#pytorch

Things I've written about #pytorch

#kafka(23)
#machine-learning(19)
#distributed-systems(17)
#performance(14)
#nanochat(12)
#llm(10)
#architecture(8)
#redpanda(8)
#zero-copy(8)
#ai(7)
#reliability(7)
#tutorial(7)
#linux(7)
#quantization(6)
#kimi-k3(6)
#practical-guide(6)
#optimization(6)
#ai-engineering(5)
#deep-learning(5)
#streaming(4)
#model-distillation(4)
#distillation(4)
#networking(4)
#inference(4)
#protocol(4)
#pytorch(4)
#transformers(4)
#training(4)
#leadership(3)
#safety(3)
#engineering(3)
#decision-making(3)
#vendor-selection(3)
#production-systems(3)
#nextjs(3)
#protocols(3)
#compatibility(3)
#model-compression(3)
#lora(3)
#mixture-of-experts(3)
#benchmarking(3)
#idempotency(3)
#consumers(3)
#benchmarks(3)
#systems(3)
#security(2)
#react(2)
#consumer-groups(2)
#platform-engineering(2)
#pruning(2)
#cost-optimization(2)
#seastar(2)
#raft(2)
#latency(2)
#replication(2)
#serialization(2)
#storage(2)
#model-audit(2)
#reproducibility(2)
#observability(2)
#distributed-inference(2)
#migration(2)
#gpu(2)
#attention(2)
#production(2)
#mermaid(2)
#pragmatic-engineering(2)
#evaluation(2)
#fine-tuning(2)
#tokenization(2)
#edge-computing(2)
#io-uring(2)
#rce(1)
#cve(1)
#javascript(1)
#web-security(1)
#debugging(1)
#rust(1)
#ibm(1)
#confluent(1)
#infrastructure(1)
#migrations(1)
#introduction(1)
#scalability(1)
#knowledge-distillation(1)
#ai-safety(1)
#mlops(1)
#synthetic-data(1)
#data-quality(1)
#python(1)
#enterprise-ai(1)
#cluster-management(1)
#controller(1)
#consensus(1)
#rpc(1)
#c(1)
#log-structured-storage(1)
#tiered-storage(1)
#object-storage(1)
#retention(1)
#wasm(1)
#transforms(1)
#data-processing(1)
#open-weights(1)
#supply-chain(1)
#ai-agents(1)
#authorization(1)
#long-context(1)
#kv-cache(1)
#retrieval(1)
#multi-gpu(1)
#tensor-parallelism(1)
#expert-parallelism(1)
#gpu-topology(1)
#kraft(1)
#operations(1)
#internals(1)
#idempotence(1)
#transactions(1)
#compression(1)
#zstd(1)
#lz4(1)
#multi-tenant(1)
#concurrency(1)
#outbox(1)
#retries(1)
#recovery(1)
#queues(1)
#hardware(1)
#api(1)
#tool-calling(1)
#context-caching(1)
#mxfp4(1)
#model-weights(1)
#checkpoint(1)
#api-pricing(1)
#self-hosting(1)
#experiments(1)
#ablation(1)
#model-architecture(1)
#case-study(1)
#build-optimization(1)
#static-site-generation(1)
#remark-plugins(1)
#custom-hooks(1)
#pan-zoom(1)
#svg-rendering(1)
#over-engineering(1)
#feature-creep(1)
#css(1)
#simplification(1)
#refactoring(1)
#mermaid-tutorial(1)
#lessons-learned(1)
#core(1)
#testing(1)
#distributed-training(1)
#zero(1)
#sft(1)
#chat(1)
#scaling-laws(1)
#memory-optimization(1)
#rlhf(1)
#reinforcement-learning(1)
#grpo(1)
#policy-gradient(1)
#bpe(1)
#tiktoken(1)
#data-engineering(1)
#beginner(1)
#tiny-llm(1)
#efficient-models(1)
#mobile-ai(1)
#mathematics(1)
#comparison(1)
#deployment-patterns(1)
#deployment(1)
#raspberry-pi(1)
#mobile(1)
#qlora(1)
#domain-adaptation(1)
#tls(1)
#dpdk(1)
#spdk(1)
#rdma(1)
#gpudirect(1)
#memory(1)

Oct 17, 2025/20 mins read
NanoChatTechnical Deep DivesPart 3/13
Distributed Muon: Custom Gradient Synchronization for Memory-Efficient Training
Oct 15, 2025/21 mins read
NanoChatTechnical Deep DivesPart 2/13
The Muon Optimizer Explained: Why Orthogonal Gradients Work
Oct 14, 2025/17 mins read
NanoChatTechnical Deep DivesPart 1/13
Build Your Own ChatGPT for $100
Sep 23, 2025/20 mins read
Tiny Language ModelsTraining & OptimizationPart 6/10
Knowledge Distillation: How to Train a 1.5B Model That Matches Your 7B

#pytorch

Distributed Muon: Custom Gradient Synchronization for Memory-Efficient Training

The Muon Optimizer Explained: Why Orthogonal Gradients Work

Build Your Own ChatGPT for $100

Knowledge Distillation: How to Train a 1.5B Model That Matches Your 7B