CHANGELOG.md
CITATION.cff
LICENSE
MANIFEST.in
NOTICE
README.md
pyproject.toml
mtplx/__init__.py
mtplx/adaptive.py
mtplx/app_settings.py
mtplx/artifacts.py
mtplx/attention_context.py
mtplx/attention_split.py
mtplx/block_attention.py
mtplx/cache_state.py
mtplx/chat_encoding.py
mtplx/cli.py
mtplx/compressed_tensors.py
mtplx/config.py
mtplx/constants.py
mtplx/daemon_client.py
mtplx/deepseek_mtp_patch.py
mtplx/default_models.py
mtplx/diagnostics.py
mtplx/draft_lm_head.py
mtplx/draft_sampling.py
mtplx/engine_session.py
mtplx/env.py
mtplx/errors.py
mtplx/expert_layout.py
mtplx/fan_mode.py
mtplx/fast_sampling.py
mtplx/gdn_capture.py
mtplx/gemma4_pair.py
mtplx/generation.py
mtplx/glm_mtp_patch.py
mtplx/graphbank.py
mtplx/hardware.py
mtplx/hf_loader.py
mtplx/kv_quant.py
mtplx/mimo_mtp_patch.py
mtplx/model_catalog.py
mtplx/model_scheduler.py
mtplx/mtp_activation_stats.py
mtplx/mtp_adapters.py
mtplx/mtp_patch.py
mtplx/native_mlp.py
mtplx/nemotron_h_mtp_patch.py
mtplx/opencode.py
mtplx/pi.py
mtplx/prefill_bench.py
mtplx/profiles.py
mtplx/proposal_reranker.py
mtplx/reasoning_codecs.py
mtplx/runtime.py
mtplx/runtime_options.py
mtplx/sampling.py
mtplx/server_urls.py
mtplx/session_bank.py
mtplx/speculative.py
mtplx/step3p5_mtp_patch.py
mtplx/swival.py
mtplx/thermal.py
mtplx/thermal_sidecar.py
mtplx/trace_parity.py
mtplx/turboquant.py
mtplx/verify_qmv.py
mtplx/version.py
mtplx.egg-info/PKG-INFO
mtplx.egg-info/SOURCES.txt
mtplx.egg-info/dependency_links.txt
mtplx.egg-info/entry_points.txt
mtplx.egg-info/requires.txt
mtplx.egg-info/top_level.txt
mtplx/backends/__init__.py
mtplx/backends/deepseek_mtp.py
mtplx/backends/descriptors.py
mtplx/backends/gemma4_assistant.py
mtplx/backends/glm_mtp.py
mtplx/backends/mimo_mtp.py
mtplx/backends/nemotron_h_mtp.py
mtplx/backends/qwen3_next.py
mtplx/backends/registry.py
mtplx/backends/step3p5_mtp.py
mtplx/batching/__init__.py
mtplx/batching/admission.py
mtplx/batching/buckets.py
mtplx/batching/scheduler.py
mtplx/batching/state.py
mtplx/benchmarks/__init__.py
mtplx/benchmarks/schema.py
mtplx/benchmarks/prompts/aime_2026.jsonl
mtplx/benchmarks/prompts/calibration_coding.jsonl
mtplx/benchmarks/prompts/default.jsonl
mtplx/benchmarks/prompts/flappy.jsonl
mtplx/benchmarks/prompts/long_code.jsonl
mtplx/benchmarks/prompts/long_code_uncapped.jsonl
mtplx/benchmarks/prompts/python_modules_long.jsonl
mtplx/benchmarks/runners/__init__.py
mtplx/benchmarks/runners/aime.py
mtplx/benchmarks/runners/batch_equivalence.py
mtplx/benchmarks/runners/capture_commit_equivalence.py
mtplx/benchmarks/runners/competitor_baselines.py
mtplx/benchmarks/runners/contract_probe.py
mtplx/benchmarks/runners/harness.py
mtplx/benchmarks/runners/mtp1_gate.py
mtplx/benchmarks/runners/mtp1_sampler_smoke.py
mtplx/benchmarks/runners/mtp_adaptive.py
mtplx/benchmarks/runners/mtp_chain_probe.py
mtplx/benchmarks/runners/mtp_depth_grid.py
mtplx/benchmarks/runners/mtp_depth_sweep.py
mtplx/benchmarks/runners/mtp_tree_probe.py
mtplx/benchmarks/runners/multi_qmv_probe.py
mtplx/benchmarks/runners/preflight.py
mtplx/benchmarks/runners/runtime_smoke.py
mtplx/benchmarks/runners/session_bank.py
mtplx/benchmarks/runners/truth.py
mtplx/benchmarks/runners/verify_profile.py
mtplx/benchmarks/runners/verify_qmm_probe.py
mtplx/benchmarks/runners/verify_ratio.py
mtplx/benchmarks/validators/__init__.py
mtplx/benchmarks/validators/aime.py
mtplx/benchmarks/validators/basic.py
mtplx/cache_bank/__init__.py
mtplx/cache_bank/codec.py
mtplx/cache_bank/cold_tier.py
mtplx/commands/__init__.py
mtplx/commands/forge.py
mtplx/commands/public.py
mtplx/correctors/__init__.py
mtplx/correctors/diagonal_affine.py
mtplx/correctors/low_rank.py
mtplx/dashboard/__init__.py
mtplx/dashboard/_static/index.html
mtplx/dashboard/_static/assets/index-DYvLRZ33.css
mtplx/dashboard/_static/assets/index-DaD2PKmU.js
mtplx/kernels/__init__.py
mtplx/kernels/copy_leaf.py
mtplx/kernels/fused_norm.py
mtplx/kernels/lm_head_topk.py
mtplx/kernels/logits_topk.py
mtplx/kernels/native_gdn_tail.py
mtplx/kernels/sdpa_2pass.py
mtplx/kernels/sdpa_2pass_paged.py
mtplx/kernels/verify_mlp_fused.py
mtplx/kpi/__init__.py
mtplx/kpi/reference_vllm.py
mtplx/kpi/runtime_kpis.py
mtplx/server/__init__.py
mtplx/server/dashboard_state.py
mtplx/server/openai.py
mtplx/server/omlx_bridge/__init__.py
mtplx/server/omlx_bridge/adapter.py
mtplx/server/omlx_bridge/thinking.py
mtplx/server/omlx_bridge/tool_calling.py
mtplx/ui/__init__.py
mtplx/ui/banner.py
mtplx/ui/chat_printer.py
mtplx/ui/download_progress.py
mtplx/ui/onboarding.py
mtplx/ui/panels.py
mtplx/ui/progress.py
scripts/agent_user_path_qa.py
scripts/aime_serve_gate.py
scripts/aime_shape_memory_bench.py
scripts/collect_mtp_activation_stats.py
scripts/collect_mtp_hidden_calib.py
scripts/combine_mtp_adapters.py
scripts/convert_step37_to_mtplx_step3p5.py
scripts/eval_mtp_corrector.py
scripts/filter_mtp_adapter.py
scripts/fresh_venv_smoke.sh
scripts/hygiene_scan.sh
scripts/install_macos.sh
scripts/install_preview_global.sh
scripts/opencode_concurrency_qa.py
scripts/phase0h_paged_verifier_exactness.py
scripts/probe_mlx_pr3026_qsdpa.py
scripts/probe_mx_compile_buckets.py
scripts/probe_paged_gqa_sdpa_routes.py
scripts/release_macos_v1.sh
scripts/run_context_degradation_diagnostics.py
scripts/serve_openai_mtplx.py
scripts/session_cache_followup_qa.py
scripts/sparkle_rehearsal_kit.sh
scripts/step_acceptance.py
scripts/step_smoke.py
scripts/train_mtp_adapter_c4.py
scripts/validate_step_mtp_injector.py
tests/test_adaptive.py
tests/test_aime_serve_gate.py
tests/test_artifacts.py
tests/test_attention_split.py
tests/test_batching_foundation.py
tests/test_cache_bank.py
tests/test_cache_state.py
tests/test_cli_parity_tools.py
tests/test_compressed_tensors.py
tests/test_config.py
tests/test_config_profile_precedence.py
tests/test_context_degradation_profiles.py
tests/test_correctors.py
tests/test_daemon_client.py
tests/test_dashboard_endpoints.py
tests/test_default_models.py
tests/test_diagnostics.py
tests/test_download_progress.py
tests/test_draft_lm_head.py
tests/test_engine_session_concurrency.py
tests/test_engine_session_env.py
tests/test_fast_sampling.py
tests/test_forge_cli.py
tests/test_generation_sustained.py
tests/test_hf_loader.py
tests/test_hygiene_scan.py
tests/test_idle_postcommit_subagent.py
tests/test_max_idle_watchdog.py
tests/test_max_lifecycle.py
tests/test_metal_memory_caps.py
tests/test_model_catalog.py
tests/test_model_scheduler.py
tests/test_mtp_activation_stats.py
tests/test_mtp_adapters.py
tests/test_mtp_depth_sweep.py
tests/test_mtp_patch.py
tests/test_no_mlx_imports.py
tests/test_omlx_bridge.py
tests/test_onboarding.py
tests/test_openai_bridge.py
tests/test_opencode.py
tests/test_persistent_replay.py
tests/test_phase0h_paged_verifier_exactness.py
tests/test_policy_fingerprint_stability.py
tests/test_postcommit_prefix_reuse.py
tests/test_postcommit_tools_plumbing.py
tests/test_postcommit_wait.py
tests/test_postcommit_wait_integration.py
tests/test_prefill_bench.py
tests/test_prefill_chunk_defaults.py
tests/test_prefill_tps_regression.py
tests/test_profiles.py
tests/test_prompt_encoding.py
tests/test_public_cli.py
tests/test_runtime_kpis.py
tests/test_sampling.py
tests/test_server_openai.py
tests/test_session_bank.py
tests/test_session_bank_env_caps.py
tests/test_step3p5_mtp_patch.py
tests/test_sustained_long_context_qa.py
tests/test_thermal.py
tests/test_thermal_sidecar.py
tests/test_tool_aware_stream_translator.py
tests/test_trace_parity.py
tests/test_turboquant.py
tests/test_turboquant_fallback.py
tests/test_ui_progress.py
tests/test_validators.py
tests/test_vllm_reference.py
vllm_metal/__init__.py
vllm_metal/metal/README.md
vllm_metal/metal/__init__.py
vllm_metal/metal/build.py
vllm_metal/metal/constants.py
vllm_metal/metal/paged_ops.cpp
vllm_metal/metal/kernels_v1/copy_blocks.metal
vllm_metal/metal/kernels_v1/float8.metal
vllm_metal/metal/kernels_v1/gather_kv_cache.metal
vllm_metal/metal/kernels_v1/kv_scale_update.metal
vllm_metal/metal/kernels_v1/pagedattention.metal
vllm_metal/metal/kernels_v1/reshape_and_cache.metal
vllm_metal/metal/kernels_v1/utils.metal
vllm_metal/metal/kernels_v2/copy_blocks.metal
vllm_metal/metal/kernels_v2/float8.metal
vllm_metal/metal/kernels_v2/gather_kv_cache.metal
vllm_metal/metal/kernels_v2/gdn_linear_attention.metal
vllm_metal/metal/kernels_v2/kv_scale_update.metal
vllm_metal/metal/kernels_v2/pagedattention.metal
vllm_metal/metal/kernels_v2/reshape_and_cache.metal
vllm_metal/metal/kernels_v2/turboquant.metal
vllm_metal/metal/kernels_v2/utils.metal