TransMLA: Multi-Head Latent Attention Is All You Need (NeurIPS 2025 Spotlight)
active 2025-10-23 → 2026-03-25 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 154 days from 2025-10-23 to 2026-03-25. Pushes: 3 total, peak 2 in a day. Pull requests: 0 total, peak 0 in a day. Issues: 1 total, peak 1 in a day. Comments: 6 total, peak 3 in a day. Stars: 14 total, peak 2 in a day.
- Pushes
- Pull requests
- Issues
- Comments
- Stars
Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.
Top contributors
Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity
| Contributor | Contributions | Pushes | PRs | Comments |
|---|---|---|---|---|
| xueliangyang-oeuler | 3 | 0 | 0 | 3 |
| fxmeng | 2 | 2 | 0 | 0 |
| Stanleytowne | 2 | 1 | 0 | 1 |
| haok1402 | 1 | 0 | 0 | 0 |
| Wfd567 | 1 | 0 | 0 | 1 |
| radna0 | 1 | 0 | 0 | 1 |
Recent activity
Latest issues, pull requests and releases
- Issue#42haok14022026-03-25 03:33Reproduction Steps for TransMLA-llama3-8b-8k
- Issue comment#40radna02025-12-28 06:32error occurred when deploying the transformed model using the vllm server
- Issue comment#38xueliangyang-oeuler2025-10-31 08:43Support Qwen3, e.g Qwen3-32B(GQA) can be tranasformed to MLA
- Issue comment#40xueliangyang-oeuler2025-10-31 08:39error occurred when deploying the transformed model using the vllm server
- Issue comment#25Wfd5672025-10-31 06:14add benchmark
- Issue comment#40Stanleytowne2025-10-29 06:20error occurred when deploying the transformed model using the vllm server
- Issue comment#40xueliangyang-oeuler2025-10-28 13:10error occurred when deploying the transformed model using the vllm server
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 14 stars here means stars gained during the window, not the repo's star count.