DDoS Protection
Resources allocated to users are fully isolated to ensure data security. GPU Mart protects against DDoS from the edge fast while ensuring legitimate traffic of Nvidia GPU cloud server is not compromised.
Express GPU Dedicated Server - P600
Express GPU Dedicated Server - P620
Basic GPU Dedicated Server - GTX 1650
Professional GPU Dedicated Server - RTX 2060
Basic GPU Dedicated Server - RTX 5060
Advanced GPU Dedicated Server - V100
Multi-GPU Dedicated Server - 2xRTX 4060
Multi-GPU Dedicated Server - 3xRTX A6000
| Use Case Type | Recommended Servers | Description |
|---|---|---|
| Chatbot / LLM Inference API | RTX 4090 / A100 / A6000 / H100 | Ideal for deploying models like Vicuna, LLaMA, Mistral, GPTQ, Exllama, DeepSeek, etc. |
| Fine-tuning / RAG Retrieval | A100 (80GB) / 2x A100 / 3x V100 / 4x A100 | For fine-tuning large models with small datasets, building embeddings, vector indexing, RAG tasks |
| AI Video Generation & Imaging | RTX 5090 / RTX 4090 / RTX 3060 Ti / RTX 5060 | Run image/video generation models like Stable Diffusion XL, RunwayML, ControlNet, AnimateDiff |
| Speech Recognition & Transcription | RTX 3060 Ti / RTX A4000 / RTX 2060 | Supports Whisper + VAD + audio separation models, suitable for real-time speech-to-text tasks |
| Research / Educational Training | RTX A4000 / RTX 2060 / GTX 1650 / V100 | Ideal for classroom demos, academic training, and development/testing environments |
| Multi-model / Multi-task Workloads | 3x V100 / 2x A100 / 4x A100 | Efficient for running concurrent inference sessions and distributed AI workloads |
| Enterprise-Level AI Computing | RTX A6000 / RTX 4090 / A100 (80GB) / H100 | Built for large-scale LLMs, generative AI, GNNs, and video big data analytics |