LICENSE
README.md
pyproject.toml
sketchssm/__init__.py
sketchssm.egg-info/PKG-INFO
sketchssm.egg-info/SOURCES.txt
sketchssm.egg-info/dependency_links.txt
sketchssm.egg-info/entry_points.txt
sketchssm.egg-info/requires.txt
sketchssm.egg-info/top_level.txt
sketchssm/calibration/__init__.py
sketchssm/calibration/__main__.py
sketchssm/calibration/bundles.py
sketchssm/calibration/calibration.py
sketchssm/calibration/run.py
sketchssm/calibration/adapters/__init__.py
sketchssm/calibration/adapters/base.py
sketchssm/calibration/adapters/gdn.py
sketchssm/calibration/adapters/kda.py
sketchssm/calibration/adapters/mamba2.py
sketchssm/calibration/collection/__init__.py
sketchssm/calibration/collection/config.py
sketchssm/calibration/collection/data.py
sketchssm/calibration/collection/engine.py
sketchssm/calibration/collection/forced_tokens.py
sketchssm/calibration/collection/models.py
sketchssm/calibration/collection/paired.py
sketchssm/calibration/collection/pipeline.py
sketchssm/calibration/collection/storage.py
sketchssm/calibration/collection/worker.py
sketchssm/calibration/collection/hooks/__init__.py
sketchssm/calibration/collection/hooks/gdn.py
sketchssm/calibration/collection/hooks/kda.py
sketchssm/calibration/collection/hooks/kda_statistics.py
sketchssm/calibration/collection/hooks/mamba2.py
sketchssm/calibration/collection/hooks/mamba_scan.py
sketchssm/calibration/collection/native/__init__.py
sketchssm/calibration/collection/native/common.py
sketchssm/calibration/collection/native/gdn.py
sketchssm/calibration/collection/native/kda.py
sketchssm/calibration/collection/native/mamba2.py
sketchssm/calibration/collection/observers/__init__.py
sketchssm/calibration/collection/observers/gdn.py
sketchssm/calibration/collection/observers/kda.py
sketchssm/calibration/collection/observers/mamba2.py
sketchssm/calibration/collection/packed/__init__.py
sketchssm/calibration/collection/packed/fp4.py
sketchssm/calibration/collection/packed/glm.py
sketchssm/calibration/collection/packed/qwen.py
sketchssm/calibration/collection/packed/qwen3_5.py
sketchssm/calibration/collection/packed/super.py
sketchssm/calibration/core/__init__.py
sketchssm/calibration/core/_solver.py
sketchssm/calibration/core/_statistics.py
sketchssm/calibration/core/allocation.py
sketchssm/calibration/core/basis.py
sketchssm/calibration/core/export.py
sketchssm/calibration/core/frames.py
sketchssm/calibration/core/reuse.py
sketchssm/calibration/core/scoring.py
sketchssm/calibration/core/traffic.py
sketchssm/calibration/tools/__init__.py
sketchssm/calibration/tools/check_gdn_hooks.py
sketchssm/calibration/tools/check_kda_collection.py
sketchssm/calibration/tools/check_mamba2_hooks.py
sketchssm/calibration/tools/make_nano_layer_subset.py
sketchssm/calibration/tools/make_toy_data.py
sketchssm/calibration/tools/verify_bundles.py
sketchssm/kernels/README.md
sketchssm/kernels/__init__.py
sketchssm/kernels/_driver.py
sketchssm/kernels/_kernel.py
sketchssm/kernels/_nvrtc.py
sketchssm/kernels/_runtime.py
sketchssm/kernels/build.py
sketchssm/kernels/gdn.py
sketchssm/kernels/kda.py
sketchssm/kernels/mamba2.py
sketchssm/kernels/types.py
sketchssm/kernels/configs/gdn/num_k_heads=16,num_v_heads=16,device_name=NVIDIA_B300_SXM6_AC.json
sketchssm/kernels/configs/gdn/num_k_heads=16,num_v_heads=16,device_name=NVIDIA_H100_80GB_HBM3.json
sketchssm/kernels/configs/gdn/num_k_heads=16,num_v_heads=32,device_name=NVIDIA_B300_SXM6_AC.json
sketchssm/kernels/configs/gdn/num_k_heads=16,num_v_heads=32,device_name=NVIDIA_H100_80GB_HBM3.json
sketchssm/kernels/configs/gdn/num_k_heads=16,num_v_heads=48,device_name=NVIDIA_B300_SXM6_AC.json
sketchssm/kernels/configs/gdn/num_k_heads=16,num_v_heads=48,device_name=NVIDIA_H100_80GB_HBM3.json
sketchssm/kernels/configs/kda/kda,num_heads=64,device_name=NVIDIA_B300_SXM6_AC.json
sketchssm/kernels/configs/kda/kda,num_heads=64,device_name=NVIDIA_H100_80GB_HBM3.json
sketchssm/kernels/configs/mamba2/head_dim=128,heads_per_group=16,state_size=256,device_name=NVIDIA_B300_SXM6_AC.json
sketchssm/kernels/configs/mamba2/head_dim=128,heads_per_group=16,state_size=256,device_name=NVIDIA_H100_80GB_HBM3.json
sketchssm/kernels/configs/mamba2/head_dim=128,heads_per_group=24,state_size=256,device_name=NVIDIA_B300_SXM6_AC.json
sketchssm/kernels/configs/mamba2/head_dim=128,heads_per_group=24,state_size=256,device_name=NVIDIA_H100_80GB_HBM3.json
sketchssm/kernels/configs/mamba2/head_dim=64,heads_per_group=80,state_size=64,device_name=NVIDIA_B300_SXM6_AC.json
sketchssm/kernels/configs/mamba2/head_dim=64,heads_per_group=80,state_size=64,device_name=NVIDIA_H100_80GB_HBM3.json
sketchssm/kernels/configs/mamba2/head_dim=80,heads_per_group=16,device_name=NVIDIA_B300_SXM6_AC.json
sketchssm/kernels/configs/mamba2/head_dim=80,heads_per_group=16,device_name=NVIDIA_H100_80GB_HBM3.json
sketchssm/kernels/configs/mamba2/head_dim=80,heads_per_group=16,window=32,device_name=NVIDIA_H100_80GB_HBM3.json
sketchssm/kernels/configs/mamba2/head_dim=80,heads_per_group=16,window=64,device_name=NVIDIA_H100_80GB_HBM3.json
sketchssm/kernels/csrc/gdn/gdn_sketch_flush.cuh
sketchssm/kernels/csrc/gdn/gdn_sketch_step.cuh
sketchssm/kernels/csrc/include/array
sketchssm/kernels/csrc/include/map
sketchssm/kernels/csrc/include/stdint.h
sketchssm/kernels/csrc/include/stdlib.h
sketchssm/kernels/csrc/include/type_traits
sketchssm/kernels/csrc/include/unordered_map
sketchssm/kernels/csrc/include/ATen/cuda/CUDAContext.h
sketchssm/kernels/csrc/include/c10/cuda/CUDAException.h
sketchssm/kernels/csrc/include/torch/extension.h
sketchssm/kernels/csrc/kda/kda_sketch_flush.cuh
sketchssm/kernels/csrc/kda/kda_sketch_step.cuh
sketchssm/kernels/csrc/mamba2/mamba2_sketch_flush.cuh
sketchssm/kernels/csrc/mamba2/mamba2_sketch_nf.cuh
sketchssm/kernels/docs/PROVENANCE_gdn.md
sketchssm/kernels/docs/PROVENANCE_kda.md
sketchssm/kernels/docs/PROVENANCE_mamba2.md
sketchssm/kernels/tools/__init__.py
sketchssm/kernels/tools/benchmark_gdn.py
sketchssm/kernels/tools/benchmark_kda.py
sketchssm/kernels/tools/benchmark_mamba2.py
sketchssm/kernels/tools/benchmark_utils.py
sketchssm/kernels/tools/tune_gdn.py
sketchssm/kernels/tools/tune_kda.py
sketchssm/kernels/tools/tune_mamba2.py
tests/test_adapters.py
tests/test_bundles.py
tests/test_calibration.py
tests/test_collection.py
tests/test_portable_calibration.py
tests/test_run.py