Self-play-pretraining-with-zero-data

Navigation

• / research
  • / { ideas, talks, papers }
• / tech
  • / { dl, talks, misc }

• fedora (138)
• linux (89)
• tensorflow (66)
• llm (45)
• dnn (43)
• nvidia (35)
• nlp (31)
• transformers (31)
• python (27)
• pytorch (26)
• paper (18)
• fundamentals (17)
• theano (16)
• cnn (15)
• gcp (15)
• gemini (12)
• mythtv (12)
• neurips (11)
• reasoning (11)
• philosophy (10)
• papertakeaway (10)
• word-embeddings (9)
• reinforcementlearning (9)
• android (9)
• opencl (9)
• workshop (8)
• llm-agents (8)
• hauppauge-pvr350 (8)
• keras (7)
• bert (7)
• git (6)
• google (6)
• deeplearning (6)
• colab (6)
• llms (6)
• transferlearning (6)
• tpu (6)
• back-to-basics (6)
• minecraft (6)
• lighttpd (6)
• encfs (6)
• scite (6)
• nginx (6)
• gpu (5)
• amd (5)
• metalearning (5)
• javascript (5)
• nodejs (5)
• imagenet (5)
• jax (5)
• gemma (5)
• truecrypt (5)
• yum (5)
• mysql (5)
• bumblebee (5)
• pylons (5)
• ner (4)
• blocks (4)
• rnn (4)
• cryptic-crosswords (4)
• cuda (4)
• rl (4)
• lasagne (4)
• virtualbox (4)
• oneshotlearning (4)
• reptile (4)
• aiday (4)
• tts (4)
• voyager (4)
• mobile (4)
• darwingodelmachines (4)
• samba (4)
• dual-screen (4)
• webmin (4)
• relation-networks (3)
• relationships (3)
• distillation (3)
• multi-hop (3)
• diffusion (3)
• stablediffusion (3)
• dall-e (3)
• rag (3)
• agents (3)
• mnist (3)
• docker (3)
• minerl (3)
• finetuning (3)
• dino (3)
• robots (3)
• sginnovate (3)
• speechrecognition (3)
• alphago (3)
• mlp (3)
• tensorflowjs (3)
• gpt2 (3)
• tflite (3)
• hyperparameters (3)
• gdsc (3)
• langchain (3)
• promptbreeder (3)
• dspy (3)
• kaggle (3)
• sprint (3)
• alphaevolve (3)
• gpu-kernel-scientist (3)
• openclaw (3)
• ai-studio (3)
• qwen (3)
• 64bit (3)
• uk (3)
• lirc (3)
• twiki (3)
• music (3)
• openvz (3)
• scss (3)
• flask (3)
• uwsgi (3)
• opencv (3)
• radeon-hd5570 (3)
• radeon (3)
• acer (3)
• optimus (3)
• matplotlib (3)
• intel (3)
• xfce (3)
• sparse (2)
• communities (2)
• unsupervised (2)
• clip (2)
• formalisation (2)
• narratives (2)
• concepts (2)
• dropout (2)
• agi (2)
• estimator (2)
• podman (2)
• containers (2)
• rlhf (2)
• tesseract (2)
• pycon (2)
• fossasia (2)
• wavenet (2)
• openai (2)
• automl (2)
• magenta (2)
• embeddings (2)
• thailand (2)
• probability (2)
• singapore (2)
• newfrontier (2)
• lamda (2)
• palm (2)
• lora (2)
• rlaif (2)
• constitutionalai (2)
• self-debug (2)
• westworld (2)
• wizardcoder (2)
• phi-1 (2)
• self-refine (2)
• evolinstruct (2)
• leandojo (2)
• demo (2)
• devfest (2)
• alphageometry (2)
• alphazero (2)
• verification (2)
• genetic-programming (2)
• genetic-algorithms (2)
• hermes (2)
• llamacpp (2)
• dell-600m (2)
• php (2)
• laptop (2)
• vista (2)
• sqlite (2)
• asterisk (2)
• ruby (2)
• firewall (2)
• security (2)
• amd-catalyst (2)
• gt-750m (2)
• ipython (2)
• fedup (2)
• dns (2)
• systemd (2)
• dnsmasq (2)
• backup (2)
• ffmpeg (2)
• negativo (2)
• jupyter (2)
• router (2)
• webcam (2)
• key-bindings (2)
• gitolite (2)
• slack (2)
• aisprint (2)
• tpuv5-1 (2)
• tpuv5-8 (2)
• nnx (2)
• tunix (2)
• golang (2)
• attention (1)
• paraphrasing (1)
• multilingual (1)
• parallelism (1)
• algorithmictrading (1)
• machinelearning (1)
• nel (1)
• pca (1)
• poster (1)
• corpus (1)
• consciousness (1)
• attention-schema (1)
• awareness (1)
• chatbots (1)
• conference (1)
• o1 (1)
• architecture (1)
• d3js (1)
• dv (1)
• bokeh (1)
• gofai (1)
• pydata (1)
• artstyletransfer (1)
• bubblebreaker (1)
• training (1)
• intrinsic-dimension (1)
• googleio (1)
• mixed-precision (1)
• explainability (1)
• interpretability (1)
• visualqa (1)
• malaysia (1)
• tfsummit (1)
• nas (1)
• neural-architecture-search (1)
• lottery-ticket-hypothesis (1)
• tensorflow-extended (1)
• tfx (1)
• albert (1)
• tensorboard (1)
• causality (1)
• nerf (1)
• perceiver (1)
• perceiverio (1)
• pondernet (1)
• gpt (1)
• beam-search (1)
• top_k (1)
• top_p (1)
• sentencebert (1)
• retro (1)
• imagen (1)
• gpt3 (1)
• asr (1)
• audiolm (1)
• soundstream (1)
• vall-e (1)
• lightning (1)
• llmint8 (1)
• peft (1)
• controlnet (1)
• minedojo (1)
• vpt (1)
• dreamerv3 (1)
• toolformer (1)
• reflexion (1)
• multimodal (1)
• visiontransformer (1)
• vit (1)
• blip (1)
• blip-2 (1)
• minigpt-4 (1)
• instructblip (1)
• blip-diffusion (1)
• mediapipe (1)
• bard (1)
• codet5 (1)
• starcoder (1)
• tinystories (1)
• gpt4 (1)
• react (1)
• openendedness (1)
• pal (1)
• maxima (1)
• planning (1)
• learning-to-program (1)
• self-rag (1)
• memgpt (1)
• jarvis-1 (1)
• flowise (1)
• langchainjs (1)
• math-ai (1)
• alphastar (1)
• alphacode (1)
• alphafold (1)
• alphacodium (1)
• self-discover (1)
• ollama (1)
• ringattention (1)
• infini-attention (1)
• transformerfam (1)
• proving (1)
• mobilellm (1)
• aimo (1)
• do-over (1)
• unsloth (1)
• huggingface (1)
• aha (1)
• latentspace (1)
• iclr (1)
• efficiency (1)
• memory (1)
• arpo (1)
• deepthink (1)
• ipho (1)
• physics (1)
• physicsolympiad (1)
• nanobanana (1)
• tinyworlds (1)
• genie (1)
• prompting (1)
• sft (1)
• sam3 (1)
• sam3d-object (1)
• sam3d-body (1)
• evolution (1)
• grpo (1)
• on-policy (1)
• deepseek (1)
• engram (1)
• mhc (1)
• nestedlearning (1)
• samsung (1)
• littlebit (1)
• undercoveragent (1)
• world-models (1)
• jepa (1)
• physical-intelligence (1)
• gemma4 (1)
• gemma4-diffusion (1)
• prime-agent (1)
• universaltransformers (1)
• rlvr (1)
• gans (1)
• paired (1)
• sakana-rlt (1)
• spiral (1)
• spade (1)
• raro (1)
• sdpo (1)
• recursive-self-improvement (1)
• rsi (1)

• self-play-pretraining-with-zero-data (1)

• gemini-4-argon (1)