lizhenyun01 2be8656c29 [BugFix] fix mtp split kv attetion (#5920) ...

* [BugFix] fix mtp split kv attetion

* clean code

* clean code

2026-01-07 04:07:31 -08:00

..

template_instantiation

【Hackathon 9th No.86】autogen MultiQueryDecoderAttention template_instantiation -part (#4383)

2025-10-16 17:08:19 +08:00

append_attention_c4_impl.cuh

Support GPT-OSS-BF16 (#4240)

2025-10-20 14:44:58 +08:00

append_attention_c8_impl.cuh

Support GPT-OSS-BF16 (#4240)

2025-10-20 14:44:58 +08:00

append_attention_c16_impl.cuh

Support GPT-OSS-BF16 (#4240)

2025-10-20 14:44:58 +08:00

append_attention_func.cuh

[BugFix] fix mtp split kv attetion (#5920)

2026-01-07 04:07:31 -08:00

append_attention_kernel.h

Support GPT-OSS-BF16 (#4240)

2025-10-20 14:44:58 +08:00

decode_attention_func.cuh

【Hackathon 9th No.86】autogen MultiQueryDecoderAttention template_instantiation -part (#4383)

2025-10-16 17:08:19 +08:00

decoder_mla_attention_kernel.cu

【Hackathon 9th No.86】autogen MoeFastHardamardImplWrapper template_instantiation (#4592)

2025-10-30 10:28:36 +08:00

decoder_mla_attention_kernel.h

【Hackathon 9th No.86】autogen MoeFastHardamardImplWrapper template_instantiation (#4592)

2025-10-30 10:28:36 +08:00

decoder_write_cache_with_rope_impl.cuh

[Feature][Optimization] Qwen Support Dynamic block_wise_fp8 cache (#5486)

2025-12-12 17:10:17 +08:00

decoder_write_cache_with_rope_kernel.cu

[Feature][Optimization] Qwen Support Dynamic block_wise_fp8 cache (#5486)

2025-12-12 17:10:17 +08:00

decoder_write_cache_with_rope_kernel.h

clean code in sttantion (#3917)

2025-09-05 20:49:01 +08:00

encoder_write_cache_with_rope_impl.cuh

[Feature][OP] Append Attn Support CUDA-PDL (#5072)

2025-11-17 20:47:33 +08:00

encoder_write_cache_with_rope_kernel.h

[Feature] Support zai-org/GLM-4.5-Air BF16 model (#3928)

2025-09-10 19:36:10 +08:00

get_block_shape_and_split_kv_block.cu

[Metax] modify wrapSize to WARP_SIZE (#5442)

2025-12-09 01:44:02 -08:00

gqa_rope_write_cache.cu

support fa3 qwen-vl rope (#5869)

2026-01-05 15:29:34 +08:00

mem_util.cuh

[Metax] adapt DeepSeek (#4498)

2025-10-24 10:14:53 +08:00

mla_cache_kernel.cu

[PD Disaggregation] Support PD deployment of DeepSeekv3. (#5251)

2025-12-02 14:11:50 +08:00

mla_cache_kernel.cuh

[Inference, rename] remove padding_offsets from atten use batch_id_per_token (#2880)

2025-07-17 18:41:31 +08:00

mma_tensor_op.cuh

…

multiquery_attention_c4_impl.cuh

[BugFix] fix mtp split kv attetion (#5920)

2026-01-07 04:07:31 -08:00

multiquery_attention_c4_kernel.h

Support GPT-OSS-BF16 (#4240)

2025-10-20 14:44:58 +08:00

multiquery_attention_c8_impl.cuh

[BugFix] fix mtp split kv attetion (#5920)

2026-01-07 04:07:31 -08:00

multiquery_attention_c8_kernel.h

Support GPT-OSS-BF16 (#4240)

2025-10-20 14:44:58 +08:00

multiquery_attention_c16_impl.cuh

[BugFix] fix mtp split kv attetion (#5920)

2026-01-07 04:07:31 -08:00

multiquery_attention_c16_kernel.h

Support GPT-OSS-BF16 (#4240)

2025-10-20 14:44:58 +08:00

multiquery_decoder_attention_impl.cuh

【Hackathon 9th No.86】autogen MultiQueryDecoderAttention template_instantiation -part (#4383)

2025-10-16 17:08:19 +08:00

multiquery_decoder_attention_kernel.h

【Hackathon 9th No.86】autogen MultiQueryDecoderAttention template_instantiation -part (#4383)

2025-10-16 17:08:19 +08:00

pre_cache_len_concat.cu

[Others] Remove useless code (#5404)

2025-12-08 13:59:46 +08:00

qwen3_rope.h

support fa3 qwen-vl rope (#5869)

2026-01-05 15:29:34 +08:00

speculate_write_cache_with_rope_impl.cuh

[BugFix][Speculative Decoding](Spend many dyas to solve)Fix write qknorm cache bug in speculative decoding (#5491)

2025-12-15 18:27:11 +08:00

speculate_write_cache_with_rope_kernel.cu

[BugFix][Speculative Decoding](Spend many dyas to solve)Fix write qknorm cache bug in speculative decoding (#5491)

2025-12-15 18:27:11 +08:00

speculate_write_cache_with_rope_kernel.h

support mtp rope_3d (#3791)

2025-09-04 17:18:05 +08:00

template_config.json

revert group size 3 (#5079)

2025-11-17 18:54:13 +08:00

utils.cuh

Support GPT-OSS-BF16 (#4240)

2025-10-20 14:44:58 +08:00