bahnland/dodoco-routing-matrices
DODOCO MoE Dispatch Matrices on Real Text Per-dispatch AlltoAll send-count matrices from expert-parallel training steps of five released Mixture-of-Experts checkpoints, on 16 GPUs, with real text as input. Each matrix records how many tokens each of the 16 ranks sends to each other rank in one MoE dispatch. The data is meant as workload input for interconnect, collective and topology studies, so that measured MoE traffic can be used without setting up the training stack.… See the full description on the dataset page: https://huggingface.co/datasets/bahnland/dodoco-routing-matrices.
| Author / organization | bahnland |
|---|---|
| License (as declared) | cc-by-sa-4.0 |
| Downloads | 0 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 17 October 2026 |
| Last updated at source | 17 October 2026 |
Connected in the knowledge graph
Recent changes
- 17 Oct 2026 Description: none → DODOCO MoE Dispatch Matrices on Real Text Per-dispatch AlltoAll send-count matrices from expert-parallel training steps of five released Mixture-of-Experts checkpoints, on 16 GPUs, with real text as input. Each matrix records how many token
- 17 Oct 2026 Topics added mixture-of-experts, expert-parallelism, alltoall, interconnect, network-traces, communication
Changes detected by Silytics between refreshes of the source.
Related papers
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.