mrhaoxx a98d544833 merge: integrate origin/main into sft branch ...

Resolved 6 conflicts:
- CMakeLists.txt: keep cpptrace + debug flag, accept flexible build type
- worker_pool.cpp: keep SFT profiling + main's block=1 spin fix
- ext_bindings.cpp: keep both SFT MOE bindings and AVX2/BF16/FP8 bindings
- common.hpp: keep gpu_experts_mask + SFT backward weight fields
- __init__.py: export both generate_gpu_experts_masks and AMXSFTMoEWrapper
- experts.py: gpu_experts_mask for inference, num_gpu_experts for SFT, new methods

2026-04-08 23:19:28 +08:00

..

.gitignore

add kt-kernel

2025-10-12 05:13:00 +00:00

bench_moe_amx_int8.py

[feat](kt-kernel): CPU-GPU experts sched (#1796)

2026-01-16 17:01:15 +08:00

compare_tp_dumps.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

configuration_deepseek_v3.py

add kt-kernel

2025-10-12 05:13:00 +00:00

debug_expert_17_24.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

modeling_deepseek_v3.py

add kt-kernel

2025-10-12 05:13:00 +00:00

repro_llamafile_re.py

fix kt-kernel installation issue (#1603)

2025-11-12 15:56:02 +08:00

test_apply_rope.py

add kt-kernel

2025-10-12 05:13:00 +00:00

test_attention.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_awq_moe_amx.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_bf16_moe.py

support GLM 4.7 (#1791)

2026-01-13 17:36:25 +08:00

test_deepseekv3_prefill_speed.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_deepseekv3_prefill.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_deepseekv3.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_fp8_moe.py

[feat](kt-kernel): CPU-GPU experts sched (#1796)

2026-01-16 17:01:15 +08:00

test_fp8_perchannel_moe.py

support GLM 4.7 (#1791)

2026-01-13 17:36:25 +08:00

test_gate.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_k2_moe_amx.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_k2_write_buffer.py

Kt minimax (#1742)

2025-12-24 15:39:44 +08:00

test_linear.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_lora_b_zero_issue.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

test_mla_qlen.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_mla_quant.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_mla_simple.py

add kt-kernel

2025-10-12 05:13:00 +00:00

test_mla_torch.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_mla.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_mlp.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_moe_amx_perf.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

test_moe_amx.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

test_moe_kernel.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_moe_kml.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_moe_sft_amx_no_tp.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

test_moe_sft_amx.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

test_moe_sft_tp_debug.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

test_moe_sft_wrapper.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

test_moe.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

test_nan_with_real_data.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

test_partition_data.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

test_rope.cpp

add kt-kernel

2025-10-12 05:13:00 +00:00

test_rope.py

add kt-kernel

2025-10-12 05:13:00 +00:00

test_skip_lora.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

test_softmax.py

add kt-kernel

2025-10-12 05:13:00 +00:00

test_write_buffer.py

support GLM 4.7 (#1791)

2026-01-13 17:36:25 +08:00

test-debug.py

[fix](test): fix import kt-kernel (#1728)

2025-12-17 19:46:32 +08:00

torch_attention.py

add kt-kernel

2025-10-12 05:13:00 +00:00

verify_pt_layout.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00