder2-15b with strong coding capabilities. Integration of image models by Perplexity AI and 3D CAD generation by PolySpectra powered by LlamaIndex were also highlighted. "SD3's win rate beats all other SOTA image gen models (except perhaps Ideogram)" and "Claude 3 models are very good at generating d3 visualizations from text descriptions."
RIP Latent Diffusion, Hello Hourglass Diffusion
gpt-4 latent-diffusion stable-diffusion meta-ai-fair openai hugging-face diffusion-models transformers image-generation model-efficiency fine-tuning quantization prompt-engineering roleplay training-optimization katherine-crowson lucidrains
Katherine Crowson from Stable Diffusion introduces a hierarchical pure transformer backbone for diffusion-based image generation that efficiently scales to megapixel resolutions with under 600 million parameters, improving upon the original ~900M parameter model. This architecture processes local and global image phenomena separately, enhancing efficiency and resolution without latent steps. Additionally, Meta's Self Rewarding LM paper has inspired lucidrains to begin an implementation. Discord summaries highlight GPT-4's robustness against quantification tricks, discussions on open-source GPT-0 alternatives, challenges in DPO training on limited VRAM with suggestions like QLoRA and rmsprop, and efforts to improve roleplay model consistency through fine-tuning and merging. Philosophical debates on AI sentience and GPT-4 customization for markdown and translation tasks were also noted.