{"title":"Sampled Muse Spark Open Recipes Document Four-Input NVIDIA Paths and Text-Only AMD Setups","description":"Documented input support depends on the checkpoint, runtime and configuration—not merely whether the model fits the hardware. Missing modality documentation does not establish missing capability.","dataset_id":"spd:muse-spark-open-on-macs-and-gpus-which-recipes-document-multimodal-support-9a750b9c","canonical_url":"https://superpowerdaily.com/research/muse-spark-open-on-macs-and-gpus-which-recipes-document-multimodal-support-9a750b9c","version_url":"https://superpowerdaily.com/research/muse-spark-open-on-macs-and-gpus-which-recipes-document-multimodal-support-9a750b9c","version":"Live","snapshot_hash":null,"date_created":null,"date_modified":"2026-10-17T20:39:25.430Z","license":{"name":"Superpower Daily data reuse terms","url":"https://superpowerdaily.com/terms"},"license_url":"https://superpowerdaily.com/terms","coverage_note":"This is a bounded documentary comparison, not an exhaustive deployment inventory. It distinguishes usable documented paths from configuration-level exclusions and unanswered documentation questions. Recovery of saved evidence does not advance the original cutoff.","measurement_technique":["Reassess only the supplied saved evidence. This synthesis is not a new collection, experiment or performance comparison.","Preserve the original collection context: October 17, 2026, 15:30:55.058 UTC, with an intended evidence cutoff of October 17 at 15:30 UTC. The packet records snapshot retrieval at 15:32:10.286 UTC and subsequent live-page checks, so an exact historical freeze is not established.","Bound the sample to Meta’s hardware and inference guides, their directly linked runtime evidence, and the supplied Apple snapshot. Cover self-hosted inference only.","Plan the evidence matrix around checkpoint-runtime-hardware configurations, with separate text, image, audio and video columns. Classify each cell as explicitly documented, explicitly excluded by the sampled configuration, or unspecified. Preserve conflicting documentation as a flagged uncertainty rather than forcing a supported/unsupported verdict.","Keep native vLLM separate from its reference plugin, distributed mlx_lm separate from oMLX, and Dynamo aggregated deployments separate from disaggregated deployments. Do not inherit backend capabilities automatically.","Attach preprocessing requirements and configuration caveats to the relevant rows. Do not score performance, calculate coverage percentages or rank hardware."],"methodology":["Reassess only the supplied saved evidence. This synthesis is not a new collection, experiment or performance comparison.","Preserve the original collection context: October 17, 2026, 15:30:55.058 UTC, with an intended evidence cutoff of October 17 at 15:30 UTC. The packet records snapshot retrieval at 15:32:10.286 UTC and subsequent live-page checks, so an exact historical freeze is not established.","Bound the sample to Meta’s hardware and inference guides, their directly linked runtime evidence, and the supplied Apple snapshot. Cover self-hosted inference only.","Plan the evidence matrix around checkpoint-runtime-hardware configurations, with separate text, image, audio and video columns. Classify each cell as explicitly documented, explicitly excluded by the sampled configuration, or unspecified. Preserve conflicting documentation as a flagged uncertainty rather than forcing a supported/unsupported verdict.","Keep native vLLM separate from its reference plugin, distributed mlx_lm separate from oMLX, and Dynamo aggregated deployments separate from disaggregated deployments. Do not inherit backend capabilities automatically.","Attach preprocessing requirements and configuration caveats to the relevant rows. Do not score performance, calculate coverage percentages or rank hardware."],"metrics":[{"label":"Verified observations","value":"0","detail":"0 measured fields"},{"label":"Supported claims","value":"10","detail":"10 material findings"},{"label":"Cited sources","value":"7","detail":"7 primary or authoritative"},{"label":"Research score","value":"83","detail":"Automated topic and evidence score"}],"columns":[{"key":"entity","label":"Entity"},{"key":"metric","label":"Metric"},{"key":"value","label":"Value"},{"key":"unit","label":"Unit"},{"key":"observed","label":"Observed"},{"key":"source","label":"Source"},{"key":"transform","label":"Transform"}],"data":[],"sources":[{"name":"Meta","title":"dev.meta.ai","url":"https://dev.meta.ai/docs/muse-spark-open/deploy","records":0},{"name":"Meta","title":"Hardware platforms","url":"https://dev.meta.ai/docs/muse-spark-open/hardware-platforms","records":0},{"name":"Apple","title":"Local and Distributed Large LLM Inference with MLX and RDMA on Mac Studio","url":"https://machinelearning.apple.com/research/mlx-distributed-inference","records":0},{"name":"NVIDIA","title":"Muse-Spark-Open-816B-A42B | NVIDIA Dynamo Documentation","url":"https://docs.nvidia.com/dynamo/dev/recipes/Muse-Spark-Open","records":0},{"name":"vLLM","title":"Muse-Spark-Open-816B-A42B on vLLM — Serve command for H200, GB200 NVL4, GB300 NVL4, MI350X, MI355X","url":"https://recipes.vllm.ai/meta-models/Muse-Spark-Open-816B-A42B","records":0},{"name":"NVIDIA","title":"TensorRT-LLM/docs/source/deployment-guide/deployment-guide-for-musespark-open-on-trtllm.md at feat/musespark-open · NVIDIA/TensorRT-LLM","url":"https://github.com/NVIDIA/TensorRT-LLM/blob/feat/musespark-open/docs/source/deployment-guide/deployment-guide-for-musespark-open-on-trtllm.md","records":0},{"name":"Hugging Face","title":"Welcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta","url":"https://huggingface.co/blog/meta-muse-spark-open","records":0}],"provenance":{"publisher":"Superpower Daily","source_count":7,"source_urls":["https://dev.meta.ai/docs/muse-spark-open/deploy","https://dev.meta.ai/docs/muse-spark-open/hardware-platforms","https://machinelearning.apple.com/research/mlx-distributed-inference","https://docs.nvidia.com/dynamo/dev/recipes/Muse-Spark-Open","https://recipes.vllm.ai/meta-models/Muse-Spark-Open-816B-A42B","https://github.com/NVIDIA/TensorRT-LLM/blob/feat/musespark-open/docs/source/deployment-guide/deployment-guide-for-musespark-open-on-trtllm.md","https://huggingface.co/blog/meta-muse-spark-open"],"methodology":["Reassess only the supplied saved evidence. This synthesis is not a new collection, experiment or performance comparison.","Preserve the original collection context: October 17, 2026, 15:30:55.058 UTC, with an intended evidence cutoff of October 17 at 15:30 UTC. The packet records snapshot retrieval at 15:32:10.286 UTC and subsequent live-page checks, so an exact historical freeze is not established.","Bound the sample to Meta’s hardware and inference guides, their directly linked runtime evidence, and the supplied Apple snapshot. Cover self-hosted inference only.","Plan the evidence matrix around checkpoint-runtime-hardware configurations, with separate text, image, audio and video columns. Classify each cell as explicitly documented, explicitly excluded by the sampled configuration, or unspecified. Preserve conflicting documentation as a flagged uncertainty rather than forcing a supported/unsupported verdict.","Keep native vLLM separate from its reference plugin, distributed mlx_lm separate from oMLX, and Dynamo aggregated deployments separate from disaggregated deployments. Do not inherit backend capabilities automatically.","Attach preprocessing requirements and configuration caveats to the relevant rows. Do not score performance, calculate coverage percentages or rank hardware."],"snapshot_hash":null},"distributions":{"csv":"https://superpowerdaily.com/api/research/muse-spark-open-on-macs-and-gpus-which-recipes-document-multimodal-support-9a750b9c?format=csv","json":"https://superpowerdaily.com/api/research/muse-spark-open-on-macs-and-gpus-which-recipes-document-multimodal-support-9a750b9c?format=json"}}