CHANGELOG.md
CONTRIBUTING.md
LICENSE
LICENSE-MIT
MANIFEST.in
README.md
pyproject.toml
bench/README.md
bench/example.json
bench/merge.py
bench/plot-bench.py
bench/serve-bench.py
bench/patches/ds4-ignore-eos.patch
docs/README.md
docs/adding-architectures.md
docs/arch-coverage.md
docs/assistant.md
docs/benchmarks.json
docs/benchmarks.md
docs/cli.md
docs/getting-started.md
docs/launch.md
docs/lora.md
docs/migrating.md
docs/performance.md
docs/python.md
docs/rag.md
docs/server-config.md
docs/serving-architecture.md
docs/speculative-batching.md
docs/streaming.md
docs/talk.md
docs/testing.md
docs/troubleshooting.md
docs/upstream-upgrades.md
docs/vlm.md
docs/assets/demo.gif
docs/assets/record-demo.sh
docs/assets/perf/fleet-ratio-dark.svg
docs/assets/perf/fleet-ratio.svg
docs/assets/perf/kimi-k3-ud-iq2xxs-car.html
docs/assets/perf/kimi-k3-ud-iq2xxs-car.png
docs/assets/perf/kimi-k3-ud-q2kxl-car.html
docs/assets/perf/kimi-k3-ud-q2kxl-car.png
docs/assets/perf/lossy-hy3-baseline.html
docs/assets/perf/lossy-hy3-baseline.png
docs/assets/perf/lossy-hy3-shed-0.07-0.93.html
docs/assets/perf/lossy-hy3-shed-0.07-0.93.png
docs/assets/perf/lossy-hy3-shed-0.10-0.90-cool.html
docs/assets/perf/lossy-hy3-shed-0.10-0.90-cool.png
docs/assets/perf/lossy-hy3-shed-0.10-0.90.html
docs/assets/perf/lossy-hy3-shed-0.10-0.90.png
docs/assets/perf/lossy-hy3-shed-0.20-0.80.html
docs/assets/perf/lossy-hy3-shed-0.20-0.80.png
docs/assets/perf/lossy-kimi-k3-shed-0.65.html
docs/assets/perf/lossy-kimi-k3-shed-0.65.png
docs/assets/perf/lossy-kimi-k3-shed-0.70.html
docs/assets/perf/lossy-kimi-k3-shed-0.70.png
docs/assets/perf/lossy-kimi-k3-shed-0.80.html
docs/assets/perf/lossy-kimi-k3-shed-0.80.png
docs/assets/perf/mtp-lift-dark.svg
docs/assets/perf/mtp-lift.svg
docs/assets/perf/per-model/deepseek-v4-flash-antirez-iq2xxs-panels-dark.svg
docs/assets/perf/per-model/deepseek-v4-flash-antirez-iq2xxs-panels.svg
docs/assets/perf/per-model/deepseek-v4-flash-unsloth-udiq3xxs-panels-dark.svg
docs/assets/perf/per-model/deepseek-v4-flash-unsloth-udiq3xxs-panels.svg
docs/assets/perf/per-model/dolphin3-llama3.1-8b-q6k-panels-dark.svg
docs/assets/perf/per-model/dolphin3-llama3.1-8b-q6k-panels.svg
docs/assets/perf/per-model/gemma-4-12b-q6k-panels-dark.svg
docs/assets/perf/per-model/gemma-4-12b-q6k-panels.svg
docs/assets/perf/per-model/gemma-4-26b-a4b-q6k-panels-dark.svg
docs/assets/perf/per-model/gemma-4-26b-a4b-q6k-panels.svg
docs/assets/perf/per-model/gemma-4-31b-q6k-panels-dark.svg
docs/assets/perf/per-model/gemma-4-31b-q6k-panels.svg
docs/assets/perf/per-model/gemma-4-e2b-q6kxl-panels-dark.svg
docs/assets/perf/per-model/gemma-4-e2b-q6kxl-panels.svg
docs/assets/perf/per-model/gemma-4-e4b-q6k-panels-dark.svg
docs/assets/perf/per-model/gemma-4-e4b-q6k-panels.svg
docs/assets/perf/per-model/gpt-oss-120b-heretic-mxfp4-panels-dark.svg
docs/assets/perf/per-model/gpt-oss-120b-heretic-mxfp4-panels.svg
docs/assets/perf/per-model/gpt-oss-20b-mxfp4-panels-dark.svg
docs/assets/perf/per-model/gpt-oss-20b-mxfp4-panels.svg
docs/assets/perf/per-model/qwen3.5-122b-a10b-ud-q5km-panels-dark.svg
docs/assets/perf/per-model/qwen3.5-122b-a10b-ud-q5km-panels.svg
docs/assets/perf/per-model/qwen3.5-9b-q6k-panels-dark.svg
docs/assets/perf/per-model/qwen3.5-9b-q6k-panels.svg
docs/assets/perf/per-model/qwen3.6-27b-q6k-panels-dark.svg
docs/assets/perf/per-model/qwen3.6-27b-q6k-panels.svg
docs/assets/perf/per-model/qwen3.6-35b-a3b-heretic-q6k-panels-dark.svg
docs/assets/perf/per-model/qwen3.6-35b-a3b-heretic-q6k-panels.svg
gmlx/__init__.py
gmlx/__main__.py
gmlx/_exitfix.py
gmlx/config.py
gmlx/envflags.py
gmlx/eval_guard.py
gmlx/spinner.py
gmlx/textfmt.py
gmlx.egg-info/PKG-INFO
gmlx.egg-info/SOURCES.txt
gmlx.egg-info/dependency_links.txt
gmlx.egg-info/entry_points.txt
gmlx.egg-info/requires.txt
gmlx.egg-info/top_level.txt
gmlx/_vendor/__init__.py
gmlx/_vendor/misaki/LICENSE
gmlx/_vendor/misaki/__init__.py
gmlx/_vendor/misaki/cutlet.py
gmlx/_vendor/misaki/en.py
gmlx/_vendor/misaki/espeak.py
gmlx/_vendor/misaki/he.py
gmlx/_vendor/misaki/ja.py
gmlx/_vendor/misaki/ko.py
gmlx/_vendor/misaki/num2kana.py
gmlx/_vendor/misaki/token.py
gmlx/_vendor/misaki/tone_sandhi.py
gmlx/_vendor/misaki/transcription.py
gmlx/_vendor/misaki/vi.py
gmlx/_vendor/misaki/zh.py
gmlx/_vendor/misaki/zh_frontend.py
gmlx/_vendor/misaki/data/__init__.py
gmlx/_vendor/misaki/data/gb_gold.json
gmlx/_vendor/misaki/data/gb_silver.json
gmlx/_vendor/misaki/data/ja_words.txt
gmlx/_vendor/misaki/data/us_gold.json
gmlx/_vendor/misaki/data/us_silver.json
gmlx/_vendor/misaki/data/vi_acronyms.json
gmlx/_vendor/misaki/data/vi_symbols.json
gmlx/_vendor/misaki/data/vi_teencode.json
gmlx/_vendor/misaki/g2pkc/__init__.py
gmlx/_vendor/misaki/g2pkc/english.py
gmlx/_vendor/misaki/g2pkc/g2pk.py
gmlx/_vendor/misaki/g2pkc/idioms.txt
gmlx/_vendor/misaki/g2pkc/numerals.py
gmlx/_vendor/misaki/g2pkc/regular.py
gmlx/_vendor/misaki/g2pkc/rules.txt
gmlx/_vendor/misaki/g2pkc/special.py
gmlx/_vendor/misaki/g2pkc/table.csv
gmlx/_vendor/misaki/g2pkc/utils.py
gmlx/_vendor/misaki/vi_cleaner/__init__.py
gmlx/_vendor/misaki/vi_cleaner/abbreviation_vi.py
gmlx/_vendor/misaki/vi_cleaner/acronym_vi.py
gmlx/_vendor/misaki/vi_cleaner/cleaner.py
gmlx/_vendor/misaki/vi_cleaner/currency_vi.py
gmlx/_vendor/misaki/vi_cleaner/datestime_vi.py
gmlx/_vendor/misaki/vi_cleaner/letter_vi.py
gmlx/_vendor/misaki/vi_cleaner/measurement_vi.py
gmlx/_vendor/misaki/vi_cleaner/num2vi.py
gmlx/_vendor/misaki/vi_cleaner/numberical_vi.py
gmlx/_vendor/misaki/vi_cleaner/passage_utils.py
gmlx/_vendor/misaki/vi_cleaner/roman_number_vi.py
gmlx/_vendor/misaki/vi_cleaner/sentence_utils.py
gmlx/_vendor/misaki/vi_cleaner/symbol_vi.py
gmlx/_vendor/misaki/zh_normalization/__init__.py
gmlx/_vendor/misaki/zh_normalization/char_convert.py
gmlx/_vendor/misaki/zh_normalization/chronology.py
gmlx/_vendor/misaki/zh_normalization/constants.py
gmlx/_vendor/misaki/zh_normalization/num.py
gmlx/_vendor/misaki/zh_normalization/phonecode.py
gmlx/_vendor/misaki/zh_normalization/quantifier.py
gmlx/_vendor/misaki/zh_normalization/text_normalization.py
gmlx/assistant/__init__.py
gmlx/assistant/brain.py
gmlx/assistant/mcp.py
gmlx/assistant/memory.py
gmlx/assistant/serve.py
gmlx/cache/__init__.py
gmlx/cache/apc_manager.py
gmlx/cache/apc_pooling.py
gmlx/cache/compat.py
gmlx/cache/fresh_gate.py
gmlx/cache/prefix_cache.py
gmlx/cache/retire_key.py
gmlx/cache/snapshot.py
gmlx/commands/__init__.py
gmlx/commands/cli.py
gmlx/commands/completion.py
gmlx/commands/doctor.py
gmlx/commands/extras.py
gmlx/commands/launch.py
gmlx/commands/manage.py
gmlx/commands/menubar.py
gmlx/commands/menubar_config.py
gmlx/commands/tool_preflight.py
gmlx/commands/train.py
gmlx/commands/wizard.py
gmlx/gen/__init__.py
gmlx/gen/benchmarks.py
gmlx/gen/diffusion.py
gmlx/gen/generation.py
gmlx/gen/prefill_decay.py
gmlx/gen/profiles.py
gmlx/gen/thinking_budget.py
gmlx/load/__init__.py
gmlx/load/adapter.py
gmlx/load/arch_table.py
gmlx/load/config_synth.py
gmlx/load/discovery.py
gmlx/load/dtypes.py
gmlx/load/gguf_meta.py
gmlx/load/headerscan.py
gmlx/load/loader.py
gmlx/load/loadlog.py
gmlx/load/memfit.py
gmlx/load/modules.py
gmlx/load/native_fp.py
gmlx/load/populate.py
gmlx/load/preflight.py
gmlx/load/remap.py
gmlx/load/remote.py
gmlx/load/tokenizer.py
gmlx/load/transforms.py
gmlx/load/vlm.py
gmlx/models/__init__.py
gmlx/models/kimi_k3.py
gmlx/models/minimax_m3.py
gmlx/models/vlm_text_only.py
gmlx/models/deepseek_v4/__init__.py
gmlx/models/deepseek_v4/cache.py
gmlx/models/deepseek_v4/dspark.py
gmlx/models/deepseek_v4/hyper_connection.py
gmlx/models/deepseek_v4/model.py
gmlx/models/deepseek_v4/mtp.py
gmlx/models/gemma4/__init__.py
gmlx/models/gemma4/batched_sdpa.py
gmlx/models/gemma4/owned.py
gmlx/models/gemma4/sync.py
gmlx/models/glm5_next/__init__.py
gmlx/models/glm5_next/model.py
gmlx/models/glm5_next/mtp.py
gmlx/models/glm5_next/vision.py
gmlx/models/glm5_next/vlm_model.py
gmlx/models/hy_v3/__init__.py
gmlx/models/hy_v3/model.py
gmlx/models/hy_v3/mtp.py
gmlx/models/hy_v3/tools.py
gmlx/models/muse_glimmer/__init__.py
gmlx/models/muse_glimmer/dflash.py
gmlx/models/muse_glimmer/model.py
gmlx/models/muse_glimmer/mtp.py
gmlx/models/muse_glimmer/tools.py
gmlx/models/muse_glimmer/vlm_model.py
gmlx/models/nemotron_h/__init__.py
gmlx/models/nemotron_h/mtp.py
gmlx/models/qwen35/__init__.py
gmlx/models/qwen35/attn.py
gmlx/models/qwen35/gdn.py
gmlx/models/qwen35/layers.py
gmlx/models/qwen35/owned.py
gmlx/models/qwen35/rope.py
gmlx/models/qwen35/verify_fold.py
gmlx/models/qwen35/verify_linear.py
gmlx/models/qwen4_exp/__init__.py
gmlx/models/qwen4_exp/model.py
gmlx/models/qwen4_exp/mtp.py
gmlx/models/qwen4_exp/vlm_model.py
gmlx/serve/__init__.py
gmlx/serve/admit_gate.py
gmlx/serve/auto_ratio.py
gmlx/serve/batch_rows.py
gmlx/serve/batch_sched.py
gmlx/serve/bridge_lm.py
gmlx/serve/bridge_vlm.py
gmlx/serve/capacity.py
gmlx/serve/cb_phase.py
gmlx/serve/decode_batch.py
gmlx/serve/embeddings.py
gmlx/serve/estimate.py
gmlx/serve/governor.py
gmlx/serve/hf_cache.py
gmlx/serve/kernel_vm.py
gmlx/serve/lifecycle.py
gmlx/serve/live_requests.py
gmlx/serve/mem_preflight.py
gmlx/serve/memory.py
gmlx/serve/memtrace.py
gmlx/serve/procname.py
gmlx/serve/queue_cap.py
gmlx/serve/rerank.py
gmlx/serve/residency.py
gmlx/serve/seed_rows.py
gmlx/serve/server.py
gmlx/serve/step_timing.py
gmlx/serve/stt.py
gmlx/serve/subservice.py
gmlx/serve/tick_guard.py
gmlx/serve/tts.py
gmlx/serve/patches/__init__.py
gmlx/serve/patches/_common.py
gmlx/serve/patches/apc.py
gmlx/serve/patches/api_contract.py
gmlx/serve/patches/capacity_routes.py
gmlx/serve/patches/chat_behavior.py
gmlx/serve/patches/completions.py
gmlx/serve/patches/hardening.py
gmlx/serve/patches/observability.py
gmlx/serve/patches/render.py
gmlx/serve/patches/request_flow.py
gmlx/serve/patches/routes.py
gmlx/serve/patches/row_failed.py
gmlx/serve/patches/sampling.py
gmlx/spec/__init__.py
gmlx/spec/dflash_drafter.py
gmlx/spec/drafter_protocol.py
gmlx/spec/engine.py
gmlx/spec/helpers.py
gmlx/spec/mtp_drafter.py
gmlx/spec/mtp_load.py
gmlx/spec/overgen.py
gmlx/spec/ragged_decode.py
gmlx/spec/speculative.py
gmlx/stream/__init__.py
gmlx/stream/decode_feeder.py
gmlx/stream/feeder_common.py
gmlx/stream/gpu_token.py
gmlx/stream/keepwarm.py
gmlx/stream/lookahead.py
gmlx/stream/moe_experts.py
gmlx/stream/pagecache.py
gmlx/stream/pin_weights.py
gmlx/stream/prefetch.py
gmlx/stream/prefill_feeder.py
gmlx/talk/__init__.py
gmlx/talk/audio.py
gmlx/talk/client.py
gmlx/talk/hotkey.py
gmlx/talk/main.py
gmlx/tui/__init__.py
gmlx/tui/chat.py
gmlx/tui/reasoning.py
gmlx/tui/render.py
gmlx/tui/sessions.py
gmlx/tui/theme.py
gmlx/upstream/__init__.py
gmlx/upstream/arrays_cache_fix.py
gmlx/upstream/attn_hd512.py
gmlx/upstream/cascade_sdpa.py
gmlx/upstream/dsv32_patches.py
gmlx/upstream/gdn_patches.py
gmlx/upstream/gpt_oss_prefill.py
gmlx/upstream/occupancy_fuse.py
gmlx/upstream/patching.py
gmlx/upstream/qkv_fuse.py
gmlx/upstream/quantized_sdpa_fix.py
gmlx/upstream/rope_batch_fix.py
gmlx/upstream/rotating_cache_fix.py
gmlx/upstream/seams.json
gmlx/upstream/seams.py
gmlx/upstream/softcap_f32.py
gmlx/upstream/sparse_sdpa.py
licenses/deepseek-v4-LICENSE
licenses/ds4.c-LICENSE
licenses/llama.cpp-LICENSE
licenses/misaki-LICENSE
licenses/mlx-lm-LICENSE
licenses/mlx-vlm-LICENSE
licenses/omlx-LICENSE
tests/conftest.py
tests/test_ascii_hygiene.py
tests/test_config.py
tests/test_docs_config.py
tests/test_docs_python.py
tests/test_e2e_checks.py
tests/test_e2e_harness_smoke.py
tests/test_envflags.py
tests/test_eval_guard.py
tests/test_exitfix.py
tests/test_import_targets.py
tests/test_pinned_invariants.py
tests/test_spinner.py
tests/assistant/test_assistant_brain.py
tests/assistant/test_assistant_mcp.py
tests/assistant/test_assistant_memory.py
tests/assistant/test_assistant_serve.py
tests/cache/test_apc_ckpt_stats.py
tests/cache/test_apc_kv_quant_compose.py
tests/cache/test_apc_manager.py
tests/cache/test_apc_pooling.py
tests/cache/test_apc_routing.py
tests/cache/test_apc_schedule_invariant.py
tests/cache/test_apc_store_eval.py
tests/cache/test_cache_snapshot.py
tests/cache/test_ckpt_cursor.py
tests/cache/test_ckpt_decode_lcp.py
tests/cache/test_ckpt_plain_path.py
tests/cache/test_ckpt_pool_pressure.py
tests/cache/test_ckpt_tier.py
tests/cache/test_exact_anchor.py
tests/cache/test_fresh_gate.py
tests/cache/test_prefix_cache_budget.py
tests/cache/test_retire_key.py
tests/cache/test_retire_key_strip.py
tests/cache/test_rotating_snapshot.py
tests/commands/test_cli_config_model.py
tests/commands/test_cli_consistency.py
tests/commands/test_cli_run_flags.py
tests/commands/test_cli_umbrella.py
tests/commands/test_cli_validation.py
tests/commands/test_cmd_profiles.py
tests/commands/test_completion.py
tests/commands/test_doctor.py
tests/commands/test_extras_pyproject.py
tests/commands/test_init_wizard.py
tests/commands/test_launch.py
tests/commands/test_manage.py
tests/commands/test_menubar.py
tests/commands/test_menubar_config.py
tests/commands/test_tool_preflight.py
tests/commands/test_train.py
tests/e2e/README.md
tests/e2e/checks.py
tests/e2e/client.py
tests/e2e/judge.py
tests/e2e/models.py
tests/e2e/prompts.py
tests/e2e/pty_session.py
tests/e2e/report.py
tests/e2e/run_apc_depth_e2e.py
tests/e2e/run_apc_disk_e2e.py
tests/e2e/run_capacity_e2e.py
tests/e2e/run_capacity_multi_e2e.py
tests/e2e/run_capacity_soak_e2e.py
tests/e2e/run_chat_pty_e2e.py
tests/e2e/run_lora_e2e.py
tests/e2e/run_residency_switch_e2e.py
tests/e2e/run_serve_harmony_e2e.py
tests/e2e/run_serve_stress_e2e.py
tests/e2e/run_server_e2e.py
tests/e2e/scenarios.py
tests/e2e/server_proc.py
tests/e2e/assets/README.md
tests/e2e/assets/cats.jpg
tests/e2e/assets/speech.wav
tests/fixtures/dflash2_reference_muse.npz
tests/fixtures/dflash2_reference_qwen38.npz
tests/fixtures/kimi_k3_template.jinja
tests/gen/test_batch_parity.py
tests/gen/test_bench_prompt_source.py
tests/gen/test_diffusion_gemma.py
tests/gen/test_generate_over.py
tests/gen/test_long_context.py
tests/gen/test_prefill_decay.py
tests/gen/test_profiles.py
tests/gen/test_run_think_echo.py
tests/gen/test_thinking_budget.py
tests/invariant_probes/_common.py
tests/invariant_probes/async_survivor.py
tests/invariant_probes/gil_release.py
tests/invariant_probes/multistream_drain.py
tests/invariant_probes/reverse_topology.py
tests/invariant_probes/sync_survivor.py
tests/invariant_probes/thread_rules.py
tests/load/test_activation_dtype.py
tests/load/test_adapter.py
tests/load/test_adapter_save.py
tests/load/test_arch_gate.py
tests/load/test_config_synth.py
tests/load/test_dequant_unattachable.py
tests/load/test_discovery.py
tests/load/test_gateup_concat.py
tests/load/test_headerscan.py
tests/load/test_loader_api.py
tests/load/test_loader_guards.py
tests/load/test_loadlog.py
tests/load/test_memfit.py
tests/load/test_modules_lora.py
tests/load/test_native_fp.py
tests/load/test_nemotron_lightning.py
tests/load/test_offload.py
tests/load/test_populate.py
tests/load/test_preflight.py
tests/load/test_remap.py
tests/load/test_remote.py
tests/load/test_rope_factors.py
tests/load/test_structured.py
tests/load/test_switch_compat.py
tests/load/test_tokenizer.py
tests/load/test_transforms.py
tests/load/test_vlm_integration.py
tests/load/test_vlm_resolve.py
tests/load/test_vlm_spec_target_swap.py
tests/load/test_wired_limit_patch.py
tests/models/test_deepseek_v4_hyper_connection.py
tests/models/test_deepseek_v4_indexer_quant.py
tests/models/test_deepseek_v4_kv_quant.py
tests/models/test_deepseek_v4_mtp.py
tests/models/test_deepseek_v4_overlap_lookback.py
tests/models/test_deepseek_v4_qat.py
tests/models/test_deepseek_v4_score_profile.py
tests/models/test_forward_contract_pins.py
tests/models/test_gemma4_batched_sdpa.py
tests/models/test_gemma4_owned.py
tests/models/test_gemma4_sync.py
tests/models/test_glm5_next.py
tests/models/test_glm5_next_mtp.py
tests/models/test_glm5_next_vlm.py
tests/models/test_hy_v3_mtp.py
tests/models/test_hy_v3_tools.py
tests/models/test_hyv3_shexp_fold.py
tests/models/test_kimi_k2.py
tests/models/test_kimi_k25_vision.py
tests/models/test_kimi_k3.py
tests/models/test_kimi_k3_template.py
tests/models/test_minimax_m3_msa.py
tests/models/test_muse_glimmer_mtp.py
tests/models/test_muse_glimmer_tools.py
tests/models/test_muse_glimmer_vision.py
tests/models/test_qwen35_attn.py
tests/models/test_qwen35_gdn.py
tests/models/test_qwen35_layers.py
tests/models/test_qwen35_mirrors.py
tests/models/test_qwen35_owned.py
tests/models/test_qwen35_rope.py
tests/models/test_qwen35_verify_fold.py
tests/models/test_qwen35_verify_linear.py
tests/models/test_qwen4exp_mtp.py
tests/models/test_qwen4exp_ple.py
tests/models/test_qwen4exp_prefill_dispatch.py
tests/models/test_qwen4exp_vlm.py
tests/models/test_threadgroup_caps.py
tests/models/test_vlm_mtp_gating.py
tests/models/test_vlm_mtp_integration.py
tests/models/test_vlm_plain_image_integration.py
tests/models/test_vlm_text_only.py
tests/serve/test_admit_gate.py
tests/serve/test_api_contract.py
tests/serve/test_auto_ratio.py
tests/serve/test_batch_rows.py
tests/serve/test_batch_sched.py
tests/serve/test_capacity.py
tests/serve/test_capacity_routes.py
tests/serve/test_capacity_stream_gate.py
tests/serve/test_cb_phase.py
tests/serve/test_completions_route.py
tests/serve/test_decode_batch.py
tests/serve/test_embeddings.py
tests/serve/test_estimate.py
tests/serve/test_faithful_history.py
tests/serve/test_governor.py
tests/serve/test_hf_cache.py
tests/serve/test_lifecycle.py
tests/serve/test_live_requests.py
tests/serve/test_mem_preflight.py
tests/serve/test_mtp_degrade.py
tests/serve/test_prefill_ticks.py
tests/serve/test_queue_cap.py
tests/serve/test_rerank.py
tests/serve/test_residency.py
tests/serve/test_residency_config.py
tests/serve/test_seed_rows.py
tests/serve/test_serve_apc_config.py
tests/serve/test_serve_apc_engagement.py
tests/serve/test_serve_config.py
tests/serve/test_serve_memtrace.py
tests/serve/test_serve_mtp.py
tests/serve/test_serve_text_embed.py
tests/serve/test_serve_vlm.py
tests/serve/test_server.py
tests/serve/test_server_bridge.py
tests/serve/test_server_memory.py
tests/serve/test_server_patches.py
tests/serve/test_stt.py
tests/serve/test_tick_guard.py
tests/serve/test_tts.py
tests/serve/test_wire_contract.py
tests/spec/test_dflash2_discovery.py
tests/spec/test_dflash2_drafter.py
tests/spec/test_dflash2_load.py
tests/spec/test_dflash_normalize.py
tests/spec/test_draft_block_depth.py
tests/spec/test_drafter_protocol.py
tests/spec/test_full_prompt_prefill.py
tests/spec/test_mrope_widen.py
tests/spec/test_mtp.py
tests/spec/test_mtp_dispatch.py
tests/spec/test_mtp_preempt_resume.py
tests/spec/test_mtp_width_cap.py
tests/spec/test_native_head_presence.py
tests/spec/test_nemotron_prefill_split.py
tests/spec/test_overgen.py
tests/spec/test_pq_log.py
tests/spec/test_ragged_decode.py
tests/spec/test_rotating_inject_extend.py
tests/spec/test_shared_kv_inject_merge.py
tests/spec/test_spec_engine_release.py
tests/spec/test_spec_kv_quant.py
tests/spec/test_stoch_accept.py
tests/stream/test_decode_feeder.py
tests/stream/test_gpu_token.py
tests/stream/test_keepwarm.py
tests/stream/test_lookahead.py
tests/stream/test_moe_experts.py
tests/stream/test_pagecache.py
tests/stream/test_pin_weights.py
tests/stream/test_prefetch.py
tests/stream/test_prefill_feeder.py
tests/talk/test_hotkey.py
tests/talk/test_talk.py
tests/talk/test_talk_audio.py
tests/talk/test_talk_client.py
tests/tui/test_chat.py
tests/tui/test_chat_assistant.py
tests/tui/test_chat_e2e.py
tests/tui/test_chat_pty.py
tests/tui/test_chat_sessions.py
tests/tui/test_reasoning.py
tests/tui/test_render.py
tests/tui/test_theme.py
tests/upstream/test_arrays_cache_fix.py
tests/upstream/test_attn_hd512_verify.py
tests/upstream/test_cascade_sdpa.py
tests/upstream/test_dsv32_indexer_rope.py
tests/upstream/test_dsv32_mask_decode.py
tests/upstream/test_gdn_ba_cat.py
tests/upstream/test_gdn_noop_mask.py
tests/upstream/test_occupancy_fuse.py
tests/upstream/test_patching.py
tests/upstream/test_quantized_sdpa_fix.py
tests/upstream/test_rope_batch_fix.py
tests/upstream/test_sparse_sdpa.py
tests/upstream/test_upstream_seams.py
tests/upstream/test_verify_ragged_mask.py