A wrong person..ย โขย 2d
Hey everyone, I recently built a workflow in ComfyUI and wanted to turn it into a small web app. I tried deploying it on RunPod, but because I didnโt understand how GPU billing works, I accidentally kept the instance running 24/7 and lost around โน5,000 in a day. ๐ After some research I found out about on-demand / pay-per-request GPUs, but Iโm still not totally sure how to set them up correctly for ComfyUI workflows. If anyone here has experience with: serverless GPU options on-demand ComfyUI deployment or best practices to avoid constant runtime charges โฆIโd really appreciate some guidance. Thanks in advance!
A wrong person..ย โขย 6d
I built an AI workflow and wanted to turn it into an app, but messed up the deployment part. Hosted it on Runpod 24/7 and wasted โน5,000 before realizing I shouldโve used per-request GPU instead. ๐ญ If anyone knows the right way to deploy ComfyUI wor
See More"I am the architect ...ย โขย 1y
NVIDIA- THE DOMINANT FORCE 1. NVIDIA is a leading force in AI and GPU technology. 2. Their GPUs, like the H100 Tensor Core, are critical for AI development, including training models like ChatGPT. 3. NVIDIA's stock has surpassed a $1 trillion ma
See MoreHey I am on Medialย โขย 4m
India might be behind in building its own AI, but we donโt have to stay out of the game. Right now, most AI video generation tools are priced in dollars โ way out of reach for Indian students and creators. While open-source AI video generators do exi
See MoreCloud | DevOps | AIย โขย 4m
Just launched ๐ LabStack, a personal project I built to let developers spin up fully isolated, cloud-based dev environments right from the browser โ in under 30 seconds โก Itโs powered by Docker ๐ณ containers running on on-demand EC2 instances, orch
See More7โs Club MD | ED 7's...ย โขย 2m
Hey everyone! I wanted to share an idea Iโve been working on and get your thoughts. Iโm exploring the possibility of creating a super affordable external GPU solution for professionals and creators here in India. The idea is to provide something like
See More
Willing to contribut...ย โขย 25d
I fine-tuned 3 models this week to understand why people fail. Used LLaMA-2-7B, Mistral-7B, and Phi-2. Different datasets. Different methods (full tuning vs LoRA vs QLoRA). Here's what I learned that nobody talks about: 1. Data quality > Data quan
See More
I help businesses to...ย โขย 8m
Full - Stack marketer๐จ A new wave has entered the marketing field, and Iโve been digging into blogs and posts about it for a few weeks. Yup, itโs a bit scary ๐ but also opening up tons of doors at the same time. Okay, keeping it simple: a Full-S
See MoreDownload the medial app to read full posts, comements and news.