Training video LoRAs on 2x RTX 5060 Ti: consumer multi-GPU feasibility
Inner_Employment_332 · reddit · 2026-09-17
Priced out of 5090/A6000-class GPUs (€5k), the author builds a dual RTX 5060 Ti system (€600 + €750) for inference and LoRA training. Key points: ComfyUI handles multi-GPU inference fine; Ostris AI Toolkit's GUI lacks multi-GPU but its CLI uses PyTorch FSDP + accelerate; diffusion-pipe uses pipeline parallelism to train video LoRAs (LTX 2.5, MiniMax H3) on consumer cards; environment is Windows 11 with WSL2 Debian. Author asks whether the setup avoids OOM, spurning cloud rentals over content policy.
More from Infra
- OpenAI's Jalapeño chip wasn't AI-designed: 100+ ex-Google TPU engineers and Broadcom were — JFPuget · 2026-09-17
- S3 is supposed to span AZs — devs question cheapest-tier AWS data loss claim — zetalyrae · 2026-09-17
- Build the agent setup first, pick the model second: a Linux + Tailscale + llama.cpp stack guide — max_paperclips · 2026-09-17
- Huawei unveils Ascend 960 chips for 2027 and UnifiedBus linking one million processors — mark_k · 2026-09-17
- Leak: CXMT Supplies TSV Embedded DRAM Dies for Chinese cHBM, Stacking Done In-House or via Local OSAT — zephyr_z9 · 2026-09-17
- OpenAI's Astra gets 3x throughput on NVIDIA Vera Rubin, plus 2x more in 72 hours — MickeySteamboat · 2026-09-17