Hubert Lu 2c562fd2d0 Fix Llama 4 with MXFP4 dynamic quant on MI35x (#9993)

2025-09-04 00:48:58 -07:00

..

attention

[feat]Ascend NPU Gemma-3-12b and Gemma-3-27b support (#8909)

2025-08-31 00:25:07 -07:00

moe

Triton 3.4.0 MoE config for Deepseek TP16 H100 (#9978)

2025-09-03 13:16:16 -07:00

quantization

Fix Llama 4 with MXFP4 dynamic quant on MI35x (#9993)

2025-09-04 00:48:58 -07:00

activation.py

[feat]Ascend NPU Gemma-3-12b and Gemma-3-27b support (#8909)

2025-08-31 00:25:07 -07:00

amx_utils.py

[CPU] refine CPU integration code (#7647)

2025-07-03 09:51:09 -07:00

communicator.py

Revert "Optimized deepseek-v3/r1 model performance on mxfp4 run (#9671)" (#9959)

2025-09-03 00:50:04 -07:00

dp_attention.py

[Bugfix] fix kv buffer register & dp attention & deepepmoe (#9327)

2025-08-19 10:09:48 -07:00

elementwise.py

Update grok.py and tiktoken tokenizer (#9532)

2025-08-23 05:40:18 -07:00

flashinfer_comm_fusion.py

Support trtllm_allreduce_fusion in flashinfer for cuda<12.8 (#9339)

2025-08-20 16:54:30 -07:00

layernorm.py

[feat]Ascend NPU Gemma-3-12b and Gemma-3-27b support (#8909)

2025-08-31 00:25:07 -07:00

linear.py

[ModelOpt] Fix Weight Loading for DSR1-FP4 Quantization (#9712)

2025-08-29 17:13:52 -07:00

logits_processor.py

feat: add original logprobs to response (#8375)

2025-08-29 11:43:57 -07:00

model_parallel.py

Move parsers under a single folder (#9912)

2025-09-02 18:25:04 -07:00

multimodal.py

[VLM] Improving multimodal tensor hash kernel (#9008)

2025-08-14 13:45:55 -07:00

parameter.py

[CPU][Qwen3 MoE] Enable fused_topk CPU fusion and enhance FP8 TP padding (#7838)

2025-07-09 02:04:21 -07:00

pooler.py

Feat/support rerank (#6058)

2025-06-16 10:50:01 -07:00

radix_attention.py

Update grok.py and tiktoken tokenizer (#9532)

2025-08-23 05:40:18 -07:00

rotary_embedding.py

Remove mrope position sync (#9460)

2025-09-03 11:40:53 -07:00

sampler.py

feat: add original logprobs to response (#8375)

2025-08-29 11:43:57 -07:00

torchao_utils.py

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

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

utils.py

move is_sm90_supported/is_sm100_supported to python/sglang/srt/utils.py (#9679)

2025-08-27 03:34:29 -07:00

vocab_parallel_embedding.py

Replace torch.jit.script with torch.compile in get_masked_input_and_mask to fix benchmark underreporting (#8733)

2025-08-03 19:02:51 -07:00