import { BackendParameter } from './index';

// Generated from the SGLang `launch_server` server arguments reference:
// https://docs.sglang.io/docs/advanced_features/server_arguments
// Grouped by the doc's sections. `options` lists the fixed choice values
// when the flag accepts a restricted set (sourced from the doc's Options column).
// Internal/experimental sections (Ktransformers, Diffusion LLM, Forward hooks,
// msProbe dump) and deprecated arguments are omitted.

const options: BackendParameter[] = [
  // Model and tokenizer
  {
    label: '--model-path',
    value: '--model-path',
    options: []
  },
  {
    label: '--model',
    value: '--model',
    options: []
  },
  {
    label: '--tokenizer-path',
    value: '--tokenizer-path',
    options: []
  },
  {
    label: '--tokenizer-mode',
    value: '--tokenizer-mode',
    options: ['auto', 'slow']
  },
  {
    label: '--tokenizer-worker-num',
    value: '--tokenizer-worker-num',
    options: []
  },
  {
    label: '--skip-tokenizer-init',
    value: '--skip-tokenizer-init',
    options: []
  },
  {
    label: '--load-format',
    value: '--load-format',
    options: [
      'auto',
      'pt',
      'safetensors',
      'npcache',
      'dummy',
      'sharded_state',
      'gguf',
      'bitsandbytes',
      'layered',
      'flash_rl',
      'remote',
      'remote_instance',
      'fastsafetensors',
      'private',
      'runai_streamer'
    ]
  },
  {
    label: '--model-loader-extra-config',
    value: '--model-loader-extra-config',
    options: []
  },
  {
    label: '--trust-remote-code',
    value: '--trust-remote-code',
    options: []
  },
  {
    label: '--context-length',
    value: '--context-length',
    options: []
  },
  {
    label: '--is-embedding',
    value: '--is-embedding',
    options: []
  },
  {
    label: '--enable-multimodal',
    value: '--enable-multimodal',
    options: []
  },
  {
    label: '--revision',
    value: '--revision',
    options: []
  },
  {
    label: '--model-impl',
    value: '--model-impl',
    options: []
  },
  // HTTP server
  {
    label: '--host',
    value: '--host',
    options: []
  },
  {
    label: '--port',
    value: '--port',
    options: []
  },
  {
    label: '--fastapi-root-path',
    value: '--fastapi-root-path',
    options: []
  },
  {
    label: '--grpc-mode',
    value: '--grpc-mode',
    options: []
  },
  {
    label: '--skip-server-warmup',
    value: '--skip-server-warmup',
    options: []
  },
  {
    label: '--warmups',
    value: '--warmups',
    options: []
  },
  {
    label: '--nccl-port',
    value: '--nccl-port',
    options: []
  },
  {
    label: '--checkpoint-engine-wait-weights-before-ready',
    value: '--checkpoint-engine-wait-weights-before-ready',
    options: []
  },
  // Quantization and data type
  {
    label: '--dtype',
    value: '--dtype',
    options: ['auto', 'half', 'float16', 'bfloat16', 'float', 'float32']
  },
  {
    label: '--quantization',
    value: '--quantization',
    options: [
      'awq',
      'fp8',
      'gptq',
      'marlin',
      'gptq_marlin',
      'awq_marlin',
      'bitsandbytes',
      'gguf',
      'modelopt',
      'modelopt_fp8',
      'modelopt_fp4',
      'petit_nvfp4',
      'w8a8_int8',
      'w8a8_fp8',
      'moe_wna16',
      'qoq',
      'w4afp8',
      'mxfp4',
      'mxfp8',
      'auto-round',
      'compressed-tensors',
      'modelslim',
      'quark_int4fp8_moe'
    ]
  },
  {
    label: '--quantization-param-path',
    value: '--quantization-param-path',
    options: []
  },
  {
    label: '--kv-cache-dtype',
    value: '--kv-cache-dtype',
    options: ['auto', 'fp8_e5m2', 'fp8_e4m3', 'bf16', 'bfloat16', 'fp4_e2m1']
  },
  {
    label: '--enable-fp32-lm-head',
    value: '--enable-fp32-lm-head',
    options: []
  },
  {
    label: '--modelopt-quant',
    value: '--modelopt-quant',
    options: []
  },
  {
    label: '--modelopt-checkpoint-restore-path',
    value: '--modelopt-checkpoint-restore-path',
    options: []
  },
  {
    label: '--modelopt-checkpoint-save-path',
    value: '--modelopt-checkpoint-save-path',
    options: []
  },
  {
    label: '--modelopt-export-path',
    value: '--modelopt-export-path',
    options: []
  },
  {
    label: '--quantize-and-serve',
    value: '--quantize-and-serve',
    options: []
  },
  {
    label: '--rl-quant-profile',
    value: '--rl-quant-profile',
    options: []
  },
  {
    label: '--enable-quant-communications',
    value: '--enable-quant-communications',
    options: []
  },
  // Memory and scheduling
  {
    label: '--mem-fraction-static',
    value: '--mem-fraction-static',
    options: []
  },
  {
    label: '--max-running-requests',
    value: '--max-running-requests',
    options: []
  },
  {
    label: '--max-queued-requests',
    value: '--max-queued-requests',
    options: []
  },
  {
    label: '--max-total-tokens',
    value: '--max-total-tokens',
    options: []
  },
  {
    label: '--chunked-prefill-size',
    value: '--chunked-prefill-size',
    options: []
  },
  {
    label: '--prefill-max-requests',
    value: '--prefill-max-requests',
    options: []
  },
  {
    label: '--enable-dynamic-chunking',
    value: '--enable-dynamic-chunking',
    options: []
  },
  {
    label: '--max-prefill-tokens',
    value: '--max-prefill-tokens',
    options: []
  },
  {
    label: '--schedule-policy',
    value: '--schedule-policy',
    options: [
      'lpm',
      'random',
      'fcfs',
      'dfs-weight',
      'lof',
      'priority',
      'routing-key'
    ]
  },
  {
    label: '--enable-priority-scheduling',
    value: '--enable-priority-scheduling',
    options: []
  },
  {
    label: '--abort-on-priority-when-disabled',
    value: '--abort-on-priority-when-disabled',
    options: []
  },
  {
    label: '--schedule-low-priority-values-first',
    value: '--schedule-low-priority-values-first',
    options: []
  },
  {
    label: '--priority-scheduling-preemption-threshold',
    value: '--priority-scheduling-preemption-threshold',
    options: []
  },
  {
    label: '--schedule-conservativeness',
    value: '--schedule-conservativeness',
    options: []
  },
  {
    label: '--page-size',
    value: '--page-size',
    options: []
  },
  {
    label: '--swa-full-tokens-ratio',
    value: '--swa-full-tokens-ratio',
    options: []
  },
  {
    label: '--disable-hybrid-swa-memory',
    value: '--disable-hybrid-swa-memory',
    options: []
  },
  {
    label: '--radix-eviction-policy',
    value: '--radix-eviction-policy',
    options: ['lru', 'lfu']
  },
  {
    label: '--enable-prefill-delayer',
    value: '--enable-prefill-delayer',
    options: []
  },
  {
    label: '--prefill-delayer-max-delay-passes',
    value: '--prefill-delayer-max-delay-passes',
    options: []
  },
  {
    label: '--prefill-delayer-token-usage-low-watermark',
    value: '--prefill-delayer-token-usage-low-watermark',
    options: []
  },
  {
    label: '--prefill-delayer-queue-min-ratio',
    value: '--prefill-delayer-queue-min-ratio',
    options: []
  },
  {
    label: '--prefill-delayer-max-delay-ms',
    value: '--prefill-delayer-max-delay-ms',
    options: []
  },
  {
    label: '--prefill-delayer-forward-passes-buckets',
    value: '--prefill-delayer-forward-passes-buckets',
    options: []
  },
  {
    label: '--prefill-delayer-wait-seconds-buckets',
    value: '--prefill-delayer-wait-seconds-buckets',
    options: []
  },
  // Runtime options
  {
    label: '--device',
    value: '--device',
    options: []
  },
  {
    label: '--tensor-parallel-size',
    value: '--tensor-parallel-size',
    options: []
  },
  {
    label: '--tp-size',
    value: '--tp-size',
    options: []
  },
  {
    label: '--pipeline-parallel-size',
    value: '--pipeline-parallel-size',
    options: []
  },
  {
    label: '--pp-size',
    value: '--pp-size',
    options: []
  },
  {
    label: '--attention-context-parallel-size',
    value: '--attention-context-parallel-size',
    options: []
  },
  {
    label: '--attn-cp-size',
    value: '--attn-cp-size',
    options: []
  },
  {
    label: '--moe-data-parallel-size',
    value: '--moe-data-parallel-size',
    options: []
  },
  {
    label: '--moe-dp-size',
    value: '--moe-dp-size',
    options: []
  },
  {
    label: '--pp-max-micro-batch-size',
    value: '--pp-max-micro-batch-size',
    options: []
  },
  {
    label: '--pp-async-batch-depth',
    value: '--pp-async-batch-depth',
    options: []
  },
  {
    label: '--stream-interval',
    value: '--stream-interval',
    options: []
  },
  {
    label: '--incremental-streaming-output',
    value: '--incremental-streaming-output',
    options: []
  },
  {
    label: '--random-seed',
    value: '--random-seed',
    options: []
  },
  {
    label: '--constrained-json-whitespace-pattern',
    value: '--constrained-json-whitespace-pattern',
    options: []
  },
  {
    label: '--constrained-json-disable-any-whitespace',
    value: '--constrained-json-disable-any-whitespace',
    options: []
  },
  {
    label: '--watchdog-timeout',
    value: '--watchdog-timeout',
    options: []
  },
  {
    label: '--soft-watchdog-timeout',
    value: '--soft-watchdog-timeout',
    options: []
  },
  {
    label: '--dist-timeout',
    value: '--dist-timeout',
    options: []
  },
  {
    label: '--download-dir',
    value: '--download-dir',
    options: []
  },
  {
    label: '--model-checksum',
    value: '--model-checksum',
    options: []
  },
  {
    label: '--base-gpu-id',
    value: '--base-gpu-id',
    options: []
  },
  {
    label: '--gpu-id-step',
    value: '--gpu-id-step',
    options: []
  },
  {
    label: '--sleep-on-idle',
    value: '--sleep-on-idle',
    options: []
  },
  {
    label: '--custom-sigquit-handler',
    value: '--custom-sigquit-handler',
    options: []
  },
  // Logging
  {
    label: '--log-level',
    value: '--log-level',
    options: []
  },
  {
    label: '--log-level-http',
    value: '--log-level-http',
    options: []
  },
  {
    label: '--log-requests',
    value: '--log-requests',
    options: []
  },
  {
    label: '--log-requests-level',
    value: '--log-requests-level',
    options: ['0', '1', '2', '3']
  },
  {
    label: '--log-requests-format',
    value: '--log-requests-format',
    options: ['text', 'json']
  },
  {
    label: '--log-requests-target',
    value: '--log-requests-target',
    options: []
  },
  {
    label: '--uvicorn-access-log-exclude-prefixes',
    value: '--uvicorn-access-log-exclude-prefixes',
    options: []
  },
  {
    label: '--crash-dump-folder',
    value: '--crash-dump-folder',
    options: []
  },
  {
    label: '--show-time-cost',
    value: '--show-time-cost',
    options: []
  },
  {
    label: '--enable-metrics',
    value: '--enable-metrics',
    options: []
  },
  {
    label: '--enable-mfu-metrics',
    value: '--enable-mfu-metrics',
    options: []
  },
  {
    label: '--enable-metrics-for-all-schedulers',
    value: '--enable-metrics-for-all-schedulers',
    options: []
  },
  {
    label: '--tokenizer-metrics-custom-labels-header',
    value: '--tokenizer-metrics-custom-labels-header',
    options: []
  },
  {
    label: '--tokenizer-metrics-allowed-custom-labels',
    value: '--tokenizer-metrics-allowed-custom-labels',
    options: []
  },
  {
    label: '--bucket-time-to-first-token',
    value: '--bucket-time-to-first-token',
    options: []
  },
  {
    label: '--bucket-inter-token-latency',
    value: '--bucket-inter-token-latency',
    options: []
  },
  {
    label: '--bucket-e2e-request-latency',
    value: '--bucket-e2e-request-latency',
    options: []
  },
  {
    label: '--collect-tokens-histogram',
    value: '--collect-tokens-histogram',
    options: []
  },
  {
    label: '--prompt-tokens-buckets',
    value: '--prompt-tokens-buckets',
    options: []
  },
  {
    label: '--generation-tokens-buckets',
    value: '--generation-tokens-buckets',
    options: []
  },
  {
    label: '--gc-warning-threshold-secs',
    value: '--gc-warning-threshold-secs',
    options: []
  },
  {
    label: '--decode-log-interval',
    value: '--decode-log-interval',
    options: []
  },
  {
    label: '--enable-request-time-stats-logging',
    value: '--enable-request-time-stats-logging',
    options: []
  },
  {
    label: '--kv-events-config',
    value: '--kv-events-config',
    options: []
  },
  {
    label: '--enable-trace',
    value: '--enable-trace',
    options: []
  },
  {
    label: '--otlp-traces-endpoint',
    value: '--otlp-traces-endpoint',
    options: []
  },
  // RequestMetricsExporter configuration
  {
    label: '--export-metrics-to-file',
    value: '--export-metrics-to-file',
    options: []
  },
  {
    label: '--export-metrics-to-file-dir',
    value: '--export-metrics-to-file-dir',
    options: []
  },
  // API related
  {
    label: '--api-key',
    value: '--api-key',
    options: []
  },
  {
    label: '--admin-api-key',
    value: '--admin-api-key',
    options: []
  },
  {
    label: '--served-model-name',
    value: '--served-model-name',
    options: []
  },
  {
    label: '--weight-version',
    value: '--weight-version',
    options: []
  },
  {
    label: '--chat-template',
    value: '--chat-template',
    options: []
  },
  {
    label: '--hf-chat-template-name',
    value: '--hf-chat-template-name',
    options: []
  },
  {
    label: '--completion-template',
    value: '--completion-template',
    options: []
  },
  {
    label: '--file-storage-path',
    value: '--file-storage-path',
    options: []
  },
  {
    label: '--enable-cache-report',
    value: '--enable-cache-report',
    options: []
  },
  {
    label: '--reasoning-parser',
    value: '--reasoning-parser',
    options: [
      'deepseek-r1',
      'deepseek-v3',
      'glm45',
      'gpt-oss',
      'kimi',
      'qwen3',
      'qwen3-thinking',
      'step3'
    ]
  },
  {
    label: '--tool-call-parser',
    value: '--tool-call-parser',
    options: [
      'deepseekv3',
      'deepseekv31',
      'glm',
      'glm45',
      'glm47',
      'gpt-oss',
      'kimi_k2',
      'llama3',
      'mistral',
      'pythonic',
      'qwen',
      'qwen25',
      'qwen3_coder',
      'step3',
      'gigachat3'
    ]
  },
  {
    label: '--tool-server',
    value: '--tool-server',
    options: []
  },
  {
    label: '--sampling-defaults',
    value: '--sampling-defaults',
    options: ['openai', 'model']
  },
  // Data parallelism
  {
    label: '--data-parallel-size',
    value: '--data-parallel-size',
    options: []
  },
  {
    label: '--dp-size',
    value: '--dp-size',
    options: []
  },
  {
    label: '--load-balance-method',
    value: '--load-balance-method',
    options: [
      'auto',
      'round_robin',
      'follow_bootstrap_room',
      'total_requests',
      'total_tokens'
    ]
  },
  // Multi-node distributed serving
  {
    label: '--dist-init-addr',
    value: '--dist-init-addr',
    options: []
  },
  {
    label: '--nccl-init-addr',
    value: '--nccl-init-addr',
    options: []
  },
  {
    label: '--nnodes',
    value: '--nnodes',
    options: []
  },
  {
    label: '--node-rank',
    value: '--node-rank',
    options: []
  },
  // Model override args
  {
    label: '--json-model-override-args',
    value: '--json-model-override-args',
    options: []
  },
  {
    label: '--preferred-sampling-params',
    value: '--preferred-sampling-params',
    options: []
  },
  // LoRA
  {
    label: '--enable-lora',
    value: '--enable-lora',
    options: []
  },
  {
    label: '--enable-lora-overlap-loading',
    value: '--enable-lora-overlap-loading',
    options: []
  },
  {
    label: '--max-lora-rank',
    value: '--max-lora-rank',
    options: []
  },
  {
    label: '--lora-target-modules',
    value: '--lora-target-modules',
    options: [
      'q_proj',
      'k_proj',
      'v_proj',
      'o_proj',
      'gate_proj',
      'up_proj',
      'down_proj',
      'qkv_proj',
      'gate_up_proj',
      'all'
    ]
  },
  {
    label: '--lora-paths',
    value: '--lora-paths',
    options: []
  },
  {
    label: '--max-loras-per-batch',
    value: '--max-loras-per-batch',
    options: []
  },
  {
    label: '--max-loaded-loras',
    value: '--max-loaded-loras',
    options: []
  },
  {
    label: '--lora-eviction-policy',
    value: '--lora-eviction-policy',
    options: ['lru', 'fifo']
  },
  {
    label: '--lora-backend',
    value: '--lora-backend',
    options: ['triton', 'csgmv', 'ascend', 'torch_native']
  },
  {
    label: '--max-lora-chunk-size',
    value: '--max-lora-chunk-size',
    options: ['16', '32', '64', '128']
  },
  {
    label: '--lora-drain-wait-threshold',
    value: '--lora-drain-wait-threshold',
    options: []
  },
  // Kernel Backends (Attention, Sampling, Grammar, GEMM)
  {
    label: '--attention-backend',
    value: '--attention-backend',
    options: [
      'triton',
      'torch_native',
      'flex_attention',
      'dsa',
      'nsa',
      'cutlass_mla',
      'fa3',
      'fa4',
      'flashinfer',
      'flashmla',
      'trtllm_mla',
      'trtllm_mha',
      'dual_chunk_flash_attn',
      'aiter',
      'wave',
      'intel_amx',
      'ascend'
    ]
  },
  {
    label: '--prefill-attention-backend',
    value: '--prefill-attention-backend',
    options: [
      'triton',
      'torch_native',
      'flex_attention',
      'dsa',
      'nsa',
      'cutlass_mla',
      'fa3',
      'fa4',
      'flashinfer',
      'flashmla',
      'trtllm_mla',
      'trtllm_mha',
      'dual_chunk_flash_attn',
      'aiter',
      'wave',
      'intel_amx',
      'ascend'
    ]
  },
  {
    label: '--decode-attention-backend',
    value: '--decode-attention-backend',
    options: [
      'triton',
      'torch_native',
      'flex_attention',
      'dsa',
      'nsa',
      'cutlass_mla',
      'fa3',
      'fa4',
      'flashinfer',
      'flashmla',
      'trtllm_mla',
      'trtllm_mha',
      'dual_chunk_flash_attn',
      'aiter',
      'wave',
      'intel_amx',
      'ascend'
    ]
  },
  {
    label: '--sampling-backend',
    value: '--sampling-backend',
    options: ['flashinfer', 'pytorch', 'ascend']
  },
  {
    label: '--grammar-backend',
    value: '--grammar-backend',
    options: ['xgrammar', 'outlines', 'llguidance', 'none']
  },
  {
    label: '--mm-attention-backend',
    value: '--mm-attention-backend',
    options: ['sdpa', 'fa3', 'fa4', 'triton_attn', 'ascend_attn', 'aiter_attn']
  },
  {
    label: '--dsa-prefill-backend',
    value: '--dsa-prefill-backend',
    options: [
      'flashmla_sparse',
      'flashmla_kv',
      'flashmla_auto',
      'fa3',
      'tilelang',
      'aiter',
      'trtllm'
    ]
  },
  {
    label: '--dsa-decode-backend',
    value: '--dsa-decode-backend',
    options: [
      'flashmla_sparse',
      'flashmla_kv',
      'fa3',
      'tilelang',
      'aiter',
      'trtllm'
    ]
  },
  {
    label: '--dsa-topk-backend',
    value: '--dsa-topk-backend',
    options: ['sgl-kernel', 'torch', 'flashinfer']
  },
  {
    label: '--fp8-gemm-backend',
    value: '--fp8-gemm-backend',
    options: [
      'auto',
      'deep_gemm',
      'flashinfer_trtllm',
      'flashinfer_cutlass',
      'flashinfer_deepgemm',
      'cutlass',
      'triton',
      'aiter'
    ]
  },
  {
    label: '--fp4-gemm-backend',
    value: '--fp4-gemm-backend',
    options: [
      'auto',
      'flashinfer_cudnn',
      'flashinfer_cutlass',
      'flashinfer_trtllm'
    ]
  },
  {
    label: '--disable-flashinfer-autotune',
    value: '--disable-flashinfer-autotune',
    options: []
  },
  // Speculative decoding
  {
    label: '--speculative-algorithm',
    value: '--speculative-algorithm',
    options: ['EAGLE', 'EAGLE3', 'NEXTN', 'STANDALONE', 'NGRAM']
  },
  {
    label: '--speculative-draft-model-path',
    value: '--speculative-draft-model-path',
    options: []
  },
  {
    label: '--speculative-draft-model',
    value: '--speculative-draft-model',
    options: []
  },
  {
    label: '--speculative-draft-model-revision',
    value: '--speculative-draft-model-revision',
    options: []
  },
  {
    label: '--speculative-draft-load-format',
    value: '--speculative-draft-load-format',
    options: []
  },
  {
    label: '--speculative-num-steps',
    value: '--speculative-num-steps',
    options: []
  },
  {
    label: '--speculative-eagle-topk',
    value: '--speculative-eagle-topk',
    options: []
  },
  {
    label: '--speculative-num-draft-tokens',
    value: '--speculative-num-draft-tokens',
    options: []
  },
  {
    label: '--speculative-accept-threshold-single',
    value: '--speculative-accept-threshold-single',
    options: []
  },
  {
    label: '--speculative-accept-threshold-acc',
    value: '--speculative-accept-threshold-acc',
    options: []
  },
  {
    label: '--speculative-token-map',
    value: '--speculative-token-map',
    options: []
  },
  {
    label: '--speculative-attention-mode',
    value: '--speculative-attention-mode',
    options: ['prefill', 'decode']
  },
  {
    label: '--speculative-draft-attention-backend',
    value: '--speculative-draft-attention-backend',
    options: []
  },
  {
    label: '--speculative-moe-runner-backend',
    value: '--speculative-moe-runner-backend',
    options: []
  },
  {
    label: '--speculative-moe-a2a-backend',
    value: '--speculative-moe-a2a-backend',
    options: []
  },
  {
    label: '--speculative-draft-model-quantization',
    value: '--speculative-draft-model-quantization',
    options: []
  },
  // Ngram speculative decoding
  {
    label: '--speculative-ngram-min-bfs-breadth',
    value: '--speculative-ngram-min-bfs-breadth',
    options: []
  },
  {
    label: '--speculative-ngram-max-bfs-breadth',
    value: '--speculative-ngram-max-bfs-breadth',
    options: []
  },
  {
    label: '--speculative-ngram-match-type',
    value: '--speculative-ngram-match-type',
    options: ['BFS', 'PROB']
  },
  {
    label: '--speculative-ngram-max-trie-depth',
    value: '--speculative-ngram-max-trie-depth',
    options: []
  },
  {
    label: '--speculative-ngram-capacity',
    value: '--speculative-ngram-capacity',
    options: []
  },
  // Multi-layer Eagle speculative decoding
  {
    label: '--enable-multi-layer-eagle',
    value: '--enable-multi-layer-eagle',
    options: []
  },
  // MoE
  {
    label: '--expert-parallel-size',
    value: '--expert-parallel-size',
    options: []
  },
  {
    label: '--ep-size',
    value: '--ep-size',
    options: []
  },
  {
    label: '--ep',
    value: '--ep',
    options: []
  },
  {
    label: '--moe-a2a-backend',
    value: '--moe-a2a-backend',
    options: ['none', 'deepep', 'mooncake', 'mori', 'nixl', 'ascend_fuseep']
  },
  {
    label: '--moe-runner-backend',
    value: '--moe-runner-backend',
    options: [
      'auto',
      'deep_gemm',
      'triton',
      'triton_kernel',
      'flashinfer_trtllm',
      'flashinfer_trtllm_routed',
      'flashinfer_cutlass',
      'flashinfer_mxfp4',
      'flashinfer_cutedsl',
      'cutlass'
    ]
  },
  {
    label: '--flashinfer-mxfp4-moe-precision',
    value: '--flashinfer-mxfp4-moe-precision',
    options: ['default', 'bf16']
  },
  {
    label: '--enable-flashinfer-allreduce-fusion',
    value: '--enable-flashinfer-allreduce-fusion',
    options: []
  },
  {
    label: '--enable-aiter-allreduce-fusion',
    value: '--enable-aiter-allreduce-fusion',
    options: []
  },
  {
    label: '--deepep-mode',
    value: '--deepep-mode',
    options: ['normal', 'low_latency', 'auto']
  },
  {
    label: '--deepep-dispatcher-output-dtype',
    value: '--deepep-dispatcher-output-dtype',
    options: ['bf16', 'fp8', 'int8', 'nvfp4', 'auto']
  },
  {
    label: '--ep-num-redundant-experts',
    value: '--ep-num-redundant-experts',
    options: []
  },
  {
    label: '--ep-dispatch-algorithm',
    value: '--ep-dispatch-algorithm',
    options: []
  },
  {
    label: '--init-expert-location',
    value: '--init-expert-location',
    options: []
  },
  {
    label: '--enable-eplb',
    value: '--enable-eplb',
    options: []
  },
  {
    label: '--eplb-algorithm',
    value: '--eplb-algorithm',
    options: []
  },
  {
    label: '--eplb-rebalance-num-iterations',
    value: '--eplb-rebalance-num-iterations',
    options: []
  },
  {
    label: '--eplb-rebalance-layers-per-chunk',
    value: '--eplb-rebalance-layers-per-chunk',
    options: []
  },
  {
    label: '--eplb-min-rebalancing-utilization-threshold',
    value: '--eplb-min-rebalancing-utilization-threshold',
    options: []
  },
  {
    label: '--expert-distribution-recorder-mode',
    value: '--expert-distribution-recorder-mode',
    options: []
  },
  {
    label: '--expert-distribution-recorder-buffer-size',
    value: '--expert-distribution-recorder-buffer-size',
    options: []
  },
  {
    label: '--enable-expert-distribution-metrics',
    value: '--enable-expert-distribution-metrics',
    options: []
  },
  {
    label: '--deepep-config',
    value: '--deepep-config',
    options: []
  },
  {
    label: '--moe-dense-tp-size',
    value: '--moe-dense-tp-size',
    options: []
  },
  {
    label: '--elastic-ep-backend',
    value: '--elastic-ep-backend',
    options: ['none', 'mooncake']
  },
  {
    label: '--enable-elastic-expert-backup',
    value: '--enable-elastic-expert-backup',
    options: []
  },
  {
    label: '--mooncake-ib-device',
    value: '--mooncake-ib-device',
    options: []
  },
  {
    label: '--enable-deepep-waterfill',
    value: '--enable-deepep-waterfill',
    options: []
  },
  {
    label: '--elastic-ep-rejoin',
    value: '--elastic-ep-rejoin',
    options: []
  },
  // Mamba Cache
  {
    label: '--max-mamba-cache-size',
    value: '--max-mamba-cache-size',
    options: []
  },
  {
    label: '--mamba-ssm-dtype',
    value: '--mamba-ssm-dtype',
    options: ['float32', 'bfloat16', 'float16']
  },
  {
    label: '--mamba-full-memory-ratio',
    value: '--mamba-full-memory-ratio',
    options: []
  },
  {
    label: '--mamba-scheduler-strategy',
    value: '--mamba-scheduler-strategy',
    options: ['auto', 'no_buffer', 'extra_buffer']
  },
  {
    label: '--mamba-track-interval',
    value: '--mamba-track-interval',
    options: []
  },
  // Hierarchical cache
  {
    label: '--enable-hierarchical-cache',
    value: '--enable-hierarchical-cache',
    options: []
  },
  {
    label: '--hicache-ratio',
    value: '--hicache-ratio',
    options: []
  },
  {
    label: '--hicache-size',
    value: '--hicache-size',
    options: []
  },
  {
    label: '--hicache-write-policy',
    value: '--hicache-write-policy',
    options: ['write_back', 'write_through', 'write_through_selective']
  },
  {
    label: '--hicache-io-backend',
    value: '--hicache-io-backend',
    options: ['direct', 'kernel', 'kernel_ascend']
  },
  {
    label: '--hicache-mem-layout',
    value: '--hicache-mem-layout',
    options: [
      'layer_first',
      'page_first',
      'page_first_direct',
      'page_first_kv_split',
      'page_head'
    ]
  },
  {
    label: '--hicache-storage-backend',
    value: '--hicache-storage-backend',
    options: ['file', 'mooncake', 'hf3fs', 'nixl', 'aibrix', 'dynamic', 'eic']
  },
  {
    label: '--hicache-storage-prefetch-policy',
    value: '--hicache-storage-prefetch-policy',
    options: ['best_effort', 'wait_complete', 'timeout']
  },
  {
    label: '--hicache-storage-backend-extra-config',
    value: '--hicache-storage-backend-extra-config',
    options: []
  },
  // Hierarchical sparse attention
  {
    label: '--hierarchical-sparse-attention-extra-config',
    value: '--hierarchical-sparse-attention-extra-config',
    options: []
  },
  // LMCache
  {
    label: '--enable-lmcache',
    value: '--enable-lmcache',
    options: []
  },
  // Offloading
  {
    label: '--cpu-offload-gb',
    value: '--cpu-offload-gb',
    options: []
  },
  {
    label: '--offload-group-size',
    value: '--offload-group-size',
    options: []
  },
  {
    label: '--offload-num-in-group',
    value: '--offload-num-in-group',
    options: []
  },
  {
    label: '--offload-prefetch-step',
    value: '--offload-prefetch-step',
    options: []
  },
  {
    label: '--offload-mode',
    value: '--offload-mode',
    options: []
  },
  // Args for multi-item scoring
  {
    label: '--multi-item-scoring-delimiter',
    value: '--multi-item-scoring-delimiter',
    options: []
  },
  // Optimization/debug options
  {
    label: '--disable-radix-cache',
    value: '--disable-radix-cache',
    options: []
  },
  {
    label: '--cuda-graph-max-bs',
    value: '--cuda-graph-max-bs',
    options: []
  },
  {
    label: '--cuda-graph-bs',
    value: '--cuda-graph-bs',
    options: []
  },
  {
    label: '--disable-cuda-graph',
    value: '--disable-cuda-graph',
    options: []
  },
  {
    label: '--disable-cuda-graph-padding',
    value: '--disable-cuda-graph-padding',
    options: []
  },
  {
    label: '--enable-profile-cuda-graph',
    value: '--enable-profile-cuda-graph',
    options: []
  },
  {
    label: '--enable-cudagraph-gc',
    value: '--enable-cudagraph-gc',
    options: []
  },
  {
    label: '--enable-layerwise-nvtx-marker',
    value: '--enable-layerwise-nvtx-marker',
    options: []
  },
  {
    label: '--enable-nccl-nvls',
    value: '--enable-nccl-nvls',
    options: []
  },
  {
    label: '--enable-symm-mem',
    value: '--enable-symm-mem',
    options: []
  },
  {
    label: '--disable-flashinfer-cutlass-moe-fp4-allgather',
    value: '--disable-flashinfer-cutlass-moe-fp4-allgather',
    options: []
  },
  {
    label: '--enable-tokenizer-batch-encode',
    value: '--enable-tokenizer-batch-encode',
    options: []
  },
  {
    label: '--disable-tokenizer-batch-decode',
    value: '--disable-tokenizer-batch-decode',
    options: []
  },
  {
    label: '--disable-outlines-disk-cache',
    value: '--disable-outlines-disk-cache',
    options: []
  },
  {
    label: '--disable-custom-all-reduce',
    value: '--disable-custom-all-reduce',
    options: []
  },
  {
    label: '--enable-mscclpp',
    value: '--enable-mscclpp',
    options: []
  },
  {
    label: '--enable-torch-symm-mem',
    value: '--enable-torch-symm-mem',
    options: []
  },
  {
    label: '--disable-overlap-schedule',
    value: '--disable-overlap-schedule',
    options: []
  },
  {
    label: '--enable-mixed-chunk',
    value: '--enable-mixed-chunk',
    options: []
  },
  {
    label: '--enable-dp-attention',
    value: '--enable-dp-attention',
    options: []
  },
  {
    label: '--enable-dp-lm-head',
    value: '--enable-dp-lm-head',
    options: []
  },
  {
    label: '--enable-two-batch-overlap',
    value: '--enable-two-batch-overlap',
    options: []
  },
  {
    label: '--enable-single-batch-overlap',
    value: '--enable-single-batch-overlap',
    options: []
  },
  {
    label: '--tbo-token-distribution-threshold',
    value: '--tbo-token-distribution-threshold',
    options: []
  },
  {
    label: '--enable-torch-compile',
    value: '--enable-torch-compile',
    options: []
  },
  {
    label: '--enable-torch-compile-debug-mode',
    value: '--enable-torch-compile-debug-mode',
    options: []
  },
  {
    label: '--disable-piecewise-cuda-graph',
    value: '--disable-piecewise-cuda-graph',
    options: []
  },
  {
    label: '--enforce-piecewise-cuda-graph',
    value: '--enforce-piecewise-cuda-graph',
    options: []
  },
  {
    label: '--piecewise-cuda-graph-tokens',
    value: '--piecewise-cuda-graph-tokens',
    options: []
  },
  {
    label: '--piecewise-cuda-graph-compiler',
    value: '--piecewise-cuda-graph-compiler',
    options: ['eager', 'inductor']
  },
  {
    label: '--torch-compile-max-bs',
    value: '--torch-compile-max-bs',
    options: []
  },
  {
    label: '--piecewise-cuda-graph-max-tokens',
    value: '--piecewise-cuda-graph-max-tokens',
    options: []
  },
  {
    label: '--torchao-config',
    value: '--torchao-config',
    options: []
  },
  {
    label: '--enable-nan-detection',
    value: '--enable-nan-detection',
    options: []
  },
  {
    label: '--enable-p2p-check',
    value: '--enable-p2p-check',
    options: []
  },
  {
    label: '--triton-attention-reduce-in-fp32',
    value: '--triton-attention-reduce-in-fp32',
    options: []
  },
  {
    label: '--triton-attention-num-kv-splits',
    value: '--triton-attention-num-kv-splits',
    options: []
  },
  {
    label: '--triton-attention-split-tile-size',
    value: '--triton-attention-split-tile-size',
    options: []
  },
  {
    label: '--num-continuous-decode-steps',
    value: '--num-continuous-decode-steps',
    options: []
  },
  {
    label: '--delete-ckpt-after-loading',
    value: '--delete-ckpt-after-loading',
    options: []
  },
  {
    label: '--enable-memory-saver',
    value: '--enable-memory-saver',
    options: []
  },
  {
    label: '--enable-weights-cpu-backup',
    value: '--enable-weights-cpu-backup',
    options: []
  },
  {
    label: '--enable-draft-weights-cpu-backup',
    value: '--enable-draft-weights-cpu-backup',
    options: []
  },
  {
    label: '--allow-auto-truncate',
    value: '--allow-auto-truncate',
    options: []
  },
  {
    label: '--enable-custom-logit-processor',
    value: '--enable-custom-logit-processor',
    options: []
  },
  {
    label: '--flashinfer-mla-disable-ragged',
    value: '--flashinfer-mla-disable-ragged',
    options: []
  },
  {
    label: '--disable-shared-experts-fusion',
    value: '--disable-shared-experts-fusion',
    options: []
  },
  {
    label: '--disable-chunked-prefix-cache',
    value: '--disable-chunked-prefix-cache',
    options: []
  },
  {
    label: '--disable-fast-image-processor',
    value: '--disable-fast-image-processor',
    options: []
  },
  {
    label: '--keep-mm-feature-on-device',
    value: '--keep-mm-feature-on-device',
    options: []
  },
  {
    label: '--enable-return-hidden-states',
    value: '--enable-return-hidden-states',
    options: []
  },
  {
    label: '--enable-return-routed-experts',
    value: '--enable-return-routed-experts',
    options: []
  },
  {
    label: '--scheduler-recv-interval',
    value: '--scheduler-recv-interval',
    options: []
  },
  {
    label: '--numa-node',
    value: '--numa-node',
    options: []
  },
  {
    label: '--enable-deterministic-inference',
    value: '--enable-deterministic-inference',
    options: []
  },
  {
    label: '--rl-on-policy-target',
    value: '--rl-on-policy-target',
    options: ['fsdp']
  },
  {
    label: '--enable-attn-tp-input-scattered',
    value: '--enable-attn-tp-input-scattered',
    options: []
  },
  {
    label: '--enable-dsa-prefill-context-parallel',
    value: '--enable-dsa-prefill-context-parallel',
    options: []
  },
  {
    label: '--dsa-prefill-cp-mode',
    value: '--dsa-prefill-cp-mode',
    options: ['in-seq-split', 'round-robin-split']
  },
  {
    label: '--enable-fused-qk-norm-rope',
    value: '--enable-fused-qk-norm-rope',
    options: []
  },
  {
    label: '--enable-precise-embedding-interpolation',
    value: '--enable-precise-embedding-interpolation',
    options: []
  },
  // Dynamic batch tokenizer
  {
    label: '--enable-dynamic-batch-tokenizer',
    value: '--enable-dynamic-batch-tokenizer',
    options: []
  },
  {
    label: '--dynamic-batch-tokenizer-batch-size',
    value: '--dynamic-batch-tokenizer-batch-size',
    options: []
  },
  {
    label: '--dynamic-batch-tokenizer-batch-timeout',
    value: '--dynamic-batch-tokenizer-batch-timeout',
    options: []
  },
  // Debug tensor dumps
  {
    label: '--debug-tensor-dump-output-folder',
    value: '--debug-tensor-dump-output-folder',
    options: []
  },
  {
    label: '--debug-tensor-dump-layers',
    value: '--debug-tensor-dump-layers',
    options: []
  },
  {
    label: '--debug-tensor-dump-input-file',
    value: '--debug-tensor-dump-input-file',
    options: []
  },
  {
    label: '--debug-tensor-dump-inject',
    value: '--debug-tensor-dump-inject',
    options: []
  },
  // PD disaggregation
  {
    label: '--disaggregation-mode',
    value: '--disaggregation-mode',
    options: ['null', 'prefill', 'decode']
  },
  {
    label: '--disaggregation-transfer-backend',
    value: '--disaggregation-transfer-backend',
    options: ['mooncake', 'nixl', 'ascend', 'fake']
  },
  {
    label: '--disaggregation-bootstrap-port',
    value: '--disaggregation-bootstrap-port',
    options: []
  },
  {
    label: '--disaggregation-ib-device',
    value: '--disaggregation-ib-device',
    options: []
  },
  {
    label: '--disaggregation-decode-enable-offload-kvcache',
    value: '--disaggregation-decode-enable-offload-kvcache',
    options: []
  },
  {
    label: '--num-reserved-decode-tokens',
    value: '--num-reserved-decode-tokens',
    options: []
  },
  {
    label: '--disaggregation-decode-polling-interval',
    value: '--disaggregation-decode-polling-interval',
    options: []
  },
  // Encode prefill disaggregation
  {
    label: '--encoder-only',
    value: '--encoder-only',
    options: []
  },
  {
    label: '--language-only',
    value: '--language-only',
    options: []
  },
  {
    label: '--encoder-transfer-backend',
    value: '--encoder-transfer-backend',
    options: ['zmq_to_scheduler', 'zmq_to_tokenizer', 'mooncake']
  },
  {
    label: '--encoder-urls',
    value: '--encoder-urls',
    options: []
  },
  // Custom weight loader
  {
    label: '--custom-weight-loader',
    value: '--custom-weight-loader',
    options: []
  },
  {
    label: '--weight-loader-disable-mmap',
    value: '--weight-loader-disable-mmap',
    options: []
  },
  {
    label: '--weight-loader-prefetch-checkpoints',
    value: '--weight-loader-prefetch-checkpoints',
    options: []
  },
  {
    label: '--weight-loader-prefetch-num-threads',
    value: '--weight-loader-prefetch-num-threads',
    options: []
  },
  {
    label: '--remote-instance-weight-loader-seed-instance-ip',
    value: '--remote-instance-weight-loader-seed-instance-ip',
    options: []
  },
  {
    label: '--remote-instance-weight-loader-seed-instance-service-port',
    value: '--remote-instance-weight-loader-seed-instance-service-port',
    options: []
  },
  {
    label: '--remote-instance-weight-loader-send-weights-group-ports',
    value: '--remote-instance-weight-loader-send-weights-group-ports',
    options: []
  },
  {
    label: '--remote-instance-weight-loader-backend',
    value: '--remote-instance-weight-loader-backend',
    options: ['transfer_engine', 'nccl']
  },
  {
    label: '--remote-instance-weight-loader-start-seed-via-transfer-engine',
    value: '--remote-instance-weight-loader-start-seed-via-transfer-engine',
    options: []
  },
  // For PD-Multiplexing
  {
    label: '--enable-pdmux',
    value: '--enable-pdmux',
    options: []
  },
  {
    label: '--pdmux-config-path',
    value: '--pdmux-config-path',
    options: []
  },
  {
    label: '--sm-group-num',
    value: '--sm-group-num',
    options: []
  },
  // Configuration file support
  {
    label: '--config',
    value: '--config',
    options: []
  },
  // For Multi-Modal
  {
    label: '--mm-max-concurrent-calls',
    value: '--mm-max-concurrent-calls',
    options: []
  },
  {
    label: '--mm-per-request-timeout',
    value: '--mm-per-request-timeout',
    options: []
  },
  {
    label: '--enable-broadcast-mm-inputs-process',
    value: '--enable-broadcast-mm-inputs-process',
    options: []
  },
  {
    label: '--mm-process-config',
    value: '--mm-process-config',
    options: []
  },
  {
    label: '--mm-enable-dp-encoder',
    value: '--mm-enable-dp-encoder',
    options: []
  },
  {
    label: '--limit-mm-data-per-request',
    value: '--limit-mm-data-per-request',
    options: []
  },
  {
    label: '--enable-mm-global-cache',
    value: '--enable-mm-global-cache',
    options: []
  },
  // For checkpoint decryption
  {
    label: '--decrypted-config-file',
    value: '--decrypted-config-file',
    options: []
  },
  {
    label: '--decrypted-draft-config-file',
    value: '--decrypted-draft-config-file',
    options: []
  },
  {
    label: '--enable-prefix-mm-cache',
    value: '--enable-prefix-mm-cache',
    options: []
  }
];

export default options;