user: george_123

9 karma | created 6 years ago

recent submissions

Loading Llama-2 70B 20x faster with Anyscale Endpoints (anyscale.com)

3 pts|2 years ago|discuss

(no title)

2 years ago|discuss

(no title)

2 years ago|discuss

How continuous batching improves LLM inference throughput 23x (twitter.com)

1 pts|2 years ago|discuss

(no title)

2 years ago|discuss

(no title)

3 years ago|discuss

Ant Group – scaling to 1.37M QPS on Ray (anyscale.com)

3 pts|3 years ago|1 comment

(no title)

3 years ago|discuss

(no title)

1 pts|3 years ago|discuss