Fine-tuning Qwen 27B for $42.98: Ras Cloud dashboard shows the full cheap LoRA run
Rasmic · x · 2026-10-21
Rasmic posted 'we fine tune models' with a Ras Cloud control-plane screenshot: a Qwen38 27B QLoRA run (4-bit NF4, LoRA on all linear layers) on a coding-sft-v3 dataset.
Key numbers from the dashboard
- Step 184/344 (53.5%), 8h01m elapsed, 5h33m ETA
- Single on-demand Daytona H100 80GB, $85/hr tier noted in logs
- Estimated cost $41.55, spent $42.98 of a $3,000 account ($2,957 left)
- Train/eval loss 0.645 / 0.441 with live curves and sandbox scheduling logs
A transparent look at how cheap observable fine-tuning of open models has become.
More from coding & agent
- Ethan Mollick's new book charts 30 months of AI progress, from cyborg to conjurer work — rseroter · 2026-10-23
- OpenJet claims Qwen3.8-27B coding runs 1.9x faster than llama.cpp on one RTX 3090 — JeremyCMorgan · 2026-10-23
- Jensen Huang details Vera Rubin agentic stack: NVL72, MGX, CMX and RTX — rohanpaul_ai · 2026-10-23
- Building an autonomous SEO agent in Hermes with three skills: programmatic pages, Keywords Everywhere, DataForSEO — EXM7777 · 2026-10-23
- Building a sports betting AI agent with onchain execution and human approval by default: architecture and lessons — Dry-Comfortable-2678 · 2026-10-23
- Capy 0.5.0 lets Claude Pro/Max subscribers reuse their plan quota, adds OIDC cloud access — garrytan · 2026-10-23