Fastllm: A LLM inference library that runs DeepSeek-V4 with 10GB VRAM
**Brief Summary (2‑3 sentences)** Fastllm is an open‑source inference library that can run the DeepSeek‑V4 large language model on a single GPU with only 10 GB
Original source**Brief Summary (2‑3 sentences)** Fastllm is an open‑source inference library that can run the DeepSeek‑V4 large language model on a single GPU with only 10 GB
Original source