LICENSE
MANIFEST.in
README.md
pyproject.toml
setup.py
configs/data/kohya/kohya_style.py
configs/moss/_base/base_full.py
configs/moss/_base/base_lora.py
configs/moss/algo/algo_dora.py
configs/moss/algo/algo_full.py
configs/moss/algo/algo_loha.py
configs/moss/algo/algo_lokr.py
configs/moss/algo/algo_lora.py
configs/moss/algo/algo_oft.py
configs/moss/data/data_folder.py
configs/moss/data/data_folder_nl.py
configs/moss/data/data_jsonl.py
configs/moss/data/data_kohya.py
configs/moss/data/data_moss_shard.py
configs/moss/data/data_parquet.py
configs/moss/data/data_webdataset.py
configs/moss/parallel/parallel_ddp.py
configs/moss/parallel/parallel_fsdp.py
configs/moss/parallel/parallel_pipeline.py
configs/moss/parallel/parallel_threaded.py
configs/moss/precision/precision_bf16.py
configs/moss/precision/precision_fp16.py
configs/moss/quality/quality_lokr16.py
configs/moss/quality/quality_lora8.py
configs/moss/recipes/recipes_dora.py
configs/moss/recipes/recipes_full_24g_offload.py
configs/moss/recipes/recipes_full_fsdp_4xv100.py
configs/moss/recipes/recipes_lora_12g_stream.py
configs/moss/recipes/recipes_lora_16g_fp16.py
configs/moss/recipes/recipes_lora_24g_bf16.py
configs/moss/recipes/recipes_lora_ddp.py
configs/moss/recipes/recipes_lora_xpu_b50.py
configs/moss/recipes/recipes_lycoris_loha.py
configs/moss/recipes/recipes_lycoris_lokr.py
configs/moss/recipes/recipes_lycoris_oft.py
configs/moss/scale/scale_full_fp16.py
configs/moss/scale/scale_full_fp16_offload.py
configs/moss/scale/scale_lora_bf16.py
configs/moss/scale/scale_lora_fp16.py
configs/moss/vram/vram_12g.py
configs/moss/vram/vram_16g.py
configs/moss/vram/vram_24g.py
configs/moss/vram/vram_24g_full.py
configs/moss/vram/vram_6g.py
configs/moss/vram/vram_8g.py
configs/moss/vram/vram_auto.py
configs/sample/_base/base_sample.py
configs/sample/precision/precision_bf16.py
configs/sample/precision/precision_fp16.py
scripts/bench/adapters/fold_base.py
scripts/bench/data/catalog.py
scripts/bench/data/loader.py
scripts/bench/infer/memory.py
scripts/bench/kernel/attention.py
scripts/bench/kernel/ffn.py
scripts/bench/kernel/fp8.py
scripts/bench/kernel/fused_linear.py
scripts/bench/kernel/norm_mod.py
scripts/bench/kernel/optim.py
scripts/bench/kernel/qk_rope.py
scripts/bench/kernel/quant.py
scripts/bench/kernel/residual.py
scripts/bench/matrix/host.py
scripts/bench/matrix/launch.py
scripts/bench/matrix/pyproject.toml
scripts/bench/matrix/src/kohakumoss_bench/__init__.py
scripts/bench/matrix/src/kohakumoss_bench/measure/__init__.py
scripts/bench/matrix/src/kohakumoss_bench/measure/ncu.py
scripts/bench/matrix/src/kohakumoss_bench/measure/probe.py
scripts/bench/matrix/src/kohakumoss_bench/measure/smi.py
scripts/bench/matrix/src/kohakumoss_bench/plan/__init__.py
scripts/bench/matrix/src/kohakumoss_bench/plan/config.py
scripts/bench/matrix/src/kohakumoss_bench/plan/coverage.py
scripts/bench/matrix/src/kohakumoss_bench/plan/coverage_runs.py
scripts/bench/matrix/src/kohakumoss_bench/plan/dry_run.py
scripts/bench/matrix/src/kohakumoss_bench/plan/hostmem.py
scripts/bench/matrix/src/kohakumoss_bench/plan/knobs.py
scripts/bench/matrix/src/kohakumoss_bench/plan/matrix.py
scripts/bench/matrix/src/kohakumoss_bench/plan/prompts.py
scripts/bench/matrix/src/kohakumoss_bench/report/__init__.py
scripts/bench/matrix/src/kohakumoss_bench/report/build.py
scripts/bench/matrix/src/kohakumoss_bench/report/cells.py
scripts/bench/matrix/src/kohakumoss_bench/report/results.py
scripts/bench/matrix/src/kohakumoss_bench/report/stage_tables.py
scripts/bench/matrix/src/kohakumoss_bench/report/tables.py
scripts/bench/matrix/src/kohakumoss_bench/run/__init__.py
scripts/bench/matrix/src/kohakumoss_bench/run/child.py
scripts/bench/matrix/src/kohakumoss_bench/run/disk.py
scripts/bench/matrix/src/kohakumoss_bench/run/filters.py
scripts/bench/matrix/src/kohakumoss_bench/run/host.py
scripts/bench/matrix/src/kohakumoss_bench/run/ncu_pass.py
scripts/bench/matrix/src/kohakumoss_bench/run/pool.py
scripts/bench/matrix/src/kohakumoss_bench/run/proc.py
scripts/bench/matrix/src/kohakumoss_bench/run/release.py
scripts/bench/matrix/src/kohakumoss_bench/run/remote.py
scripts/bench/matrix/src/kohakumoss_bench/run/stage.py
scripts/bench/offload/knobs.py
scripts/bench/offload/limits.py
scripts/bench/offload/stream.py
scripts/bench/step/step.py
scripts/bench/textenc/ops.py
scripts/bench/textenc/quant.py
scripts/bench/textenc/textenc.py
scripts/bench/train/memory.py
scripts/bench/tune/tune.py
scripts/bench/vae/vae.py
scripts/colab/kohakumoss_colab.ipynb
scripts/colab/setup.py
scripts/data/cache_latents.py
scripts/data/cache_merge.py
scripts/data/cache_text.py
scripts/data/catalog_build.py
scripts/data/catalog_cache.py
scripts/data/inspect_data.py
scripts/dist/check_comm.py
scripts/dist/scaling_report.py
scripts/doctor/doctor.py
scripts/examples/README.md
scripts/examples/grid.ps1
scripts/examples/grid.sh
scripts/examples/grid_prompts.py
scripts/examples/sample.ps1
scripts/examples/sample.sh
scripts/examples/train.ps1
scripts/examples/train.sh
scripts/export/export.py
scripts/export/fp16_fix.py
scripts/export/fp8_quant.py
scripts/export/merge.py
scripts/export/text_quant.py
scripts/install/install.ps1
scripts/install/install.sh
scripts/probe/activations.py
scripts/probe/adapter_e2e.py
scripts/probe/backward.py
scripts/probe/embed.py
scripts/probe/fp8_inference.py
scripts/probe/parity.py
scripts/probe/trajectory.py
scripts/sample/compare_grid.py
scripts/sample/sample.py
scripts/train/moss.py
src/kohakumoss/__init__.py
src/kohakumoss.egg-info/PKG-INFO
src/kohakumoss.egg-info/SOURCES.txt
src/kohakumoss.egg-info/dependency_links.txt
src/kohakumoss.egg-info/entry_points.txt
src/kohakumoss.egg-info/requires.txt
src/kohakumoss.egg-info/top_level.txt
src/kohakumoss/adapters/__init__.py
src/kohakumoss/adapters/bind/__init__.py
src/kohakumoss/adapters/bind/autotuned.py
src/kohakumoss/adapters/bind/bodies.py
src/kohakumoss/adapters/bind/compose.py
src/kohakumoss/adapters/bind/context.py
src/kohakumoss/adapters/bind/dora.py
src/kohakumoss/adapters/bind/eager.py
src/kohakumoss/adapters/bind/forwards.py
src/kohakumoss/adapters/bind/kernel_ops.py
src/kohakumoss/adapters/bind/ops.py
src/kohakumoss/adapters/bind/select.py
src/kohakumoss/adapters/bind/weight_ops.py
src/kohakumoss/adapters/build/__init__.py
src/kohakumoss/adapters/build/apply.py
src/kohakumoss/adapters/build/config.py
src/kohakumoss/adapters/build/network.py
src/kohakumoss/adapters/build/presets.py
src/kohakumoss/adapters/build/target.py
src/kohakumoss/adapters/export/__init__.py
src/kohakumoss/adapters/export/cli.py
src/kohakumoss/adapters/export/convert.py
src/kohakumoss/adapters/export/files.py
src/kohakumoss/adapters/export/fused.py
src/kohakumoss/adapters/export/merged.py
src/kohakumoss/adapters/infer/__init__.py
src/kohakumoss/adapters/infer/formats.py
src/kohakumoss/adapters/infer/inference.py
src/kohakumoss/adapters/infer/stored.py
src/kohakumoss/adapters/kinds/__init__.py
src/kohakumoss/adapters/kinds/common.py
src/kohakumoss/adapters/kinds/lowrank.py
src/kohakumoss/adapters/kinds/other.py
src/kohakumoss/adapters/kinds/record.py
src/kohakumoss/adapters/kinds/rotation.py
src/kohakumoss/adapters/lycoris/__init__.py
src/kohakumoss/adapters/lycoris/backend.py
src/kohakumoss/adapters/lycoris/tuning.py
src/kohakumoss/adapters/lycoris/version.py
src/kohakumoss/adapters/train/__init__.py
src/kohakumoss/adapters/train/adapter_set.py
src/kohakumoss/adapters/train/groups.py
src/kohakumoss/adapters/train/norms.py
src/kohakumoss/adapters/train/scope.py
src/kohakumoss/adapters/train/steps.py
src/kohakumoss/adapters/train/tlora.py
src/kohakumoss/cli/__init__.py
src/kohakumoss/cli/doctor.py
src/kohakumoss/cli/missing.py
src/kohakumoss/cli/tune.py
src/kohakumoss/cli/config/__init__.py
src/kohakumoss/cli/config/full.py
src/kohakumoss/cli/config/hardware.py
src/kohakumoss/cli/config/main.py
src/kohakumoss/cli/config/presets.py
src/kohakumoss/cli/config/starter.py
src/kohakumoss/cli/config/validate.py
src/kohakumoss/diffusion/__init__.py
src/kohakumoss/diffusion/objective/__init__.py
src/kohakumoss/diffusion/objective/flow.py
src/kohakumoss/diffusion/objective/loss.py
src/kohakumoss/diffusion/objective/normalize.py
src/kohakumoss/diffusion/objective/weighting.py
src/kohakumoss/diffusion/sampling/__init__.py
src/kohakumoss/diffusion/sampling/sample.py
src/kohakumoss/diffusion/sampling/schedule.py
src/kohakumoss/diffusion/timestep/__init__.py
src/kohakumoss/diffusion/timestep/moss.py
src/kohakumoss/diffusion/timestep/shift.py
src/kohakumoss/diffusion/timestep/simple.py
src/kohakumoss/encoders/__init__.py
src/kohakumoss/encoders/knob_names.py
src/kohakumoss/encoders/specs.py
src/kohakumoss/encoders/vae.py
src/kohakumoss/encoders/text/__init__.py
src/kohakumoss/encoders/text/encoder.py
src/kohakumoss/encoders/text/factory.py
src/kohakumoss/encoders/text/forward.py
src/kohakumoss/encoders/text/hf.py
src/kohakumoss/encoders/text/plan.py
src/kohakumoss/encoders/text/prewarm.py
src/kohakumoss/encoders/text/tokens.py
src/kohakumoss/inference/__init__.py
src/kohakumoss/inference/components.py
src/kohakumoss/inference/grid.py
src/kohakumoss/inference/knobs.py
src/kohakumoss/inference/loading.py
src/kohakumoss/inference/outputs.py
src/kohakumoss/inference/peaks.py
src/kohakumoss/inference/pipeline.py
src/kohakumoss/inference/placement_report.py
src/kohakumoss/inference/steplog.py
src/kohakumoss/inference/warm.py
src/kohakumoss/logging/__init__.py
src/kohakumoss/logging/log.py
src/kohakumoss/logging/sinks.py
src/kohakumoss/memory/__init__.py
src/kohakumoss/memory/apply.py
src/kohakumoss/memory/catalog_table.py
src/kohakumoss/memory/device_use.py
src/kohakumoss/memory/estimate.py
src/kohakumoss/memory/knob_map.py
src/kohakumoss/memory/planner.py
src/kohakumoss/memory/profiles.py
src/kohakumoss/memory/release.py
src/kohakumoss/memory/runners.py
src/kohakumoss/memory/timeline.py
src/kohakumoss/model/__init__.py
src/kohakumoss/model/backbone.py
src/kohakumoss/model/block.py
src/kohakumoss/model/build.py
src/kohakumoss/model/flops.py
src/kohakumoss/model/linear.py
src/kohakumoss/model/opset.py
src/kohakumoss/model/runner.py
src/kohakumoss/model/stream.py
src/kohakumoss/model/text_block.py
src/kohakumoss/model/components/__init__.py
src/kohakumoss/model/components/embed.py
src/kohakumoss/model/components/ffn.py
src/kohakumoss/model/components/norm.py
src/kohakumoss/model/io/__init__.py
src/kohakumoss/model/io/keymap.py
src/kohakumoss/model/io/load.py
src/kohakumoss/model/io/save.py
src/kohakumoss/model/io/source.py
src/kohakumoss/model/numerics/__init__.py
src/kohakumoss/model/numerics/fold.py
src/kohakumoss/model/numerics/fp16fix.py
src/kohakumoss/model/numerics/keep_fp32.py
src/kohakumoss/model/numerics/saturation.py
src/kohakumoss/offload/__init__.py
src/kohakumoss/offload/accel.py
src/kohakumoss/offload/activation.py
src/kohakumoss/offload/worker.py
src/kohakumoss/offload/encoders/__init__.py
src/kohakumoss/offload/encoders/placement.py
src/kohakumoss/offload/encoders/report.py
src/kohakumoss/offload/encoders/text.py
src/kohakumoss/offload/encoders/text_memory.py
src/kohakumoss/offload/encoders/vae.py
src/kohakumoss/offload/optim/__init__.py
src/kohakumoss/offload/optim/cpu_optim.py
src/kohakumoss/offload/optim/fused_step.py
src/kohakumoss/offload/optim/gpu_adamw.py
src/kohakumoss/offload/optim/grad.py
src/kohakumoss/offload/optim/host_acc.py
src/kohakumoss/offload/optim/stream_optim.py
src/kohakumoss/offload/weights/__init__.py
src/kohakumoss/offload/weights/layers.py
src/kohakumoss/offload/weights/shared.py
src/kohakumoss/offload/weights/stream.py
src/kohakumoss/offload/weights/streamed_weights.py
src/kohakumoss/offload/weights/streaming/__init__.py
src/kohakumoss/offload/weights/streaming/copier.py
src/kohakumoss/offload/weights/streaming/device.py
src/kohakumoss/offload/weights/streaming/events.py
src/kohakumoss/offload/weights/streaming/grads.py
src/kohakumoss/offload/weights/streaming/reduce.py
src/kohakumoss/offload/weights/streaming/ring.py
src/kohakumoss/offload/weights/streaming/timing.py
src/kohakumoss/parallel/__init__.py
src/kohakumoss/parallel/ddp.py
src/kohakumoss/parallel/fsdp.py
src/kohakumoss/parallel/launch.py
src/kohakumoss/parallel/pipeline/__init__.py
src/kohakumoss/parallel/pipeline/balance.py
src/kohakumoss/parallel/pipeline/distributed.py
src/kohakumoss/parallel/pipeline/parts.py
src/kohakumoss/parallel/pipeline/split.py
src/kohakumoss/parallel/pipeline/threaded.py
src/kohakumoss/parallel/telemetry/__init__.py
src/kohakumoss/parallel/telemetry/report.py
src/kohakumoss/parallel/telemetry/selftest.py
src/kohakumoss/parallel/telemetry/timeline.py
src/kohakumoss/parallel/telemetry/timing.py
src/kohakumoss/parallel/telemetry/trace.py
src/kohakumoss/precision/__init__.py
src/kohakumoss/precision/config.py
src/kohakumoss/precision/scaler.py
src/kohakumoss/probe/__init__.py
src/kohakumoss/probe/official.py
src/kohakumoss/quant/__init__.py
src/kohakumoss/quant/bind.py
src/kohakumoss/quant/comfy.py
src/kohakumoss/quant/compute.py
src/kohakumoss/quant/convert.py
src/kohakumoss/quant/stored.py
src/kohakumoss/quant/stored_text.py
src/kohakumoss/quant/text.py
src/kohakumoss/quant/weights.py
src/kohakumoss/training/__init__.py
src/kohakumoss/training/context.py
src/kohakumoss/training/entry.py
src/kohakumoss/training/full.py
src/kohakumoss/training/knobs.py
src/kohakumoss/training/layouts.py
src/kohakumoss/training/shutdown.py
src/kohakumoss/training/state.py
src/kohakumoss/training/step.py
src/kohakumoss/training/build/__init__.py
src/kohakumoss/training/build/kernels.py
src/kohakumoss/training/build/loop.py
src/kohakumoss/training/build/model.py
src/kohakumoss/training/build/objective.py
src/kohakumoss/training/build/parallel.py
src/kohakumoss/training/build/pipelines.py
src/kohakumoss/training/build/shapes.py
src/kohakumoss/training/build/text_estimate.py
src/kohakumoss/training/callbacks/__init__.py
src/kohakumoss/training/callbacks/evaluate.py
src/kohakumoss/training/callbacks/finite.py
src/kohakumoss/training/callbacks/kernels.py
src/kohakumoss/training/callbacks/monitor.py
src/kohakumoss/training/callbacks/progress.py
src/kohakumoss/training/callbacks/repro.py
src/kohakumoss/training/callbacks/runner.py
src/kohakumoss/training/callbacks/sampler.py
src/kohakumoss/training/callbacks/startup.py
src/kohakumoss/training/callbacks/transfer.py
src/kohakumoss/training/checkpoint/__init__.py
src/kohakumoss/training/checkpoint/io.py
src/kohakumoss/training/checkpoint/manager.py
src/kohakumoss/training/checkpoint/stream.py
src/kohakumoss/training/checks/__init__.py
src/kohakumoss/training/checks/encoders.py
src/kohakumoss/training/checks/objective.py
src/kohakumoss/training/checks/outputs.py
src/kohakumoss/training/checks/parallel.py
src/kohakumoss/training/drivers/__init__.py
src/kohakumoss/training/drivers/common.py
src/kohakumoss/training/drivers/fsdp.py
src/kohakumoss/training/drivers/lightning.py
src/kohakumoss/training/drivers/native.py
src/kohakumoss/training/drivers/pipeline.py
src/kohakumoss/training/ema/__init__.py
src/kohakumoss/training/ema/averagers.py
src/kohakumoss/training/ema/posthoc.py
src/kohakumoss/training/memory/__init__.py
src/kohakumoss/training/memory/optimizers.py
src/kohakumoss/training/memory/session.py
src/kohakumoss/training/optim/__init__.py
src/kohakumoss/training/optim/groups.py
src/kohakumoss/training/optim/schedule.py
src/kohakumoss/training/optim/bundle/__init__.py
src/kohakumoss/training/optim/bundle/base.py
src/kohakumoss/training/optim/bundle/build.py
src/kohakumoss/training/optim/bundle/fused.py
src/kohakumoss/training/optim/bundle/mirror.py
src/kohakumoss/training/optim/bundle/nonfinite.py
src/kohakumoss/training/optim/optimizers/__init__.py
src/kohakumoss/training/optim/optimizers/adamw16.py
src/kohakumoss/training/optim/optimizers/adamw8bit.py
src/kohakumoss/training/optim/optimizers/factory.py
src/kohakumoss/training/optim/optimizers/kwargs.py
src/kohakumoss/training/optim/optimizers/muon.py
src/kohakumoss/training/optim/optimizers/rounding.py
src/kohakumoss/training/optim/optimizers/skip.py
src/kohakumoss/training/textenc/__init__.py
src/kohakumoss/training/textenc/ahead.py
src/kohakumoss/training/vae/__init__.py
src/kohakumoss/training/vae/ahead.py
src/kohakumoss/tuning/__init__.py
src/kohakumoss/tuning/lengths.py
src/kohakumoss/tuning/prewarm.py
src/kohakumoss/types/__init__.py
src/kohakumoss/types/arch.py
src/kohakumoss/types/batch.py
src/kohakumoss/types/context.py
src/kohakumoss/types/plans.py
src/kohakumoss/types/protocols.py
src/kohakumoss/types/seqinfo.py
src/kohakumoss/types/step.py
src/kohakumoss_common/__init__.py
src/kohakumoss_common/calls.py
src/kohakumoss_common/devices.py
src/kohakumoss_common/imports.py
src/kohakumoss_common/io.py
src/kohakumoss_common/knobs.py
src/kohakumoss_common/log.py
src/kohakumoss_common/markdown.py
src/kohakumoss_common/registry.py
src/kohakumoss_common/units.py
src/kohakumoss_data/__init__.py
src/kohakumoss_data/config.py
src/kohakumoss_data/dataset.py
src/kohakumoss_data/importer.py
src/kohakumoss_data/registry.py
src/kohakumoss_data/scripting.py
src/kohakumoss_data/seeds.py
src/kohakumoss_data/stats.py
src/kohakumoss_data/types.py
src/kohakumoss_data/cache/__init__.py
src/kohakumoss_data/cache/build.py
src/kohakumoss_data/cache/keys.py
src/kohakumoss_data/cache/settings.py
src/kohakumoss_data/cache/store.py
src/kohakumoss_data/cache/tensor_cache.py
src/kohakumoss_data/cache/writer.py
src/kohakumoss_data/caption/__init__.py
src/kohakumoss_data/caption/json_text.py
src/kohakumoss_data/caption/tags.py
src/kohakumoss_data/caption/template.py
src/kohakumoss_data/catalog/__init__.py
src/kohakumoss_data/catalog/table.py
src/kohakumoss_data/catalog/text.py
src/kohakumoss_data/catalog/vault.py
src/kohakumoss_data/format/__init__.py
src/kohakumoss_data/format/catalog.py
src/kohakumoss_data/format/header.py
src/kohakumoss_data/format/shard.py
src/kohakumoss_data/loader/__init__.py
src/kohakumoss_data/loader/decode.py
src/kohakumoss_data/loader/stateless.py
src/kohakumoss_data/resolution/__init__.py
src/kohakumoss_data/resolution/base.py
src/kohakumoss_data/resolution/batching.py
src/kohakumoss_data/resolution/bucket.py
src/kohakumoss_data/resolution/geometry.py
src/kohakumoss_data/resolution/keep.py
src/kohakumoss_data/resolution/overlong.py
src/kohakumoss_data/resolution/resize.py
src/kohakumoss_data/resolution/square.py
src/kohakumoss_data/scan/__init__.py
src/kohakumoss_data/scan/lock.py
src/kohakumoss_data/scan/manage.py
src/kohakumoss_data/scan/paths.py
src/kohakumoss_data/scan/runner.py
src/kohakumoss_data/scan/walk.py
src/kohakumoss_data/sources/__init__.py
src/kohakumoss_data/sources/base.py
src/kohakumoss_data/sources/folder.py
src/kohakumoss_data/sources/jsonl.py
src/kohakumoss_data/sources/kohya.py
src/kohakumoss_data/sources/moss_shard.py
src/kohakumoss_data/sources/parquet.py
src/kohakumoss_data/sources/webdataset.py
src/kohakumoss_kernels/__init__.py
src/kohakumoss_kernels/cache_tag.py
src/kohakumoss_kernels/constraints.py
src/kohakumoss_kernels/dtypes.py
src/kohakumoss_kernels/error_stats.py
src/kohakumoss_kernels/errors.py
src/kohakumoss_kernels/ops.py
src/kohakumoss_kernels/recording.py
src/kohakumoss_kernels/registry.py
src/kohakumoss_kernels/run_state.py
src/kohakumoss_kernels/saved_ops.py
src/kohakumoss_kernels/status.py
src/kohakumoss_kernels/backends/__init__.py
src/kohakumoss_kernels/backends/_fused_linear.py
src/kohakumoss_kernels/backends/_master.py
src/kohakumoss_kernels/backends/_memo.py
src/kohakumoss_kernels/backends/varlen_choice.py
src/kohakumoss_kernels/backends/compile/__init__.py
src/kohakumoss_kernels/backends/compile/choice.py
src/kohakumoss_kernels/backends/compile/ops.py
src/kohakumoss_kernels/backends/compile/prewarm.py
src/kohakumoss_kernels/backends/eager/__init__.py
src/kohakumoss_kernels/backends/eager/_acc.py
src/kohakumoss_kernels/backends/eager/attention.py
src/kohakumoss_kernels/backends/eager/ffn.py
src/kohakumoss_kernels/backends/eager/loss.py
src/kohakumoss_kernels/backends/eager/norm.py
src/kohakumoss_kernels/backends/eager/rope.py
src/kohakumoss_kernels/backends/eager/textenc.py
src/kohakumoss_kernels/backends/tilelang/__init__.py
src/kohakumoss_kernels/backends/tilelang/optim.py
src/kohakumoss_kernels/backends/tilelang/quant.py
src/kohakumoss_kernels/backends/tilelang/attention/__init__.py
src/kohakumoss_kernels/backends/tilelang/attention/_kernels.py
src/kohakumoss_kernels/backends/tilelang/attention/entry.py
src/kohakumoss_kernels/backends/tilelang/attention/varlen.py
src/kohakumoss_kernels/backends/tilelang/blocks/__init__.py
src/kohakumoss_kernels/backends/tilelang/blocks/_blockops.py
src/kohakumoss_kernels/backends/tilelang/blocks/_gemm_prims.py
src/kohakumoss_kernels/backends/tilelang/blocks/ffn.py
src/kohakumoss_kernels/backends/tilelang/blocks/fused_linear.py
src/kohakumoss_kernels/backends/tilelang/blocks/loss.py
src/kohakumoss_kernels/backends/tilelang/blocks/norm.py
src/kohakumoss_kernels/backends/tilelang/blocks/residual.py
src/kohakumoss_kernels/backends/tilelang/blocks/rope.py
src/kohakumoss_kernels/backends/tilelang/textenc/__init__.py
src/kohakumoss_kernels/backends/tilelang/textenc/_kernels.py
src/kohakumoss_kernels/backends/tilelang/textenc/ops.py
src/kohakumoss_kernels/backends/triton/__init__.py
src/kohakumoss_kernels/backends/triton/_common.py
src/kohakumoss_kernels/backends/triton/_status.py
src/kohakumoss_kernels/backends/triton/quant.py
src/kohakumoss_kernels/backends/triton/attention/__init__.py
src/kohakumoss_kernels/backends/triton/attention/_kernels.py
src/kohakumoss_kernels/backends/triton/attention/entry.py
src/kohakumoss_kernels/backends/triton/attention/varlen.py
src/kohakumoss_kernels/backends/triton/blocks/__init__.py
src/kohakumoss_kernels/backends/triton/blocks/_blockops.py
src/kohakumoss_kernels/backends/triton/blocks/_gemm_kernels.py
src/kohakumoss_kernels/backends/triton/blocks/ffn.py
src/kohakumoss_kernels/backends/triton/blocks/fused_linear.py
src/kohakumoss_kernels/backends/triton/blocks/loss.py
src/kohakumoss_kernels/backends/triton/blocks/norm.py
src/kohakumoss_kernels/backends/triton/blocks/residual.py
src/kohakumoss_kernels/backends/triton/blocks/rope.py
src/kohakumoss_kernels/backends/triton/fp8/__init__.py
src/kohakumoss_kernels/backends/triton/fp8/fused.py
src/kohakumoss_kernels/backends/triton/fp8/gemm.py
src/kohakumoss_kernels/backends/triton/fp8/linear.py
src/kohakumoss_kernels/backends/triton/fp8/quantize.py
src/kohakumoss_kernels/backends/triton/optim/__init__.py
src/kohakumoss_kernels/backends/triton/optim/adamw.py
src/kohakumoss_kernels/backends/triton/optim/multi.py
src/kohakumoss_kernels/backends/triton/textenc/__init__.py
src/kohakumoss_kernels/backends/triton/textenc/_attention_kernels.py
src/kohakumoss_kernels/backends/triton/textenc/_delta_kernels.py
src/kohakumoss_kernels/backends/triton/textenc/_delta_short.py
src/kohakumoss_kernels/backends/triton/textenc/attention.py
src/kohakumoss_kernels/backends/triton/textenc/delta.py
src/kohakumoss_kernels/bench/__init__.py
src/kohakumoss_kernels/bench/blockops.py
src/kohakumoss_kernels/fp8/__init__.py
src/kohakumoss_kernels/fp8/formats.py
src/kohakumoss_kernels/fp8/op.py
src/kohakumoss_kernels/fp8/train.py
src/kohakumoss_kernels/hardware/__init__.py
src/kohakumoss_kernels/hardware/arch.py
src/kohakumoss_kernels/hardware/current.py
src/kohakumoss_kernels/hardware/device.py
src/kohakumoss_kernels/hardware/memory.py
src/kohakumoss_kernels/hardware/toolchain.py
src/kohakumoss_kernels/hardware/xpu_triton.py
src/kohakumoss_kernels/optim/__init__.py
src/kohakumoss_kernels/optim/choice.py
src/kohakumoss_kernels/optim/reference.py
src/kohakumoss_kernels/quant/__init__.py
src/kohakumoss_kernels/tuning/__init__.py
src/kohakumoss_kernels/tuning/buckets.py
src/kohakumoss_kernels/tuning/decisions.py
src/kohakumoss_kernels/tuning/manifest.py
src/kohakumoss_kernels/tuning/plan.py
src/kohakumoss_kernels/tuning/prewarm.py
src/kohakumoss_kernels/tuning/replay.py
src/kohakumoss_kernels/tuning/timing.py
src/kohakumoss_kernels/tuning/tuner.py
src/kohakumoss_kernels/tuning/window.py