Question about online system

bytedance / HLLM

HLLM: Enhancing Sequential Recommendations via Hierarchical Large Language Models for Item and User Modeling

Apache License 2.0

172 stars 21 forks source link

Closed ccccjunkang closed 4 weeks ago

ccccjunkang commented 1 month ago

Could you please disclose the actual number of samples, number of GPUs used, and number of model parameters in the online system?

ssyzeChen commented 4 weeks ago

Hi. We use a model with the same parameter size as described in the paper, and trained on hundreds of millions of samples.