# ==========================================
# ACTUAL TEST CASES
# ==========================================

test_cases:
  - name: "MiniMax-M2.5-w8a8"
    model: "Eco-Tech/MiniMax-M2.5-w8a8-QuaRot"
    envs:
      HCCL_BUFFSIZE: "512"
      HCCL_OP_EXPANSION_MODE: "AIV"
      PYTORCH_NPU_ALLOC_CONF: "expandable_segments:True"
      VLLM_ASCEND_ENABLE_FLASHCOMM1: "1"
      OMP_NUM_THREADS: "1"
      TASK_QUEUE_ENABLE: "1"
      VLLM-ASCEND_BALANCE_SCHEDULING: "1"
      HCCL_INTRA_PCIE_ENABLE: "1"
      HCCL_INTRA_ROCE_ENABLE: "0"
      OMP_PROC_BIND: "false"
      VLLM_TORCH_PROFILER_WITH_STACK: "0"
      VLLM_TORCH_PROFILER_DIR: "./profile"
      VLLM_USE_MODELSCOPE: "true"
      SERVER_PORT: "DEFAULT_PORT"
    server_cmd:
      - "--tensor-parallel-size"
      - "8"
      - "--port"
      - "$SERVER_PORT"
      - "--trust-remote-code"
      - "--gpu-memory-utilization"
      - "0.9"
      - "--quantization"
      - "ascend"
      - "--additional-config"
      - '{"enable_cpu_binding":true}'
      - "--model-loader-extra-config"
      - '{"enable_multithread_load":true,"num_threads":16}'
      - "--speculative_config"
      - '{"method":"eagle3","model":"vllm-ascend/MiniMax-M2.5-eagle-model-0318","num_speculative_tokens":3}'
      - "--enable-expert-parallel"
      - "--enable-chunked-prefill"
      - "--enable-prefix-caching"
      - "--max-num-seqs"
      - "100"
      - "--max-model-len"
      - "196608"
      - "--seed"
      - "1024"
      - "--max-num-batched-tokens"
      - "6144"
      - "--enable-auto-tool-choice"
      - "--tool-call-parser"
      - "minimax_m2"
      - "--reasoning-parser"
      - "minimax_m2_append_think"
      - "--enable-force-include-usage"
      - "--profiler-config"
      - '{"profiler":"torch","torch_profiler_dir":"./profile","torch_profiler_with_stack":false}'
      - "--compilation-config"
      - '{"cudagraph_mode":"FULL_DECODE_ONLY","cudagraph_capture_sizes":[4,16,40,80,160,256,400]}'
    benchmarks:
      acc:
        case_type: accuracy
        dataset_path: vllm-ascend/gpqa
        request_conf: vllm_api_general_chat
        dataset_conf: gsm8k/gpqa_gen_0_shot_str
        max_out_len: 131072
        batch_size: 64
        baseline: 83
        threshold: 5
        bos_token_id: 200019
        do_sample: true
        eos_token_id: 200020
        temperature: 1.0
        top_p: 0.95
        top_k: 40
        transformers_version: 4.46.1
        ignore_eos: false
      perf:
        case_type: performance
        dataset_path: vllm-ascend/GSM8K-in3500-bs2800
        request_conf: vllm_api_stream_chat
        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
        num_prompts: 360
        max_out_len: 1500
        batch_size: 120
        request_rate: 0
        baseline: 2042
        threshold: 0.97