JohnQinAMD 6dfd87f596 [ROCm][DSv4.1] Keep the Engram tables in host memory on ROCm (#57491)

2026-09-21 10:17:11 +08:00

..

__init__.py

[Refactor] Normalize DeepSeek V4.1 model package naming (#56741)

2026-09-13 23:17:50 -07:00

dspark.py

[Model][DSv4.1] Store the whole KV in MXFP8 (FlashMLA V4.1 record) (#56893)

2026-09-14 20:12:55 -07:00

model.py

[ROCm][DSv4.1] Keep the Engram tables in host memory on ROCm (#57491)

2026-09-21 10:17:11 +08:00

rocm.py

[ROCm][DSv4.1][Perf] Reuse the decode topk ragged metadata across layers (#57434)

2026-09-20 11:46:33 +08:00

vl_model.py

[DSV4.1] Add encoder cuda graph support for deepseek-v4.1-flash (#56625)

2026-09-20 21:51:13 +08:00