ContentPosts from @superfin..
Link
@faun shared a link, 1 month, 4 weeks ago

Run the Full DeepSeek-R1-0528 Model Locally

DeepSeek-R1-0528's nanized form chops space needs down to162GB. But here's the kicker—without a solid GPU, it's like waiting for paint to dry...

Run the Full DeepSeek-R1-0528 Model Locally
Link
@faun shared a link, 1 month, 4 weeks ago

Training a Rust 1.5B Coder LM with Reinforcement Learning (GRPO)

DeepSeek-R1flips the script on training LLMs. Armed withGRPO, it challenges the industry heavies like OpenAI's o1 by playing smart with custom data and cleverly designed rewards. Imagine this: a humble 1.5B model, running on merely asingle H100, clocks in at an 80% build pass rate. It’s nibbling at ..

Training a Rust 1.5B Coder LM with Reinforcement Learning (GRPO)
Link
@faun shared a link, 1 month, 4 weeks ago

AWS' custom chip strategy is showing results, and cutting into Nvidia's AI dominance

Graviton4just cranked up the juice to600 Gbps. In the grand race of public cloud champions, it's gunning straight for Nvidia's AI kingdom, powered by the formidableProject Rainier...

AWS' custom chip strategy is showing results, and cutting into Nvidia's AI dominance
Link
@faun shared a link, 1 month, 4 weeks ago

End to End Argo-Workflow for CI/CD

Argo Workflowsisn't just another tool; it sings for Kubernetes-native CI/CD. It juggles complex workflows as DAGs, brings dynamic execution to life with CRDs and parameters. Got a weekly CI? Automate it withCronWorkflows. Secure those Docker pushes using Kubernetes secrets, and let shared volumes ha..

End to End Argo-Workflow for CI/CD
Link
@faun shared a link, 1 month, 4 weeks ago

The Ultimate Guide to Running Kubernetes in a Home Lab

K3sandMicroK8sshine in makeshift home labs with minimal hardware. Throw inLonghornfor storage andVelerofor backup bliss. Now that's a recipe for tech nirvana...

Link
@faun shared a link, 1 month, 4 weeks ago

Kubernetes 1.33: Resizing Pods Without the Drama (Finally!)

Kubernetes 1.33brings in-place pod vertical scaling, allowing you to adjust CPU and memory without restarting pods, a game-changer for seamless resource management in production workloads. This feature simplifies vertical pod autoscaling especially for stateful workloads like databases...

Kubernetes 1.33: Resizing Pods Without the Drama (Finally!)
Link
@faun shared a link, 1 month, 4 weeks ago

A Journey Through Kafkian SplitDNS in a Multitenant Kubernetes Offering

SCHIPfaced off with tenant demands for serverless Kafka. Their weapon of choice? A crafty DNS trick usingCoreDNSand a few clevernode-local DNSadjustments. They kept multitenancy alive and kicking without wearing out the ops team. Nice move...

A Journey Through Kafkian SplitDNS in a Multitenant Kubernetes Offering
Link
@faun shared a link, 1 month, 4 weeks ago

GitOps for Kubernetes With Nixidy and ArgoCD

Nixidyturns Kubernetes YAMLs into sleek, declarative Nix setups. It offers a robust, repeatable config flow—even for those complex Helm charts. Spice up your deployment by pairingArgoCDwith encrypted secrets viasops-secrets-operator. Now you can wrangle sensitive data in Git with style—and security...

GitOps for Kubernetes With Nixidy and ArgoCD
Link
@faun shared a link, 1 month, 4 weeks ago

Interesting Kubernetes application demos

Kubeappsis your backstage pass to deploying and controllingK8sapps with style. Dive into a treasure chest ofHelmcharts ready to roll. For those looking to jazz up a demo, unleashKubedoomorKubevaders. Obliteratepodsfor stress-testing, or just because you can. Craving some retro-futuristic fun? Check ..

Link
@faun shared a link, 1 month, 4 weeks ago

How We Designed Model Runner and What’s Next

Docker's just unleashed a new gadget with Desktop4.40. Meet theModel Runner, your ticket to running AI models on your local machine. Imagine it as the Peacekeeper of container-host diplomacy. It’s powered byllama.cppand can ride GPUs like a pro skater. Oh, and it plays nice with theOpenAI API. Model..

How We Designed Model Runner and What’s Next