aiconfigurator
Offline optimization of your disaggregated Dynamo graph
File Explorer
Download Latest Version (.zip)Showing a partial file list β download the zip above to see everything.
- openai.yaml
- custom-mapping.md
- adapt_config.py
- SKILL.md
- case_authoring.md
- failure_handling.md
- layer_permissions.md
- config_schema.md
- cross_module_impact.md
- debugging.md
- guard_rails.md
- new_backend_version.md
- render_path_policy.md
- rule_authoring.md
- template_authoring.md
- testing.md
- parity.md
- generator-development.md
- repo-guide.md
- SKILL.md
- openai.yaml
- SKILL.md
- SKILL.md
- action.yml
- areas.yaml
- build_codeowners.py
- codeowners_match.py
- emit_codeowners.py
- external_contributors.yaml
- min_approvers.py
- README.md
- test_aic_policy.py
- test_codeowners.py
- test_codeowners_validation.py
- who_owns.py
- support_matrix_request.md
- accuracy-regression-testing.yml
- build-platform-wheels-copied-pr.yml
- build-platform-wheels.yml
- build-test.yml
- cherry-pick.yml
- codeowners.yml
- collector-check.yml
- comment-charts.yml
- comment-perf-parquet-diff.yml
- common.ps1
- copyright-check.ps1
- copyright-checks.yml
- create-charts.yml
- lint-pr-title.yaml
- lint.yml
- perf-data-reuse-manifest-daily-run.yml
- perf-parquet-diff.yml
- prediction-regression-gate.yml
- random-review-assignment.yml
- support-matrix-daily-run.yml
- support-matrix-main.yml
- support-matrix-release.yml
- validate-platform-wheels.yml
- actionlint.yaml
- copy-pr-bot.yaml
- dco.yml
- pull_request_template.md
- release.yml
- design_doc.html
- parity-audit-2026-07-14.md
- parity-scan-report.md
- parity-scan-runbook.md
- perf-speedup-report.md
- python-dedup-plan.md
- compile_engine.json
- engine_step.json
- per_op.json
- benchmark_engine_step.py
- gen_moe_a2a_oracle.py
- gen_moe_expert_compute_oracle.py
- gen_op_oracle.py
- pin_goldens.py
- README.md
- test_compile_engine_parity.py
- test_engine_step_parity.py
- enums.rs
- error.rs
- mod.rs
- system_spec.rs
- mod.rs
- runtime.rs
- spec.rs
- correction.rs
- metrics.rs
- mod.rs
- model.rs
- options.rs
- regression.rs
- samples.rs
- tests.rs
- op_oracle.json
- attention.rs
- base.rs
- communication.rs
- dsa.rs
- dsv4.rs
- elementwise.rs
- embedding.rs
- fpm_forward.rs
- fpm_sol.rs
- gemm.rs
- mamba.rs
- mhc.rs
- mla.rs
- mod.rs
- moe.rs
- moe_a2a.rs
- moe_dispatch.rs
- moe_expert_compute.rs
- msa.rs
- op.rs
- overlap.rs
- util_empirical.rs
- vision.rs
- wideep_mla.rs
- moe_a2a_oracle.json
- moe_expert_compute_oracle.json
- attention.rs
- axis_curve.rs
- communication.rs
- dsa.rs
- dsv4.rs
- dsv4_megamoe.rs
- fpm_forward.rs
- gemm.rs
- interpolation.rs
- mhc.rs
- mla.rs
- mod.rs
- moe.rs
- moe_a2a.rs
- moe_expert_compute.rs
- moe_index.rs
- msa.rs
- parquet_loader.rs
- perf_interp.rs
- source_resolution.rs
- state_space.rs
- table_view.rs
- trtllm_alltoall.rs
- wideep_mla.rs
- config.rs
- lib.rs
- memory.rs
- py.rs
- py_ops.rs
- session.rs
- embedded_round_trip.rs
- memory_round_trip.rs
- pyo3_smoke.rs
- .gitignore
- Cargo.lock
- Cargo.toml
- deny.toml
- README.md
- lib.rs
- Cargo.toml
- deepseek-ai--DeepSeek-R1_config.json
- deepseek-ai--DeepSeek-V3.2_config.json
- deepseek-ai--DeepSeek-V3_config.json
- deepseek-ai--DeepSeek-V4-Flash_config.json
- deepseek-ai--DeepSeek-V4-Pro_config.json
- google--gemma-4-26B-A4B_config.json
- meta-llama--Llama-4-Maverick-17B-128E-Instruct_config.json
- meta-llama--Llama-4-Scout-17B-16E-Instruct_config.json
- meta-llama--Meta-Llama-3.1-405B_config.json
- meta-llama--Meta-Llama-3.1-70B_config.json
- meta-llama--Meta-Llama-3.1-8B_config.json
- MiniMaxAI--MiniMax-M2.5_config.json
- MiniMaxAI--MiniMax-M2.7_config.json
- MiniMaxAI--MiniMax-M3_config.json
- mistralai--Mixtral-8x22B-v0.1_config.json
- mistralai--Mixtral-8x7B-v0.1_config.json
- moonshotai--Kimi-K2-Instruct_config.json
- moonshotai--Kimi-K2.5_config.json
- moonshotai--Kimi-K3_config.json
- nvidia--DeepSeek-V3.1-NVFP4_config.json
- nvidia--DeepSeek-V3.1-NVFP4_hf_quant_config.json
- nvidia--DeepSeek-V3.2-NVFP4_config.json
- nvidia--DeepSeek-V3.2-NVFP4_hf_quant_config.json
- nvidia--DeepSeek-V4-Flash-NVFP4_config.json
- nvidia--DeepSeek-V4-Flash-NVFP4_hf_quant_config.json
- nvidia--DeepSeek-V4-Pro-NVFP4_config.json
- nvidia--DeepSeek-V4-Pro-NVFP4_hf_quant_config.json
- nvidia--Gemma-4-26B-A4B-NVFP4_config.json
- nvidia--Gemma-4-26B-A4B-NVFP4_hf_quant_config.json
- nvidia--Gemma-4-31B-IT-NVFP4_config.json
- nvidia--Gemma-4-31B-IT-NVFP4_hf_quant_config.json
- nvidia--GLM-5-NVFP4_config.json
- nvidia--GLM-5-NVFP4_hf_quant_config.json
- nvidia--GLM-5.1-NVFP4_config.json
- nvidia--GLM-5.1-NVFP4_hf_quant_config.json
- nvidia--GLM-5.2-NVFP4_config.json
- nvidia--GLM-5.2-NVFP4_hf_quant_config.json
- nvidia--Kimi-K2.5-NVFP4_config.json
- nvidia--Kimi-K2.6-NVFP4_config.json
- nvidia--Kimi-K2.6-NVFP4_hf_quant_config.json
- nvidia--Kimi-K2.7-Code-NVFP4_config.json
- nvidia--Kimi-K2.7-Code-NVFP4_hf_quant_config.json
- nvidia--Llama-3.1-70B-Instruct-FP8_config.json
- nvidia--Llama-3.1-70B-Instruct-FP8_hf_quant_config.json
- nvidia--Llama-3_3-Nemotron-Super-49B-v1_config.json
- nvidia--MiniMax-M2.5-NVFP4_config.json
- nvidia--MiniMax-M2.5-NVFP4_hf_quant_config.json
- nvidia--MiniMax-M2.7-NVFP4_config.json
- nvidia--MiniMax-M2.7-NVFP4_hf_quant_config.json
- nvidia--MiniMax-M3-NVFP4_config.json
- nvidia--MiniMax-M3-NVFP4_hf_quant_config.json
- nvidia--Nemotron-H-56B-Base-8K_config.json
- nvidia--NVIDIA-Nemotron-3-Nano-30B-A3B-BF16_config.json
- nvidia--NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP4_config.json
- nvidia--NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP4_hf_quant_config.json
- nvidia--NVIDIA-Nemotron-3-Super-120B-A12B-FP8_config.json
- nvidia--NVIDIA-Nemotron-3-Super-120B-A12B-FP8_hf_quant_config.json
- nvidia--NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4_config.json
- nvidia--NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4_hf_quant_config.json
- nvidia--NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16_config.json
- nvidia--NVIDIA-Nemotron-3-Ultra-550B-A55B-FP8_config.json
- nvidia--NVIDIA-Nemotron-3-Ultra-550B-A55B-FP8_hf_quant_config.json
- nvidia--NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4_config.json
- nvidia--NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4_hf_quant_config.json
- nvidia--NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4_config.json
- nvidia--NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4_hf_quant_config.json
- nvidia--Qwen3-235B-A22B-NVFP4_config.json
- nvidia--Qwen3-235B-A22B-NVFP4_hf_quant_config.json
- nvidia--Qwen3.5-122B-A10B-NVFP4_config.json
- nvidia--Qwen3.5-122B-A10B-NVFP4_hf_quant_config.json
- nvidia--Qwen3.5-397B-A17B-NVFP4_config.json
- nvidia--Qwen3.5-397B-A17B-NVFP4_hf_quant_config.json
- nvidia--Qwen3.6-27B-NVFP4_config.json
- nvidia--Qwen3.6-27B-NVFP4_hf_quant_config.json
- nvidia--Qwen3.6-35B-A3B-NVFP4_config.json
- nvidia--Qwen3.6-35B-A3B-NVFP4_hf_quant_config.json
- openai--gpt-oss-120b_config.json
- openai--gpt-oss-20b_config.json
- Qwen--Qwen3-0.6B_config.json
- Qwen--Qwen3-1.7B_config.json
- Qwen--Qwen3-235B-A22B-FP8_config.json
- Qwen--Qwen3-235B-A22B_config.json
- Qwen--Qwen3-30B-A3B-FP8_config.json
- Qwen--Qwen3-30B-A3B_config.json
- Qwen--Qwen3-32B-FP8-Static-PerTensor_config.json
- Qwen--Qwen3-32B-FP8-Static-PerTensor_hf_quant_config.json
- Qwen--Qwen3-32B-FP8_config.json
- Qwen--Qwen3-32B_config.json
- Qwen--Qwen3-8B_config.json
- Qwen--Qwen3-Coder-480B-A35B-Instruct_config.json
- Qwen--Qwen3-VL-235B-A22B-Instruct_config.json
- Qwen--Qwen3-VL-2B-Instruct_config.json
- Qwen--Qwen3-VL-30B-A3B-Instruct_config.json
- Qwen--Qwen3-VL-32B-Instruct_config.json
- Qwen--Qwen3-VL-32B-Thinking_config.json
- Qwen--Qwen3-VL-4B-Instruct_config.json
- Qwen--Qwen3-VL-8B-Instruct_config.json
- Qwen--Qwen3.5-122B-A10B_config.json
- Qwen--Qwen3.5-27B_config.json
- Qwen--Qwen3.5-35B-A3B_config.json
- Qwen--Qwen3.5-397B-A17B_config.json
- sgl-project--DeepSeek-V4-Flash-FP8_config.json
- sgl-project--DeepSeek-V4-Pro-FP8_config.json
- stepfun-ai--Step-3.7-Flash-FP8_config.json
- stepfun-ai--Step-3.7-Flash_config.json
- XiaomiMiMo--MiMo-7B-Base_config.json
- XiaomiMiMo--MiMo-V2-Flash_config.json
- zai-org--GLM-5-FP8_config.json
- zai-org--GLM-5.1-FP8_config.json
- zai-org--GLM-5.1_config.json
- zai-org--GLM-5.2-FP8_config.json
- zai-org--GLM-5.2_config.json
- zai-org--GLM-5_config.json
- base_backend.py
- factory.py
- sglang_backend.py
- trtllm_backend.py
- vllm_backend.py
- __init__.py
- moe.py
- vit.py
- __init__.py
- base.py
- deepseek.py
- deepseek_v32.py
- deepseek_v4.py
- gemma4.py
- gpt.py
- helpers.py
- hybrid_moe.py
- kimi_k3.py
- llama.py
- minimax_m3.py
- moe.py
- nemotron_h.py
- nemotron_nas.py
- qwen35.py
- qwen3vl.py
- README.md
- step3p7.py
- vit_ops.py
- __init__.py
- afd_transfer.py
- attention.py
- base.py
- communication.py
- dsa.py
- dsv4.py
- elementwise.py
- embedding.py
- fpm_forward.py
- gemm.py
- mamba.py
- mla.py
- moe.py
- moe_comm.py
- msa.py
- overlap.py
- README.md
- util_empirical.py
- __init__.py
- afd_partition.py
- common.py
- config.py
- config_builders.py
- engine.py
- engine_table_view.py
- errors.py
- inference_summary.py
- memory.py
- perf_database.py
- performance_result.py
- rust_engine_step.py
- step_estimate.py
- system_spec.py
- utils.py
- .gitkeep
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- .gitkeep
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- wideep_deepep_ll_perf.parquet
- wideep_deepep_normal_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- kda_perf.parquet
- collection_meta.yaml
- kda_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- mhc_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mhc_module_perf.parquet
- collection_meta.yaml
- mhc_module_perf.parquet
- reuse.yaml
- reuse.yaml
- collection_meta.yaml
- mhc_module_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- wideep_context_mla_perf.parquet
- wideep_generation_mla_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- wideep_context_moe_perf.parquet
- wideep_generation_moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- dsv4_megamoe_module_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- wideep_moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- reuse.yaml
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- reuse.yaml
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_attn_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_csa_topk_calib_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- glm5_dsa_attn_module_perf.parquet
- glm5_mqa_logits_module_perf.parquet
- glm5_topk_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_attn_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- wideep_deepep_ll_perf.parquet
- wideep_deepep_normal_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- nccl_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- kda_perf.parquet
- collection_meta.yaml
- kda_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- mhc_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mhc_module_perf.parquet
- collection_meta.yaml
- mhc_module_perf.parquet
- reuse.yaml
- reuse.yaml
- collection_meta.yaml
- mhc_module_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- wideep_context_mla_perf.parquet
- wideep_generation_mla_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- wideep_context_mla_perf.parquet
- wideep_generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- wideep_context_moe_perf.parquet
- wideep_generation_moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- wideep_moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- reuse.yaml
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_attn_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_csa_topk_calib_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- glm5_dsa_attn_module_perf.parquet
- glm5_mqa_logits_module_perf.parquet
- glm5_topk_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_attn_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- oneccl_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- wideep_deepep_ll_perf.parquet
- wideep_deepep_normal_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- trtllm_alltoall_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- kda_perf.parquet
- collection_meta.yaml
- kda_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- mhc_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mhc_module_perf.parquet
- collection_meta.yaml
- mhc_module_perf.parquet
- reuse.yaml
- reuse.yaml
- collection_meta.yaml
- mhc_module_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- wideep_context_mla_perf.parquet
- wideep_generation_mla_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- dsv4_megamoe_module_perf.parquet
- moe_perf.parquet
- wideep_context_moe_perf.parquet
- wideep_generation_moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- wideep_moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- reuse.yaml
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_attn_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_csa_topk_calib_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- glm5_dsa_attn_module_perf.parquet
- glm5_mqa_logits_module_perf.parquet
- glm5_topk_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_attn_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- wideep_deepep_ll_perf.parquet
- wideep_deepep_normal_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- kda_perf.parquet
- collection_meta.yaml
- kda_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- mhc_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mhc_module_perf.parquet
- collection_meta.yaml
- mhc_module_perf.parquet
- reuse.yaml
- reuse.yaml
- collection_meta.yaml
- mhc_module_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- wideep_context_mla_perf.parquet
- wideep_generation_mla_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- dsv4_megamoe_module_perf.parquet
- moe_perf.parquet
- wideep_context_moe_perf.parquet
- wideep_generation_moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- wideep_moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- reuse.yaml
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_attn_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_csa_topk_calib_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- glm5_dsa_attn_module_perf.parquet
- glm5_mqa_logits_module_perf.parquet
- glm5_topk_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_attn_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- .gitkeep
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- wideep_deepep_ll_perf.parquet
- wideep_deepep_normal_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- wideep_deepep_ll_perf.parquet
- wideep_deepep_normal_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- gemm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gemm_perf.parquet
- collection_meta.yaml
- kda_perf.parquet
- collection_meta.yaml
- kda_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- mamba2_perf.parquet
- collection_meta.yaml
- gdn_perf.parquet
- reuse.yaml
- collection_meta.yaml
- gdn_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mhc_module_perf.parquet
- collection_meta.yaml
- mhc_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mhc_module_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- wideep_context_mla_perf.parquet
- wideep_generation_mla_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- wideep_context_mla_perf.parquet
- wideep_generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- wideep_context_mla_perf.parquet
- wideep_generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- context_mla_perf.parquet
- generation_mla_perf.parquet
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_context_module_perf.parquet
- mla_generation_module_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- mla_bmm_perf.parquet
- collection_meta.yaml
- mla_bmm_perf.parquet
- reuse.yaml
- collection_meta.yaml
- wideep_context_mlp_perf.parquet
- wideep_generation_mlp_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- wideep_context_moe_perf.parquet
- wideep_generation_moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- wideep_context_moe_perf.parquet
- wideep_generation_moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- wideep_moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- moe_perf.parquet
- reuse.yaml
- collection_meta.yaml
- moe_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- collection_meta.yaml
- msa_context_module_perf.parquet
- msa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- reuse.yaml
- collection_meta.yaml
- computescale_perf.parquet
- scale_matrix_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_csa_topk_calib_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- glm5_dsa_attn_module_perf.parquet
- glm5_mqa_logits_module_perf.parquet
- glm5_topk_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- reuse.yaml
- collection_meta.yaml
- dsa_context_module_perf.parquet
- dsa_generation_module_perf.parquet
- dsv4_csa_context_module_perf.parquet
- dsv4_csa_generation_module_perf.parquet
- dsv4_hca_attn_module_perf.parquet
- dsv4_hca_context_module_perf.parquet
- dsv4_hca_generation_module_perf.parquet
- dsv4_paged_mqa_logits_module_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- context_attention_perf.parquet
- generation_attention_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- nccl_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- wideep_deepep_ll_perf.parquet
- wideep_deepep_normal_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- wideep_deepep_ll_perf.parquet
- wideep_deepep_normal_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- custom_allreduce_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- encoder_attention_perf.parquet
- collection_meta.yaml
- encoder_attention_perf.parquet
- reuse.yaml
- collection_meta.yaml
- a100_pcie.yaml
- a100_sxm.yaml
- a30.yaml
- b200_sxm.yaml
- b300_sxm.yaml
- b60.yaml
- __init__.py
- _aiconfigurator_core.pyi
- py.typed
- API.md
- pyproject.toml
- .coderabbit.yaml
- .dockerignore
- .gitattributes
- .gitignore
- .pre-commit-config.yaml
- AGENTS.md
- ATTRIBUTIONS.md
- Cargo.toml
- CODE_OF_CONDUCT.md
- CODEOWNERS
- CONTRIBUTING.md
- CONTRIBUTORS.md
- DEVELOPMENT.md
- LICENSE
- README.md
- SECURITY.md
# Installation Guide
git clone https://github.com/ai-dynamo/aiconfigurator
Downloads the entire project code from GitHub to your computer.
cd aiconfigurator
Moves into the project folder you just downloaded.
2. Official Install Script
Easy Recommended- Python 3 Python is required to use pip.
pip3 install aiconfigurator
Installs the package published on PyPI directly β no need to clone the source.
pip3 install ./aic-core
Installs the package published on PyPI directly β no need to clone the source.
pip3 install .
Installs the package published on PyPI directly β no need to clone the source.
Pulled directly from this repo's README.
3. Python
Easypip3 install aiconfigurator
Installs the package published on PyPI directly β no need to clone the source.
python3 -m pip uninstall -y aiconfigurator aiconfigurator-core
Runs the Python script (or module).
python3 -m pip install 'aiconfigurator==0.11.0'
Installs the Python libraries listed in requirements.txt (or similar).
python3 -m pip install --force-reinstall --no-deps 'aiconfigurator-core==0.11.0'
Installs the Python libraries listed in requirements.txt (or similar).
python3 -m venv myenv && source myenv/bin/activate # (requires Python 3.11-3.13)
Runs the Python script (or module).
Pulled directly from this repo's README.
4. Rust
Medium- Git Needed to download the project code from GitHub.
- Rust (rustup) Installing via rustup also installs cargo.
cargo build --release
Compiles the Rust project.
cargo run
Builds and then immediately runs the program.
