You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Unsloth LoRA/QLoRA fine-tuning with full checkpoint versioning on Backblaze B2 — every dataset, per-epoch checkpoint, adapter, config, and metric is keyed under runs/<run_id>/. Real local fine-tunes of Llama/Qwen/Gemma on GPU (Unsloth) or CPU (transformers+PEFT); $0 per run, only B2 credentials.
NotebookLM-style podcast generator built on Backblaze B2: turn PDFs and docs into a two-host AI audio overview with OpenAI script generation and TTS, storing sources, transcripts, and episode audio in B2 cloud storage via the S3-compatible API.
Python For GCP is a project aimed at simplifying the interaction with Google Cloud Platform (GCP) services using Python. This repository provides code examples and scripts that help you manage and automate various GCP resources such as BigQuery, Cloud Storage, BigTable, Compute Engine, and more entirely through Python.
This Terraform module sets up a workflow where files stored in Cloud Storage trigger events in Eventarc, which then processes the files and stores the data in Spanner.
Full-stack sample app that runs MONAI medical-image segmentation on DICOM/NIfTI volumes and stores the source scans, segmentation masks, and preview slices in Backblaze B2 via the S3-compatible API. Next.js 16 + FastAPI; models run on-device (CPU by default, GPU auto-detected).