Uploaded image for project: 'Red Hat Enterprise Linux AI'
  1. Red Hat Enterprise Linux AI
  2. RHELAI-2610

mmlu eval and mmlu-branch eval failing on RHEL AI 1.3

XMLWordPrintable

    • Critical

      Run MMLU eval on the granite-8b-starter model

      ilab model evaluate --model /var/mnt/instg1/instructlab/models/granite-8b-starter/ --benchmark mmlu --gpus 8 --enable-serving-output

      It will fail with a 500 Internal Server error: issue documented in community issue: https://github.com/instructlab/eval/issues/195

      Expected behavior

      • Expect MMLU eval to successfully run and output results 

      Screenshots

      • Attached Image 

      Device Info (please complete the following information):

      • Hardware Specs: 8xA100 IBM Cloud VSI 
      • OS Version: RHEL AI 1.3
      • InstructLab Version: 0.21.0
      • Provide the output of these two commands:
        • "registry.redhat.io/rhelai1/bootc-ibm-nvidia-rhel9:1.3"

      ```

        • [root@tyler-machine-boot-6 ~]# ilab system info

      Platform:

        sys.version: 3.11.7 (main, Oct  9 2024, 00:00:00) [GCC 11.4.1 20231218 (Red Hat 11.4.1-3)]

        sys.platform: linux

        os.name: posix

        platform.release: 5.14.0-427.42.1.el9_4.x86_64

        platform.machine: x86_64

        platform.node: tyler-machine-boot-6

        platform.python_version: 3.11.7

        os-release.ID: rhel

        os-release.VERSION_ID: 9.4

        os-release.PRETTY_NAME: Red Hat Enterprise Linux 9.4 (Plow)

        memory.total: 1259.87 GB

        memory.available: 1246.53 GB

        memory.used: 5.41 GB

       

      InstructLab:

        instructlab.version: 0.21.0

        instructlab-dolomite.version: 0.2.0

        instructlab-eval.version: 0.4.1

        instructlab-quantize.version: 0.1.0

        instructlab-schema.version: 0.4.1

        instructlab-sdg.version: 0.6.1

        instructlab-training.version: 0.6.1

       

      Torch:

        torch.version: 2.4.1

        torch.backends.cpu.capability: AVX512

        torch.version.cuda: 12.4

        torch.version.hip: None

        torch.cuda.available: True

        torch.backends.cuda.is_built: True

        torch.backends.mps.is_built: False

        torch.backends.mps.is_available: False

        torch.cuda.bf16: True

        torch.cuda.current.device: 0

        torch.cuda.0.name: NVIDIA A100-SXM4-80GB

        torch.cuda.0.free: 78.7 GB

        torch.cuda.0.total: 79.1 GB

        torch.cuda.0.capability: 8.0 (see https://developer.nvidia.com/cuda-gpus#compute)

        torch.cuda.1.name: NVIDIA A100-SXM4-80GB

        torch.cuda.1.free: 78.7 GB

        torch.cuda.1.total: 79.1 GB

        torch.cuda.1.capability: 8.0 (see https://developer.nvidia.com/cuda-gpus#compute)

        torch.cuda.2.name: NVIDIA A100-SXM4-80GB

        torch.cuda.2.free: 78.7 GB

        torch.cuda.2.total: 79.1 GB

        torch.cuda.2.capability: 8.0 (see https://developer.nvidia.com/cuda-gpus#compute)

        torch.cuda.3.name: NVIDIA A100-SXM4-80GB

        torch.cuda.3.free: 78.7 GB

        torch.cuda.3.total: 79.1 GB

        torch.cuda.3.capability: 8.0 (see https://developer.nvidia.com/cuda-gpus#compute)

        torch.cuda.4.name: NVIDIA A100-SXM4-80GB

        torch.cuda.4.free: 78.7 GB

        torch.cuda.4.total: 79.1 GB

        torch.cuda.4.capability: 8.0 (see https://developer.nvidia.com/cuda-gpus#compute)

        torch.cuda.5.name: NVIDIA A100-SXM4-80GB

        torch.cuda.5.free: 78.7 GB

        torch.cuda.5.total: 79.1 GB

        torch.cuda.5.capability: 8.0 (see https://developer.nvidia.com/cuda-gpus#compute)

        torch.cuda.6.name: NVIDIA A100-SXM4-80GB

        torch.cuda.6.free: 78.7 GB

        torch.cuda.6.total: 79.1 GB

        torch.cuda.6.capability: 8.0 (see https://developer.nvidia.com/cuda-gpus#compute)

        torch.cuda.7.name: NVIDIA A100-SXM4-80GB

        torch.cuda.7.free: 78.7 GB

        torch.cuda.7.total: 79.1 GB

        torch.cuda.7.capability: 8.0 (see https://developer.nvidia.com/cuda-gpus#compute)

       

      llama_cpp_python:

        llama_cpp_python.version: 0.2.79

        llama_cpp_python.supports_gpu_offload: True
      ```

      Additional context

      • <your text here>

              osilkin@redhat.com Oleg Silkin
              lisowskiibm Tyler Lisowski
              Mustafa Eyceoz
              Votes:
              0 Vote for this issue
              Watchers:
              7 Start watching this issue

                Created:
                Updated: