{"default":{"hw":"m_24","mode":"server","variant":"default"},"family":"afmoe","features":[{"default":"off","id":"api_key","modes":["server","disaggregated"]},{"default":"off","id":"estimate_memory","modes":["cli_run","cli_generate"]},{"default":"off","id":"lora","modes":["server","cli_run","cli_generate","disaggregated"]},{"default":"off","id":"metrics","modes":["server","disaggregated"]},{"default":"on","id":"tool_calling","modes":["server","cli_run","cli_generate","disaggregated"]}],"hardware":{"a100_80":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"cuda","commands":2,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"a100_80"}}]},"cli_run":{"backend":"cuda","commands":1,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"a100_80"}}]},"disaggregated":{"backend":"cuda","commands":3,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"a100_80"}}]},"server":{"backend":"cuda","commands":0,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"a100_80"}}]}}},"a100_80x2":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"cuda","commands":2,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80x2","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"a100_80x2"}}]},"cli_run":{"backend":"cuda","commands":1,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80x2","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"a100_80x2"}}]},"disaggregated":{"backend":"cuda","commands":3,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80x2","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"a100_80x2"}}]},"server":{"backend":"cuda","commands":0,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80x2","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"a100_80x2"}}]}}},"a100_80x4":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"cuda","commands":2,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":288,"drafterGb":0,"headroomGb":23.04,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80x4","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"a100_80x4"}}]},"cli_run":{"backend":"cuda","commands":1,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":288,"drafterGb":0,"headroomGb":23.04,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80x4","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"a100_80x4"}}]},"disaggregated":{"backend":"cuda","commands":3,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":288,"drafterGb":0,"headroomGb":23.04,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80x4","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"a100_80x4"}}]},"server":{"backend":"cuda","commands":0,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":288,"drafterGb":0,"headroomGb":23.04,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"a100_80x4","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"a100_80x4"}}]}}},"gb10_122":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"cuda","commands":6,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":91.5,"drafterGb":0,"headroomGb":7.32,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"gb10_122","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]},"cli_run":{"backend":"cuda","commands":5,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":91.5,"drafterGb":0,"headroomGb":7.32,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"gb10_122","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]},"disaggregated":{"backend":"cuda","commands":7,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":91.5,"drafterGb":0,"headroomGb":7.32,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"gb10_122","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"}]},"server":{"backend":"cuda","commands":4,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":91.5,"drafterGb":0,"headroomGb":7.32,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"gb10_122","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]}}},"h100_80":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"cuda","commands":2,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"h100_80"}}]},"cli_run":{"backend":"cuda","commands":1,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"h100_80"}}]},"disaggregated":{"backend":"cuda","commands":3,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"h100_80"}}]},"server":{"backend":"cuda","commands":0,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"h100_80"}}]}}},"h100_80x2":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"cuda","commands":2,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80x2","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"h100_80x2"}}]},"cli_run":{"backend":"cuda","commands":1,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80x2","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"h100_80x2"}}]},"disaggregated":{"backend":"cuda","commands":3,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80x2","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"h100_80x2"}}]},"server":{"backend":"cuda","commands":0,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80x2","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"h100_80x2"}}]}}},"h100_80x8":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"cuda","commands":2,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":576,"drafterGb":0,"headroomGb":46.08,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80x8","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"h100_80x8"}}]},"cli_run":{"backend":"cuda","commands":1,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":576,"drafterGb":0,"headroomGb":46.08,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80x8","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"h100_80x8"}}]},"disaggregated":{"backend":"cuda","commands":3,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":576,"drafterGb":0,"headroomGb":46.08,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80x8","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"h100_80x8"}}]},"server":{"backend":"cuda","commands":0,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":576,"drafterGb":0,"headroomGb":46.08,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"h100_80x8","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"h100_80x8"}}]}}},"m1ultra_128":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m1ultra_128","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m1ultra_128","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m1ultra_128","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m1ultra_128","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]}}},"m5max_128":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m5max_128","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m5max_128","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m5max_128","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m5max_128","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]}}},"m_128":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_128","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_128"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_128","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_128"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_128","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_128"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":96,"drafterGb":0,"headroomGb":7.68,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_128","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_128"}}]}}},"m_16":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":10.72,"drafterGb":0,"headroomGb":1,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_16","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_16"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":10.72,"drafterGb":0,"headroomGb":1,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_16","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_16"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":10.72,"drafterGb":0,"headroomGb":1,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_16","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_16"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":10.72,"drafterGb":0,"headroomGb":1,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_16","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_16"}}]}}},"m_192":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_192","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_192"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_192","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_192"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_192","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_192"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":144,"drafterGb":0,"headroomGb":11.52,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_192","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_192"}}]}}},"m_24":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":16.08,"drafterGb":0,"headroomGb":1.29,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_24","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_24"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":16.08,"drafterGb":0,"headroomGb":1.29,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_24","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_24"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":16.08,"drafterGb":0,"headroomGb":1.29,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_24","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_24"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":16.08,"drafterGb":0,"headroomGb":1.29,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_24","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_24"}}]}}},"m_256":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":192,"drafterGb":0,"headroomGb":15.36,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_256","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_256"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":192,"drafterGb":0,"headroomGb":15.36,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_256","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_256"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":192,"drafterGb":0,"headroomGb":15.36,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_256","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_256"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":192,"drafterGb":0,"headroomGb":15.36,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_256","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_256"}}]}}},"m_32":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":21.44,"drafterGb":0,"headroomGb":1.72,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_32","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_32"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":21.44,"drafterGb":0,"headroomGb":1.72,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_32","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_32"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":21.44,"drafterGb":0,"headroomGb":1.72,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_32","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_32"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":21.44,"drafterGb":0,"headroomGb":1.72,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_32","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_32"}}]}}},"m_36":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":24.12,"drafterGb":0,"headroomGb":1.93,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_36","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_36"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":24.12,"drafterGb":0,"headroomGb":1.93,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_36","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_36"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":24.12,"drafterGb":0,"headroomGb":1.93,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_36","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_36"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":24.12,"drafterGb":0,"headroomGb":1.93,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_36","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_36"}}]}}},"m_48":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":36,"drafterGb":0,"headroomGb":2.88,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_48","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_48"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":36,"drafterGb":0,"headroomGb":2.88,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_48","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_48"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":36,"drafterGb":0,"headroomGb":2.88,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_48","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_48"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":36,"drafterGb":0,"headroomGb":2.88,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_48","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_48"}}]}}},"m_512":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":384,"drafterGb":0,"headroomGb":30.72,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_512","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_512"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":384,"drafterGb":0,"headroomGb":30.72,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_512","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_512"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":384,"drafterGb":0,"headroomGb":30.72,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_512","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_512"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":384,"drafterGb":0,"headroomGb":30.72,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_512","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_512"}}]}}},"m_64":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":48,"drafterGb":0,"headroomGb":3.84,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_64","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_64"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":48,"drafterGb":0,"headroomGb":3.84,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_64","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_64"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":48,"drafterGb":0,"headroomGb":3.84,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_64","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_64"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":48,"drafterGb":0,"headroomGb":3.84,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_64","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_64"}}]}}},"m_8":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":5.36,"drafterGb":0,"headroomGb":1,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_8","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_8"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":5.36,"drafterGb":0,"headroomGb":1,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_8","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_8"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":5.36,"drafterGb":0,"headroomGb":1,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_8","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_8"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":5.36,"drafterGb":0,"headroomGb":1,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_8","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_8"}}]}}},"m_96":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"metal","commands":10,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_96","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_96"}}]},"cli_run":{"backend":"metal","commands":9,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_96","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_96"}}]},"disaggregated":{"backend":"metal","commands":11,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_96","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"m_96"}}]},"server":{"backend":"metal","commands":8,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":72,"drafterGb":0,"headroomGb":5.76,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"m_96","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"m_96"}}]}}},"rtx_4090_24":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"cuda","commands":2,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":21.6,"drafterGb":0,"headroomGb":1.73,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_4090_24","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"rtx_4090_24"}}]},"cli_run":{"backend":"cuda","commands":1,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":21.6,"drafterGb":0,"headroomGb":1.73,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_4090_24","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"rtx_4090_24"}}]},"disaggregated":{"backend":"cuda","commands":3,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":21.6,"drafterGb":0,"headroomGb":1.73,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_4090_24","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"rtx_4090_24"}}]},"server":{"backend":"cuda","commands":0,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":21.6,"drafterGb":0,"headroomGb":1.73,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_4090_24","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"rtx_4090_24"}}]}}},"rtx_4090_24x2":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"cuda","commands":2,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":43.2,"drafterGb":0,"headroomGb":3.46,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_4090_24x2","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"rtx_4090_24x2"}}]},"cli_run":{"backend":"cuda","commands":1,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":43.2,"drafterGb":0,"headroomGb":3.46,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_4090_24x2","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"rtx_4090_24x2"}}]},"disaggregated":{"backend":"cuda","commands":3,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":43.2,"drafterGb":0,"headroomGb":3.46,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_4090_24x2","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"rtx_4090_24x2"}}]},"server":{"backend":"cuda","commands":0,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":43.2,"drafterGb":0,"headroomGb":3.46,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_4090_24x2","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"rtx_4090_24x2"}}]}}},"rtx_5090_32":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"cuda","commands":2,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":28.8,"drafterGb":0,"headroomGb":2.3,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_5090_32","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"rtx_5090_32"}}]},"cli_run":{"backend":"cuda","commands":1,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":28.8,"drafterGb":0,"headroomGb":2.3,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_5090_32","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"rtx_5090_32"}}]},"disaggregated":{"backend":"cuda","commands":3,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":28.8,"drafterGb":0,"headroomGb":2.3,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_5090_32","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"},{"code":"not_measured_on_tier","params":{"hw":"rtx_5090_32"}}]},"server":{"backend":"cuda","commands":0,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":28.8,"drafterGb":0,"headroomGb":2.3,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"rtx_5090_32","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"not_measured_on_tier","params":{"hw":"rtx_5090_32"}}]}}},"v100_32":{"default":{"mode":"server","variant":"default"},"features":0,"renderings":{"cli_generate":{"backend":"cuda","commands":2,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":28.8,"drafterGb":0,"headroomGb":2.3,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"v100_32","mode":"cli_generate","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]},"cli_run":{"backend":"cuda","commands":1,"expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":28.8,"drafterGb":0,"headroomGb":2.3,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"v100_32","mode":"cli_run","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]},"disaggregated":{"backend":"cuda","commands":3,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit \\\n  mlxcel-server --node-role decode --serving-bind 192.168.1.21:9100","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":28.8,"drafterGb":0,"headroomGb":2.3,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"v100_32","mode":"disaggregated","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}},{"code":"disagg_placeholder_addresses"}]},"server":{"backend":"cuda","commands":0,"envText":"LLAMA_ARG_MODEL=mlx-community/Trinity-Nano-Preview-4bit LLAMA_ARG_ALIAS=trinity-nano-preview-4bit LLAMA_ARG_CTX_SIZE=8192 LLAMA_ARG_N_PARALLEL=4 \\\n  mlxcel-server","expect":{"kind":"none"},"features":["tool_calling"],"fit":{"alternatives":[],"budgetGb":28.8,"drafterGb":0,"headroomGb":2.3,"kvGb":0.44,"source":"approx","state":"fits","weightsGb":0},"hw":"v100_32","mode":"server","snippets":0,"variant":"default","warnings":[{"code":"parameter_count_unknown","params":{"model_id":"mlx-community/Trinity-Nano-Preview-4bit"}}]}}}},"i18n":{"ko":{"description":"Trinity-Nano-Preview-4bit: mlxcel의 텍스트 생성·API 서빙 명령을 구성하는 레시피입니다.","title":"Trinity-Nano-Preview-4bit"}},"model_id":"mlx-community/Trinity-Nano-Preview-4bit","modes":["server","cli_run","cli_generate","disaggregated"],"schema_version":2,"shared":{"commands":[{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-x86_64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-x86_64-cuda13/bin:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-x86_64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-x86_64-cuda13/bin:$PATH\" \\\n  && mlxcel --version","title":"Linux CUDA archive"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-linux","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (Linux)"}],"run":[{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--alias","trinity-nano-preview-4bit","--ctx-size","8192","--n-parallel","4"],"env":{"LLAMA_ARG_ALIAS":"trinity-nano-preview-4bit","LLAMA_ARG_CTX_SIZE":"8192","LLAMA_ARG_MODEL":"mlx-community/Trinity-Nano-Preview-4bit","LLAMA_ARG_N_PARALLEL":"4"},"id":"server","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit --alias trinity-nano-preview-4bit \\\n  --ctx-size 8192 --n-parallel 4","title":"mlxcel-server"}]},{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-x86_64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-x86_64-cuda13/bin:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-x86_64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-x86_64-cuda13/bin:$PATH\" \\\n  && mlxcel --version","title":"Linux CUDA archive"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-linux","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (Linux)"}],"run":[{"argv":["mlxcel","run","mlx-community/Trinity-Nano-Preview-4bit"],"id":"cli_run","text":"mlxcel run mlx-community/Trinity-Nano-Preview-4bit","title":"mlxcel run"}]},{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-x86_64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-x86_64-cuda13/bin:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-x86_64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-x86_64-cuda13/bin:$PATH\" \\\n  && mlxcel --version","title":"Linux CUDA archive"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-linux","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (Linux)"}],"run":[{"argv":["mlxcel","generate","-m","mlx-community/Trinity-Nano-Preview-4bit","-p","Explain unified memory in one paragraph.","-n","256"],"id":"cli_generate","text":"mlxcel generate -m mlx-community/Trinity-Nano-Preview-4bit -p 'Explain unified memory in one paragraph.' -n 256","title":"mlxcel generate"}]},{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-x86_64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-x86_64-cuda13/bin:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-x86_64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-x86_64-cuda13/bin:$PATH\" \\\n  && mlxcel --version","title":"Linux CUDA archive"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-linux","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (Linux)"}],"run":[{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--node-role","decode","--serving-bind","192.168.1.21:9100"],"env":{"LLAMA_ARG_MODEL":"mlx-community/Trinity-Nano-Preview-4bit"},"id":"disagg-decode","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit \\\n  --node-role decode --serving-bind 192.168.1.21:9100","title":"Decode node"},{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--node-role","prefill","--serving-bind","192.168.1.20:9100","--decode-peers","192.168.1.21:9100"],"id":"disagg-prefill","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit \\\n  --node-role prefill --serving-bind 192.168.1.20:9100 --decode-peers 192.168.1.21:9100","title":"Prefill node"},{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--alias","trinity-nano-preview-4bit","--port","8080","--node-role","router","--serving-bind","192.168.1.10:9100","--prefill-peers","192.168.1.20:9100","--decode-peers","192.168.1.21:9100","--no-warmup"],"id":"disagg-router","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit --alias trinity-nano-preview-4bit \\\n  --port 8080 \\\n  --node-role router --serving-bind 192.168.1.10:9100 --prefill-peers 192.168.1.20:9100 --decode-peers 192.168.1.21:9100 --no-warmup","title":"Router"}]},{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-aarch64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-aarch64-cuda13/bin:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-aarch64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-aarch64-cuda13/bin:$PATH\" \\\n  && mlxcel --version","title":"Linux CUDA archive"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-linux","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (Linux)"}],"run":[{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--alias","trinity-nano-preview-4bit","--ctx-size","8192","--n-parallel","4"],"env":{"LLAMA_ARG_ALIAS":"trinity-nano-preview-4bit","LLAMA_ARG_CTX_SIZE":"8192","LLAMA_ARG_MODEL":"mlx-community/Trinity-Nano-Preview-4bit","LLAMA_ARG_N_PARALLEL":"4"},"id":"server","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit --alias trinity-nano-preview-4bit \\\n  --ctx-size 8192 --n-parallel 4","title":"mlxcel-server"}]},{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-aarch64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-aarch64-cuda13/bin:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-aarch64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-aarch64-cuda13/bin:$PATH\" \\\n  && mlxcel --version","title":"Linux CUDA archive"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-linux","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (Linux)"}],"run":[{"argv":["mlxcel","run","mlx-community/Trinity-Nano-Preview-4bit"],"id":"cli_run","text":"mlxcel run mlx-community/Trinity-Nano-Preview-4bit","title":"mlxcel run"}]},{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-aarch64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-aarch64-cuda13/bin:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-aarch64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-aarch64-cuda13/bin:$PATH\" \\\n  && mlxcel --version","title":"Linux CUDA archive"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-linux","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (Linux)"}],"run":[{"argv":["mlxcel","generate","-m","mlx-community/Trinity-Nano-Preview-4bit","-p","Explain unified memory in one paragraph.","-n","256"],"id":"cli_generate","text":"mlxcel generate -m mlx-community/Trinity-Nano-Preview-4bit -p 'Explain unified memory in one paragraph.' -n 256","title":"mlxcel generate"}]},{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-aarch64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-aarch64-cuda13/bin:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-linux-aarch64-cuda13.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-linux-aarch64-cuda13/bin:$PATH\" \\\n  && mlxcel --version","title":"Linux CUDA archive"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-linux","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features cuda && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (Linux)"}],"run":[{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--node-role","decode","--serving-bind","192.168.1.21:9100"],"env":{"LLAMA_ARG_MODEL":"mlx-community/Trinity-Nano-Preview-4bit"},"id":"disagg-decode","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit \\\n  --node-role decode --serving-bind 192.168.1.21:9100","title":"Decode node"},{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--node-role","prefill","--serving-bind","192.168.1.20:9100","--decode-peers","192.168.1.21:9100"],"id":"disagg-prefill","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit \\\n  --node-role prefill --serving-bind 192.168.1.20:9100 --decode-peers 192.168.1.21:9100","title":"Prefill node"},{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--alias","trinity-nano-preview-4bit","--port","8080","--node-role","router","--serving-bind","192.168.1.10:9100","--prefill-peers","192.168.1.20:9100","--decode-peers","192.168.1.21:9100","--no-warmup"],"id":"disagg-router","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit --alias trinity-nano-preview-4bit \\\n  --port 8080 \\\n  --node-role router --serving-bind 192.168.1.10:9100 --prefill-peers 192.168.1.20:9100 --decode-peers 192.168.1.21:9100 --no-warmup","title":"Router"}]},{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-macos-aarch64.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-macos-aarch64.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64:$PATH\" \\\n  && mlxcel --version","title":"macOS archive"},{"argv":["sh","-lc","brew tap lablup/tap && brew install mlxcel"],"id":"brew","text":"brew tap lablup/tap && brew install mlxcel","title":"Homebrew"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features metal,accelerate && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-macos","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features metal,accelerate && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (macOS)"},{"argv":["brew","install","mlxcel"],"id":"brew-command","text":"brew install mlxcel","title":"Brew install"}],"run":[{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--alias","trinity-nano-preview-4bit","--ctx-size","8192","--n-parallel","4"],"env":{"LLAMA_ARG_ALIAS":"trinity-nano-preview-4bit","LLAMA_ARG_CTX_SIZE":"8192","LLAMA_ARG_MODEL":"mlx-community/Trinity-Nano-Preview-4bit","LLAMA_ARG_N_PARALLEL":"4"},"id":"server","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit --alias trinity-nano-preview-4bit \\\n  --ctx-size 8192 --n-parallel 4","title":"mlxcel-server"}]},{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-macos-aarch64.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-macos-aarch64.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64:$PATH\" \\\n  && mlxcel --version","title":"macOS archive"},{"argv":["sh","-lc","brew tap lablup/tap && brew install mlxcel"],"id":"brew","text":"brew tap lablup/tap && brew install mlxcel","title":"Homebrew"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features metal,accelerate && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-macos","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features metal,accelerate && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (macOS)"},{"argv":["brew","install","mlxcel"],"id":"brew-command","text":"brew install mlxcel","title":"Brew install"}],"run":[{"argv":["mlxcel","run","mlx-community/Trinity-Nano-Preview-4bit"],"id":"cli_run","text":"mlxcel run mlx-community/Trinity-Nano-Preview-4bit","title":"mlxcel run"}]},{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-macos-aarch64.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-macos-aarch64.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64:$PATH\" \\\n  && mlxcel --version","title":"macOS archive"},{"argv":["sh","-lc","brew tap lablup/tap && brew install mlxcel"],"id":"brew","text":"brew tap lablup/tap && brew install mlxcel","title":"Homebrew"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features metal,accelerate && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-macos","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features metal,accelerate && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (macOS)"},{"argv":["brew","install","mlxcel"],"id":"brew-command","text":"brew install mlxcel","title":"Brew install"}],"run":[{"argv":["mlxcel","generate","-m","mlx-community/Trinity-Nano-Preview-4bit","-p","Explain unified memory in one paragraph.","-n","256"],"id":"cli_generate","text":"mlxcel generate -m mlx-community/Trinity-Nano-Preview-4bit -p 'Explain unified memory in one paragraph.' -n 256","title":"mlxcel generate"}]},{"download":{"argv":["mlxcel","download","mlx-community/Trinity-Nano-Preview-4bit"],"id":"download","text":"mlxcel download mlx-community/Trinity-Nano-Preview-4bit","title":"Download checkpoint"},"install":[{"argv":["sh","-lc","MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-macos-aarch64.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64:$PATH\" \\\n  && mlxcel --version"],"id":"archive","text":"MLXCEL_INSTALL_DIR=\"${MLXCEL_INSTALL_DIR:-$HOME/.local/share/mlxcel}\" \\\n  && mkdir -p \"$MLXCEL_INSTALL_DIR\" \\\n  && curl -fL -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" https://github.com/lablup/mlxcel/releases/download/v0.8.1/mlxcel-macos-aarch64.zip \\\n  && unzip -o \"$MLXCEL_INSTALL_DIR/mlxcel.zip\" -d \"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64\" \\\n  && export PATH=\"$MLXCEL_INSTALL_DIR/mlxcel-macos-aarch64:$PATH\" \\\n  && mlxcel --version","title":"macOS archive"},{"argv":["sh","-lc","brew tap lablup/tap && brew install mlxcel"],"id":"brew","text":"brew tap lablup/tap && brew install mlxcel","title":"Homebrew"},{"argv":["sh","-lc","git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features metal,accelerate && export PATH=\"$PWD/target/release:$PATH\""],"id":"source-macos","text":"git clone https://github.com/lablup/mlxcel.git && cd mlxcel && cargo build --release --features metal,accelerate && export PATH=\"$PWD/target/release:$PATH\"","title":"Build from source (macOS)"},{"argv":["brew","install","mlxcel"],"id":"brew-command","text":"brew install mlxcel","title":"Brew install"}],"run":[{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--node-role","decode","--serving-bind","192.168.1.21:9100"],"env":{"LLAMA_ARG_MODEL":"mlx-community/Trinity-Nano-Preview-4bit"},"id":"disagg-decode","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit \\\n  --node-role decode --serving-bind 192.168.1.21:9100","title":"Decode node"},{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--node-role","prefill","--serving-bind","192.168.1.20:9100","--decode-peers","192.168.1.21:9100"],"id":"disagg-prefill","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit \\\n  --node-role prefill --serving-bind 192.168.1.20:9100 --decode-peers 192.168.1.21:9100","title":"Prefill node"},{"argv":["mlxcel-server","-m","mlx-community/Trinity-Nano-Preview-4bit","--alias","trinity-nano-preview-4bit","--port","8080","--node-role","router","--serving-bind","192.168.1.10:9100","--prefill-peers","192.168.1.20:9100","--decode-peers","192.168.1.21:9100","--no-warmup"],"id":"disagg-router","text":"mlxcel-server -m mlx-community/Trinity-Nano-Preview-4bit --alias trinity-nano-preview-4bit \\\n  --port 8080 \\\n  --node-role router --serving-bind 192.168.1.10:9100 --prefill-peers 192.168.1.20:9100 --decode-peers 192.168.1.21:9100 --no-warmup","title":"Router"}]}],"features":[[{"allowed_modes":["server","disaggregated"],"cli_args":[],"id":"api_key","memory_effect":"none","server_args":["--api-key","<key>"]},{"allowed_modes":["cli_run","cli_generate"],"cli_args":["--estimate-memory"],"id":"estimate_memory","memory_effect":"none","server_args":[]},{"allowed_modes":["server","cli_run","cli_generate","disaggregated"],"cli_args":["--adapter","<path>"],"id":"lora","memory_effect":"none","server_args":["--adapter","<path>"]},{"allowed_modes":["server","disaggregated"],"cli_args":[],"id":"metrics","memory_effect":"none","server_args":["--metrics"]},{"allowed_modes":["server","cli_run","cli_generate","disaggregated"],"cli_args":[],"id":"tool_calling","memory_effect":"none","server_args":[]}]],"snippets":[{"anthropic":"curl http://127.0.0.1:8080/v1/messages \\\n  -H \"Content-Type: application/json\" \\\n  -d '{\"model\":\"trinity-nano-preview-4bit\",\"max_tokens\":128,\"messages\":[{\"role\":\"user\",\"content\":\"Hello\"}]}'","curl":"curl http://127.0.0.1:8080/v1/chat/completions \\\n  -H \"Content-Type: application/json\" \\\n  -d '{\"model\":\"trinity-nano-preview-4bit\",\"messages\":[{\"role\":\"user\",\"content\":\"Hello\"}]}'","javascript":"import OpenAI from \"openai\";\n\nconst client = new OpenAI({ baseURL: \"http://127.0.0.1:8080/v1\", apiKey: \"not-needed\" });\nconsole.log(await client.chat.completions.create({ model: \"trinity-nano-preview-4bit\", messages: [{ role: \"user\", content: \"Hello\" }] }));","python":"from openai import OpenAI\n\nclient = OpenAI(base_url=\"http://127.0.0.1:8080/v1\", api_key=\"not-needed\")\nprint(client.chat.completions.create(model=\"trinity-nano-preview-4bit\", messages=[{\"role\": \"user\", \"content\": \"Hello\"}]))","responses":"curl http://127.0.0.1:8080/v1/responses \\\n  -H \"Content-Type: application/json\" \\\n  -d '{\"model\":\"trinity-nano-preview-4bit\",\"input\":[{\"role\":\"user\",\"content\":\"Hello\"}]}'"}]},"variants":[{"backends":["cuda","metal","rocm"],"gated":false,"id":"default","model_id":"mlx-community/Trinity-Nano-Preview-4bit","precision":"q4"}]}
