Overview
I am looking for a GPU server to run a short batch processing task. I only need a solution that can be quickly deployed, run batch scripts, and then be destroyed after completion. I do not need a monthly dedicated server rental or a long-term contract.Specific requirements:
- Hourly billing
- Minimum 32 GB VRAM
- RTX 3090 / 4090 / A5000 or similar GPUs
- Preferably 64GB if the price is right.
- Pre-built images with CUDA + vLLM or Ollama included
- Ideally, an OpenAI-compatible API should be provided.
- Qwen 14B/32B or similar models
- Good disk/network speed is beneficial for model downloads
- Initial usage time is approximately 6–20 hours.
