{"record":{"id":"793778300a09e632","repo":"sgl-project/sglang","slug":"minimax-h3-quality-high-is-validated-only-for-th","errorCode":null,"errorMessage":"MiniMax-H3 quality=\"high\" is validated only for the strict 4xH200 fl2va deployment; mismatches: {mismatches}","messagePattern":"MiniMax-H3 quality=\"high\" is validated only for the strict 4xH200 fl2va deployment; mismatches: (.+?)","errorType":"validation","errorClass":"ValueError","httpStatus":null,"severity":"error","filePath":"python/sglang/multimodal_gen/configs/pipeline_configs/minimax_h3.py","lineNumber":221,"sourceCode":"            name: {\"expected\": wanted, \"actual\": actual[name]}\n            for name, wanted in expected.items()\n            if (\n                actual[name] not in wanted\n                if isinstance(wanted, set)\n                else actual[name] != wanted\n            )\n        }\n        if (\n            not current_platform.is_cuda()\n            or \"H200\" not in device_name.upper()\n            or capability_int != 90\n        ):\n            mismatches[\"device\"] = {\n                \"expected\": \"NVIDIA H200 (compute capability 9.0)\",\n                \"actual\": f\"{device_name} (compute capability {capability_int})\",\n            }\n        if mismatches:\n            raise ValueError(\n                'MiniMax-H3 quality=\"high\" is validated only for '\n                f\"the strict 4xH200 fl2va deployment; mismatches: {mismatches}\"\n            )\n\n    def validate_server_args(self, server_args) -> None:\n        # Reject known-inexact VAE modes before any large component download.\n        self.vae_config.resolved_parallel_decode_mode()\n        if current_platform.is_mps():\n            required_components = (\n                \"transformer\",\n                \"text_encoder\",\n                \"video_vae\",\n                \"audio_vae\",\n            )\n            missing_components = [\n                component\n                for component in required_components\n                if server_args.residency_mode(component) != LAYERWISE_OFFLOAD","sourceCodeStart":203,"sourceCodeEnd":239,"githubUrl":"https://github.com/sgl-project/sglang/blob/0132848349585cfe6aae51c4941cbae872505f8a/python/sglang/multimodal_gen/configs/pipeline_configs/minimax_h3.py#L203-L239","documentation":"MiniMax-H3 quality=\"high\" mode is only validated on a strict 4x NVIDIA H200 deployment using the fl2va stack. validate_quality_deployment collects hardware/backend mismatches (device model, compute capability) and raises with a dict of them when the actual environment differs. This guards against silently degraded or numerically wrong high-quality output on unvalidated hardware.","triggerScenarios":"Setting quality=\"high\" on MiniMax-H3 and running on anything other than 4x H200 GPUs (e.g. A100s, H100s, different GPU count, or non-fl2va backend) — mismatches are reported in the error dict.","commonSituations":"Trying quality=\"high\" on available cluster hardware (A100/H100), CI/dev machines, or after changing attention/VAE backends away from fl2va; also triggered deliberately by a regression test rejecting transformer weight overrides.","solutions":["Use the default (non-high) quality setting on non-H200 hardware","Run on a strict 4x H200 deployment with the fl2va configuration if quality=\"high\" is required","Check the mismatches dict in the message to see exactly which expectation (device, capability) failed and address that specific dimension"],"exampleFix":"# before\nconfig = MiniMaxH3Config(quality=\"high\")  # on 8xA100\n\n# after\nconfig = MiniMaxH3Config(quality=\"default\")  # validated on this hardware","handlingStrategy":"validation","validationCode":"import torch\nif quality == \"high\":\n    ok = torch.cuda.device_count() == 4 and all(torch.cuda.get_device_capability(i)[0] == 9 for i in range(4))\n    assert ok, \"quality='high' requires strict 4xH200\"","typeGuard":null,"tryCatchPattern":null,"preventionTips":["Gate quality=\"high\" behind a hardware capability check in your launcher","Default to standard quality outside validated deployments"],"tags":["minimax-h3","hardware-validation","quality-mode","video-generation"],"backgroundTag":"hardware-requirements-not-met","analyzedSha":"0132848349585cfe6aae51c4941cbae872505f8a","analyzedAt":"2026-08-28T05:10:05.995Z","schemaVersion":2},"datasetVersion":"2026-08-28T06:17:29.519Z"}