Back    Zoom +    Zoom -
LENOVO GROUP Unveils 1st AMD-Powered AI Server with Full DeepSeek Deployment on Single Machine
Recommend
6
Positive
10
Negative
8
LENOVO GROUP (00992.HK) announced that its first AMD-chip-powered AI foundation model training server, the "Lenovo Wentian WA7785a G3", achieved a peak throughput of 6708 tokens per second when deploying the full 671B version of the DeepSeek foundation model on a single machine. This milestone marks a new high in the performance of a single server running foundation models.

High-performance servers are pivotal for training foundation models. The exceptional capabilities of the "Lenovo Wentian WA7785a G3" signify higher efficiency and capacity in handling complex AI tasks, potentially paving the way for breakthroughs in research and commercial applications in related fields.

Related NewsM Stanley Lists CN Stocks as Direct AI Beneficiaries (Table)

AAStocks Financial News