oldcap

3 ms·

1. 1.▲
   Amazon's Exabyte-Scale Migration from Apache Spark to Ray (Saving $120M / Year)(aws.amazon.com)

   2 points by oldcap 2y ago|0 comments
2. 2.▲
   Uber's Journey from Predictive to Generative AI(uber.com)

   1 points by oldcap 2y ago|0 comments
3. 3.▲
   by oldcap 3y ago

   GH: bit.ly/llm-perf Blog: https://www.anyscale.com/blog/comparing-llm-performance-intr...
4. 4.▲
   by oldcap 3y ago

   Short link https://bit.ly/load-llm
5. 5.▲
   Loading LLM (Llama-2 70B) 20x faster with Anyscale Endpoints(anyscale.com)

   5 points by oldcap 3y ago|2 comments
6. 6.▲
   by oldcap 3y ago

   How long does it take to download Llama2 70B? On the 4x 25 Gbps NICs that aws.p4de's have, it should take ~10s. Yet in production we've observed much higher times, which makes autoscaling less responsive + more expensive. This blo
7. 7.▲
   by oldcap 4y ago

   AWS actually has a _higher_ unit cost than Alibaba Cloud
8. 8.▲
   by oldcap 6y ago

   "We believe that Ray will continue to play an increasingly important role in bringing much needed common infrastructure and standardization to the production machine learning ecosystem, both within Uber and the industry at large."