Day-0 support for @thinkymachines Inkling 2 Small on TokenSpeed.
On the @SemiAnalysis_ AgentX benchmark, DSPARK deliver | Hanami
Day-0 support for @thinkymachines Inkling 2 Small on TokenSpeed.
On the @SemiAnalysis_ AgentX benchmark, DSPARK delivers 476 tokens/s per user on 4× NVIDIA B300 GPUs—75% faster decoding than the target model alone.
Also supported on AMD Instinct MI350X. More technical details 👇
📖 https://t.co/XvGOCslWvw
⭐️ https://t.co/XNzaNRclQ2
Retweeted by @SemiAnalysis_ from @lightseekorg