lattice
Run, quantize, and fine-tune LLMs on Apple Silicon. Pure Rust, no Python, no CUDA, no ONNX
File Explorer
- audit.toml
- pre-commit
- action.yml
- action.yml
- bug_report.md
- config.yml
- feature_request.md
- adr064-gpu-decode-nightly.yml
- app-binaries.yml
- bench-evidence.yml
- bench-update.yml
- cargo-audit.yml
- ci.yml
- e2e-parity.yml
- embed-parity-release.yml
- npm-prebuild.yml
- perf-postmerge-gate.yml
- release-binaries.yml
- actionlint.yaml
- dependabot.yml
- PULL_REQUEST_TEMPLATE.md
- INSTRUMENT_SCOPE.md
- lattice-logo.svg
- lattice-tray-template.svg
- LatticeStudio.icns
- generate-icon.py
- package-app.sh
- upload-release-assets.sh
- LatticeStudioApp.swift
- Drivers.swift
- LatticeBridge.swift
- LatticeEvents.swift
- RunRegistry.swift
- ServeController.swift
- Badges.swift
- ButtonStyles.swift
- CLIFooter.swift
- CommandBar.swift
- ContrastPair.swift
- CopyButton.swift
- DataTable.swift
- EmptyStateView.swift
- FaderToggle.swift
- Field.swift
- GatePill.swift
- HeroNumber.swift
- InspectorShell.swift
- KeyCapChip.swift
- MarkdownText.swift
- MassBars.swift
- OpaquePanel.swift
- ParamRow.swift
- ReadoutWell.swift
- RuntimeBar.swift
- StripChart.swift
- ChatFinalization.swift
- ChatScreen.swift
- GetModelsSheet.swift
- InspectTab.swift
- QuantizeTab.swift
- ScreenScaffold.swift
- ServeTab.swift
- TrainTab.swift
- ContentView.swift
- MainColumn.swift
- ModelSidebar.swift
- AppStore.swift
- DomainModels.swift
- Theme.swift
- ChatFinalizationTests.swift
- LatticeEventsTests.swift
- .gitignore
- DESIGN.md
- DISTRIBUTION.md
- INSTALL.md
- Package.swift
- 2026-07.jsonl
- README.md
- lattice__393.json
- lattice__699.json
- lattice__716.json
- manifest.json
- estimates.json
- estimates.json
- estimates.json
- estimates.json
- estimates.json
- estimates.json
- estimates.json
- estimates.json
- embeddings.rs
- README.md
- simd.rs
- simd_bench.rs
- simd_opt_bench.rs
- simsimd_comparison.rs
- backfill.md
- design.md
- INDEX.md
- migration.md
- model.md
- service.md
- simd.md
- basic_embed.rs
- dump_parity_embeddings.rs
- similarity.rs
- tests.txt
- coordinator.rs
- mod.rs
- tests.rs
- types.rs
- embed-drift.rs
- embed.rs
- controller.rs
- mod.rs
- tests.rs
- types.rs
- cached.rs
- mod.rs
- native.rs
- tests.rs
- binary.rs
- cosine.rs
- distance.rs
- dot_product.rs
- int4.rs
- manhattan.rs
- mod.rs
- normalize.rs
- quantized.rs
- tests.rs
- tier.rs
- cache.rs
- drift.rs
- error.rs
- lib.rs
- model.rs
- types.rs
- vision.rs
- wasm.rs
- all_minilm_l6_v2.json
- bge_small_en_v15.json
- multilingual_e5_small.json
- paraphrase_multilingual_minilm_l12_v2.json
- all_minilm_l6_v2.json
- bge_small_en_v15.json
- long_input_case.json
- multilingual_e5_small.json
- paraphrase_multilingual_minilm_l12_v2.json
- qwen3_embedding_0_6b.json
- embed_parity_wasm.mjs
- simd128_parity_wasm.mjs
- drift_api.rs
- embed_cli_image_args.rs
- embed_drift_baseline.rs
- embed_drift_cli.rs
- embed_parity_vs_hf.rs
- f16_feature_load_guard.rs
- inference_dot_product_parity.rs
- integration.rs
- simd_edge_cases.rs
- simd_precision.rs
- tokenizer_parity_e2e.rs
- vision_single_file_real_model.rs
- BASELINE.toml
- Cargo.toml
- CHANGELOG.md
- CONFIG.md
- DESIGN.md
- README.md
- router_online.rs
- design.md
- gpu.md
- INDEX.md
- network.md
- training.md
- fann_xor.rs
- buffer.rs
- circuit_breaker.rs
- context.rs
- error.rs
- mod.rs
- network.rs
- shader_manager.rs
- shaders.rs
- builder.rs
- mod.rs
- serialization.rs
- backprop.rs
- ewc.rs
- gradient.rs
- mod.rs
- rloo.rs
- activation.rs
- error.rs
- layer.rs
- lib.rs
- lora.rs
- cpu_gpu_parity.rs
- proptest_network.rs
- shader_validation.rs
- Cargo.toml
- README.md
- attention_dispatch_bench.rs
- attn_opt_bench.rs
- backward_stage0.rs
- cross_turn_prefix_cache_bench.rs
- decode_attn_bench.rs
- differential_attention_bench.rs
- e2e_bench.rs
- elementwise_bench.rs
- elementwise_cpu_bench.rs
- f16_convert_bench.rs
- gated_attention_bench.rs
- grammar_mask_bench.rs
- inference_bench.rs
- inference_perf.rs
- kv_cache_f16_bench.rs
- kv_cache_layout_bench.rs
- lm_head_bench.rs
- metal_decode_bench.rs
- metrics_bench.rs
- mtp_decode.rs
- native_sparse_attention_bench.rs
- pruning_bench.rs
- quarot_hadamard_bench.rs
- tokenizer_bench.rs
- topk_readback.rs
- golden_16q2kv_32768_ref.f32
- golden_16q2kv_4096_ref.f32
- golden_16q2kv_512_ref.f32
- golden_8q2kv_32768_ref.f32
- golden_8q2kv_4096_ref.f32
- golden_8q2kv_512_ref.f32
- baseline.json
- bench_concurrent.rs
- bench_dispatch.rs
- bench_dispatch2.rs
- bench_embed_quality.rs
- bench_embedding.rs
- bench_gdn.rs
- bench_gdn_decode.rs
- bench_gdn_prefill_ab.rs
- bench_gdn_state.rs
- bench_gpu.rs
- bench_metal.rs
- bench_mps_gemm.rs
- bench_persistent_state.rs
- bench_profile.rs
- bench_pruning.rs
- bench_q4_prefill.rs
- bench_q8_prefill.rs
- bench_quality.rs
- bench_simdgroup.rs
- bench_stability.rs
- bench_suite.rs
- decode_profile.rs
- diff_attn_layer23.rs
- diff_gdn_layer.rs
- layer_sweep.rs
- profile_metal.rs
- profile_metal_decode.rs
- decode.rs
- differential.rs
- flash.rs
- flash_causal.rs
- gated.rs
- gdn.rs
- gdn_backward.rs
- gdn_fused.rs
- gqa.rs
- mod.rs
- native_sparse.rs
- softmax_row.rs
- standard.rs
- mod.rs
- attention_gqa.rs
- gradcheck.rs
- mod.rs
- ops.rs
- tape.rs
- config.rs
- mod.rs
- scheduler.rs
- sequence.rs
- worker.rs
- chat.rs
- doctor.rs
- main.rs
- prune_score.rs
- serve.rs
- backfill_qwen3.rs
- bench_decode_ab.rs
- bench_decode_slopefit.rs
- bench_gdn_prefill_ab.rs
- bench_logit_dump.rs
- bench_lora_mixture.rs
- bench_vision_prefill_ab.rs
- chat_metal.rs
- dump_quarot_q4_golden.rs
- eval_perplexity.rs
- gramperf_profile.rs
- gramtime_profile.rs
- lattice_serve.rs
- moe_admission_sim.rs
- ppl_metal.rs
- quantize_q4.rs
- quantize_quarot.rs
- qwen35_debug.rs
- qwen35_generate.rs
- activation.rs
- arch_kernels.rs
- blas.rs
- elementwise.rs
- gemm_validate.rs
- matmul.rs
- mod.rs
- norm.rs
- simd.rs
- softmax.rs
- tests.rs
- tiled.rs
- tiled_avx2.rs
- tiled_neon.rs
- api.rs
- bind_groups.rs
- buffers.rs
- dims.rs
- dispatch.rs
- params.rs
- pipelines.rs
- shaders.rs
- state.rs
- tests.rs
- util.rs
- inner.rs
- mod.rs
- gdn.rs
- gqa.rs
- mod.rs
- dispatch.rs
- layers.rs
- dispatch.rs
- gdn_state.rs
- mtp_weights.rs
- decode_reference.metal
- flash_attention.metal
- gemm.metal
- gemm_q3_tiled.metal
- gemm_q4_tiled.metal
- gemm_q8_tiled.metal
- qwen35.metal
- rms_reduce.metal
- add.wgsl
- attention_context.wgsl
- attention_scores.wgsl
- attention_softmax.wgsl
- copy.wgsl
- gemm.wgsl
- matmul_bt.wgsl
- mul.wgsl
- rms_norm.wgsl
- rope.wgsl
- silu.wgsl
- batch_prefill.rs
- bitnet_kernel.rs
- cpu_f16.rs
- cpu_q8.rs
- gdn_chunk_ref.rs
- gpu_gemm.rs
- metal.rs
- metal_gemm.rs
- metal_qwen35.rs
- mod.rs
- moe_expert_cache.rs
- neon.rs
- neon_forward.rs
- signpost.rs
- engine.rs
- gbnf.rs
- json_schema.rs
- mod.rs
- pda.rs
- spec.rs
- trie.rs
- vocab_partition.rs
- cross_turn.rs
- flat.rs
- mod.rs
- paged.rs
- prefix.rs
- cache.rs
- debug.rs
- embed.rs
- eval.rs
- forward.rs
- generation.rs
- generation_setup.rs
- loading.rs
- mod.rs
- model.rs
- moe.rs
- norm.rs
- rerank.rs
- sampling.rs
- stop_strings.rs
- test_support.rs
- tests.rs
- weights.rs
- bert.rs
- bitnet_config.rs
- cross_encoder.rs
- gemma4_cache.rs
- gemma4_config.rs
- gemma4_loading.rs
- gemma4_model.rs
- gemma4_ops.rs
- gemma4_preflight.rs
- gemma4_weights.rs
- mod.rs
- qwen.rs
- qwen35_config.rs
- convert.rs
- forward_equivalence.rs
- hadamard.rs
- io.rs
- lm_head.rs
- lora.rs
- mod.rs
- pipeline.rs
- plan.rs
- r3_reference.rs
- rmsnorm_fusion.rs
- rotation.rs
- mod.rs
- q4_manifest.rs
- contract.rs
- embeddings.rs
- metal_worker.rs
- metrics.rs
- mod.rs
- bpe.rs
- common.rs
- detokenize.rs
- gemma_bpe.rs
- mod.rs
- sentencepiece.rs
- wordpiece.rs
- checkpoint.rs
- config.rs
- merger.rs
- mod.rs
- multimodal.rs
- pooled_embed.rs
- preprocess.rs
- qwen35_merger.rs
- qwen35_mrope.rs
- qwen35_vit.rs
- qwen35_vit_metal.rs
- vit.rs
- f16_encode.rs
- f16_weights.rs
- f32_weights.rs
- half_bits.rs
- ingress.rs
- mmap_trust.rs
- mod.rs
- q3_weights.rs
- q4_weights.rs
- q8_weights.rs
- safetensors_layout.rs
- download.rs
- error.rs
- lib.rs
- lora_hook.rs
- measurement.rs
- metrics.rs
- mixture.rs
- model_format.rs
- moe_admission.rs
- pool.rs
- pruning.rs
- rope.rs
- sampling.rs
- speculative.rs
- stop_reason.rs
- stop_token_contract.rs
- candidate_selection.json
- reference.json
- case_seed11_len384_chunk128.json
- case_seed23_len130_chunk64.json
- case_seed31_len160_chunk64_strongdecay.json
- case_seed7_len191_chunk64.json
- generate.py
- dual_rope.json
- dual_rope_global_input.bin
- dual_rope_global_inv_freq.bin
- dual_rope_global_output.bin
- dual_rope_local_input.bin
- dual_rope_local_inv_freq.bin
- dual_rope_local_output.bin
- geglu_mlp.json
- geglu_mlp_down_proj_weight.bin
- geglu_mlp_gate_proj_weight.bin
- geglu_mlp_input.bin
- geglu_mlp_output.bin
- geglu_mlp_up_proj_weight.bin
- logit_softcap.json
- logit_softcap_input.bin
- logit_softcap_output.bin
- manifest.json
- qk_norm_v_unscaled.json
- qk_norm_v_unscaled_k_input.bin
- qk_norm_v_unscaled_k_norm_weight.bin
- qk_norm_v_unscaled_k_output.bin
- qk_norm_v_unscaled_q_input.bin
- qk_norm_v_unscaled_q_norm_weight.bin
- qk_norm_v_unscaled_q_output.bin
- qk_norm_v_unscaled_v_input.bin
- qk_norm_v_unscaled_v_output.bin
- rms_norm.json
- rms_norm_input.bin
- rms_norm_output.bin
- rms_norm_weight.bin
- rms_norm_wide.json
- rms_norm_wide_input.bin
- rms_norm_wide_output.bin
- rms_norm_wide_weight.bin
- scaled_embedding.json
- scaled_embedding_embed_weight.bin
- scaled_embedding_output.bin
- boundary_golden.json
- e2e_golden.json
- chat_template.jinja
- chat_template_golden.json
- corpus_goldens.json
- decode_goldens.json
- expansion_goldens.json
- manifest.json
- processor_config.json
- tokenizer.json
- tokenizer_config.json
- e2b_config.json
- e2b_tensor_manifest.json
- golden.json
- gate_proj_golden.q4
- k_proj_golden.q4
- qwen35_0_8b_quarot_q4_greedy_tokens.json
- tokenizer.json
- tokenizer_config.json
- merges.txt
- vocab.json
- tokenizer.json
- tokenizer_config.json
- tokenizer.json
- qwen35_0_8b_config.json
- qwen36_config.json
- audit_tokenizer_parity.rs
- bert_lora_bounds_regression.rs
- differential_attention_test.rs
- examples_no_preset_fallback.rs
- gated_attention_test.rs
- gdn_decay_value_head_diff.py
- gemma4_e2e_forward_test.rs
- gemma4_tensor_manifest_test.rs
- gemma4_tokenizer_goldens_test.rs
- grammar_json_schema_regression.rs
- grammar_tbv_probe.rs
- legacy_generate_removed.rs
- metal_measurement_lock_contract.rs
- mmap_trust_boundary_contract.rs
- native_sparse_attention_test.rs
- quantize_q4_reader_refactor.rs
- quarot_q4_composed_golden.rs
- vision_goldens_fixture_test.rs
- vision_s3_vit_forward_test.rs
- vision_s3b_vit_metal_gate_test.rs
- vision_s4_merger_test.rs
- vision_s5b_e2e_gate_test.rs
- vision_serve_e2e_test.rs
- wgsl_shader_validation.rs
- BASELINE.toml
- Cargo.toml
- CHANGELOG.md
- CONFIG.md
- DESIGN.md
- METAL_TRACE.md
- README.md
- round0_transport.rs
- design.md
- INDEX.md
- drift_compare.rs
- sinkhorn_ot.rs
- barycenter.rs
- cost.rs
- divergence.rs
- drift.rs
- lib.rs
- logsumexp.rs
- math.rs
- online_drift.rs
- reference_cases.rs
- sinkhorn.rs
- sinkhorn_log.rs
- tests.rs
- transport_plan.rs
- unbalanced.rs
- input_validation.rs
- Cargo.toml
- README.md
- ADR-001-finetuning-pipeline.md
- ADR-003-knowledge-distillation.md
- ADR-004-lora-adapter-management.md
- ADR-005-training-callbacks.md
- ADR-006-jit-adaptation.md
- ADR-007-dataset-pipeline.md
- ADR-001-teacher-providers.md
- ADR-002-model-registry.md
- data.md
- design.md
- distill.md
- INDEX.md
- lora-core.md
- lora-io.md
- lora-router.md
- registry.md
- train.md
- main.rs
- generate_lora.rs
- train_grad.rs
- train_grad_full.rs
- dataset.rs
- example.rs
- mod.rs
- distill.rs
- mod.rs
- tests.rs
- types.rs
- config.rs
- mod.rs
- security.rs
- tests.rs
- mod.rs
- apply.rs
- blend.rs
- loader.rs
- manifest.rs
- mod.rs
- online.rs
- optimizer.rs
- router_update.rs
- safetensors.rs
- train.rs
- train_core.rs
- backends.rs
- mod.rs
- query.rs
- registry.rs
- tests.rs
- live_model.rs
- mod.rs
- model.rs
- rollback.rs
- safetensors_io.rs
- shadow.rs
- early_stopping.rs
- lr_schedule.rs
- mod.rs
- optimizer.rs
- regularization.rs
- builder.rs
- mod.rs
- optimizers.rs
- state.rs
- tests.rs
- uniforms.rs
- callbacks.rs
- checkpoint.rs
- metrics.rs
- mod.rs
- state.rs
- jit.rs
- mod.rs
- full_driver.rs
- error.rs
- lib.rs
- train_support.rs
- legacy_adapter.safetensors
- legacy_run.txt
- train.jsonl
- bench_backward_737.rs
- bert_cross_encoder_over_complete_lora.rs
- gdn_lora_wiring.rs
- layer23_golden_bridge.rs
- lora_apply_no_alloc.rs
- proptest_softmax.rs
- training_pipeline.rs
- Cargo.toml
- CHANGELOG.md
- README.md
- ADR_TEMPLATE.md
- RELEASE.md
- ADR-001-pure-rust-transformer-engine.md
- ADR-002-simd-dispatch.md
- ADR-003-safetensors-loading.md
- ADR-004-kv-cache.md
- ADR-005-tokenizer-architecture.md
- ADR-006-speculative-decoding.md
- ADR-007-rope-positional-encoding.md
- ADR-008-lora-injection.md
- ADR-009-model-architectures.md
- ADR-010-attention-mechanisms.md
- ADR-011-sampling-strategies.md
- ADR-012-simd-strategy.md
- ADR-013-simd-foundation.md
- ADR-014-embedding-service.md
- ADR-015-embedding-cache.md
- ADR-016-model-variants.md
- ADR-017-native-embedding-service.md
- ADR-018-quantized-vectors.md
- ADR-019-backfill-pipeline.md
- ADR-020-neural-net-primitives.md
- ADR-021-zero-alloc.md
- ADR-022-gradient-guards.md
- ADR-023-activation-functions.md
- ADR-024-network-builder.md
- ADR-025-gpu-backend.md
- ADR-026-training-loop.md
- ADR-027-finetuning-pipeline.md
- ADR-028-teacher-providers.md
- ADR-029-model-registry.md
- ADR-030-knowledge-distillation.md
- ADR-031-lora-adapter-management.md
- ADR-032-training-callbacks.md
- ADR-033-jit-adaptation.md
- ADR-034-dataset-pipeline.md
- ADR-035-sinkhorn-solver.md
- ADR-036-log-domain-stability.md
- ADR-037-cost-matrices.md
- ADR-038-barycenters.md
- ADR-039-divergence-measures.md
- ADR-040-gated-attention.md
- ADR-041-differential-attention.md
- ADR-042-native-sparse-attention.md
- ADR-043-lora-serving-verification.md
- ADR-044-quarot-rotated-quantization.md
- ADR-045-quarot-lora-composition.md
- ADR-046-structured-output.md
- ADR-047-paged-kv-cache.md
- ADR-048-continuous-batching.md
- ADR-049-vision-encoder.md
- ADR-050-rejection-sampling.md
- ADR-051-quarot-mtp-rotation.md
- ADR-052-gdn-speculative-state.md
- ADR-053-moe-metal-dispatch.md
- ADR-054-rolora-rotation-aware-lora.md
- ADR-055-online-drift-detection.md
- ADR-056-lora-tuning-pipeline.md
- ADR-057-lora-consumer-api.md
- ADR-058-cpu-perf-regression-ci.md
- ADR-059-composable-layer-architecture.md
- ADR-060-pruning-toolbox.md
- ADR-061-inference-metrics-infrastructure.md
- ADR-062-metal-fa2-prefill.md
- ADR-063-serving-architecture.md
- ADR-064-ci-gate-taxonomy.md
- ADR-065-feature-promotion-gates.md
- ADR-066-output-correctness-gate-architecture.md
- ADR-067-gpu-decode-perf-regression-gates.md
- ADR-068-grammar-wire-contract.md
- ADR-069-vision-encoder-qwen35-recalibration.md
- ADR-070-cpu-embedding-batched-encode.md
- ADR-071-gdn-state-object-priority.md
- ADR-072-weight-quantization-priority.md
- ADR-073-kv-cache-quantization-priority.md
- ADR-074-mtp-speculative-decoding-priority.md
- ADR-075-moe-expert-offload-priority.md
- ADR-076-adaptive-reasoning-priority.md
- ADR-077-pruning-distill-priority.md
- ADR-078-multimodal-serving-deferral-priority.md
- ADR-079-adaptive-micro-lora-brain.md
- ADR-080-consolidation-duplicated-contracts.md
- ADR-081-prefill-gemm-optimization-priority.md
- ADR-082-gemma4-e2b-support.md
- ADR-083-backward-simd-dispatch.md
- ADR-084-fallible-inference-validation-helpers.md
- ADR-085-explicit-quantization-simd.md
- ADR-086-retire-legacy-qwen-decode-api.md
- ADR-087-amendment-1-population-snapshot.md
- ADR-087-bench-compare-gate-calibration-and-coverage.md
- ADR-ALIGNMENT-AUDIT.md
- INDEX.md
- a2a_raw.tsv
- agentic_1000tok.json
- agentic_1k_compare.json
- agentic_2000tok.json
- agentic_4000tok.json
- agentic_sweep.json
- backward_simd_stage0_m2_max.md
- bench_data.json
- context_scaling.tsv
- fused_rmsnorm_batch_ab.txt
- lattice_benchmark.png
- perplexity.tsv
- wiki.test.raw
- v0.2.3.md
- v0.2.4.md
- v0.2.5.md
- v0.3.0.md
- v0.4.0.md
- v0.4.1.md
- v0.4.2.md
- v0.5.0.md
- v0.5.1.md
- v0.6.0.md
- v0.6.1.md
- v0.7.0.md
- v0.7.1.md
- v0.7.2.md
- v0.8.0.md
- v0.9.0.md
- architecture.md
- benchmarks.md
- capability-matrix.md
- cli-tools.md
- CONTRIBUTING.md
- cross-turn-cache.md
- e2e-parity-frozen-reference.md
- examples.md
- forward-pass.md
- gemma4.md
- generation-entrypoint-matrix.md
- getting-started.md
- inference-usage.md
- models.md
- q4-quantization.md
- RELEASE.md
- ROADMAP.md
- safety.md
- serve-http-api.md
- assert-platform-packlist.spec.mjs
- embedding.spec.mjs
- package.json
- assert-packlist.mjs
- assert-platform-packlist.mjs
- assert-prebuild-matrix.mjs
- lib.rs
- .gitignore
- binding.js
- build.rs
- Cargo.toml
- index.d.ts
- index.js
- LICENSE-APACHE
- LICENSE-MIT
- package-lock.json
- package.json
- README.md
- smoke.mjs
- assert-simd128-dispatch.mjs
- build-wasm.sh
- .gitignore
- index.mjs
- init-race.test.mjs
- package.json
- README.md
- registry.mjs
- resolve.mjs
- smoke.mjs
- CANONICAL_POLICY_IDENTITY.md
- POLICY_SHA_TRANSITIONS.md
- report_ctx1024.json
- report_ctx512.json
- bench-compare-impl.sh
- bench-host-id.py
- bench-informational-targets.sh
- bench-locks.py
- bench-python.sh
- bench-quick-informational-targets.txt
- bench-supervision.sh
- bench_supervision.py
- ensure-noindex-marker.sh
- machine-state-probe.py
- quiet-probe.py
- adr064-gpu-decode-gate.py
- bench-ci.sh
- bench-command.sh
- bench-compare.sh
- bench-gate.sh
- bench-measurements.toml
- bench_context_scaling_chart.py
- bench_cpu_flagship_supervisor.py
- bench_decode_adapters_agentic.py
- bench_decode_adapters_apples_precise.py
- bench_decode_adapters_apples_to_apples.py
- bench_decode_adapters_context_scaling.py
- bench_decode_adapters_q4_apples.py
- bench_decode_harness.py
- bench_decode_profiles.toml
- bench_decode_slopefit.py
- bench_decode_slopefit.sh
- bench_disposition_check.sh
- bench_expected_cells.toml
- bench_gate_math.py
- bench_linkedin_post.py
- bench_quality.sh
- bench_quality_selftest.sh
- bench_wasm_simd.mjs
- build-app-bins.sh
- check-capability-matrix.sh
- ci-changed-files.sh
- ci-engine-filter-selftest.sh
- ci.sh
- compare_logits.py
- e2e-parity-local.sh
- e2e_parity_check.py
- ensure-noindex-marker-selftest.sh
- fake_quant_pilot.py
- gemma4_stage3_goldens.py
- gemma4_stage5_e2e_golden.py
- gemma4_tensor_manifest.py
- gemma4_tokenizer_goldens.py
- gen_embed_parity_goldens.py
- gen_quarot_q4_composed_golden.py
- lint-absolute-paths.sh
- lint-docs.sh
- lint-source-markers.sh
- make_bench_chart.py
- perf-baselines-readme.py
- perf-bench-gate.py
- perf-bench-snapshot.py
- perf-policy.toml
- perf-postmerge-base.py
- perf_governor.py
- perf_governor.README.md
- ppl_gate_check.py
- ppl_gate_check_selftest.sh
- publish-npm-preflight-selftest.sh
- publish-npm.sh
- publish.sh
- semver-checks-disclose.py
- vision_goldens_qwen35.py
- wasm-parity.sh
- wasm-simd128-parity.sh
- ci-log-ansi-zero.txt
- real-arm-a-healthy.txt
- real-arm-b-zero.txt
- golden_image.png
- greedy_tokens.json
- input_ids.json
- manifest.json
- position_ids.json
- README.md
- vit_post_merger_f32.bin
- vit_pre_merger_f32.bin
- test_adr064_gpu_decode_gate.py
- test_bench_compare_measurement.py
- test_bench_cpu_flagship_supervisor.py
- test_bench_decode_adapters_agentic.py
- test_bench_decode_adapters_apples_precise.py
- test_bench_decode_adapters_apples_to_apples.py
- test_bench_decode_adapters_context_scaling.py
- test_bench_decode_adapters_q4_apples.py
- test_bench_decode_harness.py
- test_bench_disposition.py
- test_bench_gate_math.py
- test_bench_locks.py
- test_bench_measurement_supervision.py
- test_bench_run_record.py
- test_bench_targets.py
- test_ci_changed_files.py
- test_e2e_parity_reference.py
- test_lint_docs.py
- test_macos_app_release.py
- test_perf_bench_gate_resolution.py
- test_perf_bench_gate_status.py
- test_perf_governor.py
- test_perf_postmerge_base.py
- test_perf_postmerge_workflow.py
- test_semver_checks_disclose.py
- .gitignore
- _typos.toml
- AGENTS.md
- Cargo.lock
- Cargo.toml
- CLAUDE.md
- deno.json
- deny.toml
- LICENSE-APACHE
- LICENSE-MIT
- Makefile
- pyproject.toml
- README.md
- rust-toolchain.toml
# Use via CDN
jsDelivrjsDelivr serves any public GitHub repository as a CDN with zero setup. Pick a version and a file to get a ready-to-paste link and snippet.
Command Glossary
Commands referenced in this DOCs, explained below.
cargo bench
View Details ▼
cargo bench
Compile and execute benchmarks.
cargo bench
Execute all benchmarks of a package:
cargo bench --no-fail-fast
Don't stop when a benchmark fails:
cargo bench --no-run
Compile, but don't run benchmarks:
cargo build
View Details ▼
cargo build
Compile a local package and all of its dependencies.
cargo {{[b|build]}}
Build the package or packages defined by the `Cargo.toml` manifest file in the local path:
cargo {{[b|build]}} {{[-r|--release]}}
Build artifacts in release mode, with optimizations:
cargo {{[b|build]}} --locked
Require that `Cargo.lock` is up to date:
cargo install
View Details ▼
cargo install
Build and install a Rust binary.
cargo install {{package}}@{{version}}
Install a package from <https://crates.io> (the version is optional - latest by default):
cargo install --git {{repo_url}}
Install a package from the specified Git repository:
cargo install --git {{repo_url}} --{{branch|tag|rev}} {{branch_name|tag|commit_hash}}
Build from the specified branch/tag/commit when installing from a Git repository:
git clone
View Details ▼
git clone
Clone an existing repository.
git clone {{remote_repository_location}} {{path/to/directory}}
Clone an existing repository into a new directory (the default directory is the repository name):
git clone --recursive {{remote_repository_location}}
Clone an existing repository and its submodules:
git clone {{[-n|--no-checkout]}} {{remote_repository_location}}
Clone only the `.git` directory of an existing repository:
huggingface-cli
View Details ▼
huggingface-cli
This command is an alias of `hf`.
tldr hf
View documentation for the original command:
pip install
View Details ▼
pip install
Install Python packages.
pip install {{package1 package2 ...}}
Install one or more packages:
pip install {{package1 package2 ...}} {{[-U|--upgrade]}}
Upgrade all specified packages to the latest version, installing any that are not already present:
pip install {{package}}=={{version}}
Install a specific version of a package:
rustup update
View Details ▼
rustup update
Update Rust toolchains and `rustup` itself (if not installed using a package manager).
rustup update
Update all installed toolchains and `rustup`:
rustup update {{toolchain}}
Install or update a specific toolchain (see `rustup help toolchain` for more information):
serve
View Details ▼
serve
Static file serving and directory listing.
serve
Start an HTTP server listening on the default port to serve the current directory:
serve -p {{port}} {{path/to/directory}}
Start an HTTP server on a specific [p]ort to serve a specific directory:
serve {{[-C|--cors]}}
Start an HTTP server with CORS enabled by including the `Access-Control-Allow-Origin: *` header in all responses:
shasum
View Details ▼
shasum
Calculate SHA cryptographic checksums.
shasum {{path/to/file1 path/to/file2 ...}}
Calculate the SHA1 checksum for one or more files:
shasum --algorithm {{1|224|256|384|512|512224|512256}} {{path/to/file1 path/to/file2 ...}}
Calculate the SHA checksum for one or more files with the specified algorithm:
{{command}} | shasum
Calculate a SHA1 checksum from `stdin`:
tar
View Details ▼
tar
Archiving utility.
Often combined with a compression method, such as `gzip` or `bzip2`.
tar cf {{path/to/target.tar}} {{path/to/file1 path/to/file2 ...}}
[c]reate an archive and write it to a [f]ile:
tar czf {{path/to/target.tar.gz}} {{path/to/file1 path/to/file2 ...}}
[c]reate a g[z]ipped archive and write it to a [f]ile:
tar czf {{path/to/target.tar.gz}} {{[-C|--directory]}} {{path/to/directory}} .
[c]reate a g[z]ipped (compressed) archive from a directory using relative paths:
uv run
View Details ▼
uv run
Run a command or script in the project environment.
uv run {{path/to/script.py}}
Run a Python script:
uv run {{[-m|--module]}} {{module_name}}
Run a Python module:
uv run {{[-w|--with]}} {{package}} {{command}}
Run a command with additional packages installed temporarily:
xattr
View Details ▼
xattr
Utility to work with extended filesystem attributes.
xattr -l {{file}}
List key:value extended attributes for a given file:
xattr -w {{attribute_key}} {{attribute_value}} {{file}}
Write an attribute for a given file:
xattr -d {{com.apple.quarantine}} {{file}}
Delete an attribute from a given file:
