Hardware5h ago
DeepSeek DSpark Cuts Response Latency by 85%
The DSpark framework increases per-user response speeds by 60 to 85 percent. It uses a small model to propose token candidates that a larger model verifies in batches. This architecture extracts more performance from limited chip counts. DeepSeek effectively reduces its reliance on high-end US hardware through software optimization.