[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6},"actionData","errors","routes/docs",{"_7":8,"_9":10,"_11":12,"_13":14},"entry",{"_32":33,"_34":35,"_36":37,"_38":39,"_40":41,"_109":-7},"html","

Unreleased native NuTorch shell reference. Generated by\nscripts/gen-ops-reference.ts; every op also documents itself with\ntorch <op> --help. Tensor operands are native values, not strings.

manual_seed

seed the random number generator

Signature (placeholders describe arguments; this is not a runnable example):

use torch\ntorch manual_seed <seed>

","items",[15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31],"indexable",true,{"_32":105,"_34":106,"_36":107,"_38":108,"_40":78},{"_32":101,"_34":102,"_36":103,"_38":104,"_40":96},{"_32":97,"_34":98,"_36":99,"_38":100,"_40":96},{"_32":92,"_34":93,"_36":94,"_38":95,"_40":96},{"_32":88,"_34":89,"_36":90,"_38":91,"_40":87},{"_32":83,"_34":84,"_36":85,"_38":86,"_40":87},{"_32":79,"_34":80,"_36":81,"_38":82,"_40":78},{"_32":74,"_34":75,"_36":76,"_38":77,"_40":78},{"_32":70,"_34":71,"_36":72,"_38":73,"_40":41},{"_32":66,"_34":67,"_36":68,"_38":69,"_40":41},{"_32":62,"_34":63,"_36":64,"_38":65,"_40":41},{"_32":58,"_34":59,"_36":60,"_38":61,"_40":41},{"_32":54,"_34":55,"_36":56,"_38":57,"_40":41},{"_32":50,"_34":51,"_36":52,"_38":53,"_40":41},{"_32":46,"_34":47,"_36":48,"_38":49,"_40":41},{"_32":42,"_34":43,"_36":44,"_38":45,"_40":41},{"_32":33,"_34":35,"_36":37,"_38":39,"_40":41},"slug","reference/utility","title","Utility ops","description","The 1 utility operation, generated from the op table.","order",28,"section","Reference","reference/autograd","Autograd ops","The 4 autograd operations, generated from the op table.",27,"reference/loss","Loss ops","The 9 loss operations, generated from the op table.",26,"reference/shape","Shape ops","The 23 shape operations, generated from the op table.",25,"reference/linalg","Linear algebra ops","The 17 linalg operations, generated from the op table.",24,"reference/reduction","Reduction ops","The 21 reduction operations, generated from the op table.",23,"reference/comparison","Comparison ops","The 25 comparison operations, generated from the op table.",22,"reference/pointwise","Pointwise ops","The 71 pointwise operations, generated from the op table.",21,"reference/creation","Creation ops","The 14 creation operations, generated from the op table.",20,"install-from-source","Install from source","Build NuTorch from the Astrohacker checkout for local development.",8,"Start","nushell","The NuTorch shell","Launch NuTorch, import torch and work with native tensors in a full Nushell-based shell.",7,"neural-networks","Neural networks","Compose shared native modules, train with live optimizers and load compatible safetensors state files.",6,"Deep learning","autograd","Autograd","Differentiate native tensor computations while preserving shared leaves and independent gradient snapshots.",5,"ops","Operations","Use all 185 native tensor operations through shared command metadata and structured pipelines.",4,"Core","tensors","Tensors","Create native MPS tensors, share their ownership and explicitly convert their data.",3,"daemon","Native ownership","The NuTorch shell owns native tensors, modules and optimizers without a separate daemon.",2,"getting-started","Getting started","Install NuTorch with Homebrew and run native GPU tensor pipelines.",1,"content"]