Xiaoyu Zhang 38076dea84 apply fused moe gate in ds v3/r1 (#5371) ...

Co-authored-by: Yineng Zhang <me@zhyncs.com>

2025-04-14 16:24:26 -07:00

..

attention

feat: use fa3 mla by default on hopper (#5210)

2025-04-12 01:09:25 -07:00

moe

apply fused moe gate in ds v3/r1 (#5371)

2025-04-14 16:24:26 -07:00

quantization

[perf] experimental enhance fp8 per-tensor quant (#5370)

2025-04-14 12:35:43 -07:00

activation.py

cleanup deps 2/n (#4464)

2025-03-15 23:06:17 -07:00

dp_attention.py

Fix DeepSeek DP Attention + torch compile (#5367)

2025-04-14 01:07:58 -07:00

elementwise.py

Fix run time error in ROCm platform (#5147)

2025-04-07 22:49:40 -07:00

layernorm.py

Urgent model support: support gemma-3-it (#4424)

2025-03-16 17:37:32 -07:00

linear.py

FP4 weight loading and inference (2/2) (#3972)

2025-04-08 17:26:21 -07:00

logits_processor.py

Reduce computation and communication in DP attention (#4521)

2025-03-18 13:41:36 -07:00

parameter.py

[Feature] Integrate DeepEP into SGLang (#4232)

2025-03-19 08:16:31 -07:00

pooler.py

Rename InputMetadata -> ForwardBatch (#1543)

2024-09-30 02:41:11 -07:00

radix_attention.py

Fix loading KV quantization scale; Enable modelopt kv cache (#4686)

2025-04-08 09:11:35 -07:00

rotary_embedding.py

Apply deepseek cuda rope (#5385)

2025-04-14 15:22:43 -07:00

sampler.py

Remove redundant type conversion (#4513)

2025-03-17 05:57:35 -07:00

torchao_utils.py

Support loading of larger models with on-the-fly quantization (#3061)

2025-01-22 21:33:17 -08:00

vocab_parallel_embedding.py

Clean up fp8 support (#4230)

2025-03-09 21:46:35 -07:00