*   Added a distributed runtime using GRPC
*   Move skflow to `contrib/learn`
*   Better linear optimizer in `contrib/linear_optimizer`
*   Random forest implementation in `contrib/tensor_forest`
*   CTC loss and decoders in `contrib/ctc`
*   Basic support for `half` data type
*   Better support for loading user ops (see examples in `contrib/`)
*   Allow use of (non-blocking) Eigen threadpool with

## 0.101.0

### Features

* ✨ Enable Pydantic's serialization mode for responses, add support for Pydantic's `computed_field`, better OpenAPI for response models, proper required attributes, better generated clients. PR [#10011](https://github.com/tiangolo/fastapi/pull/10011) by [@tiangolo](https://github.com/tiangolo).

### Refactors

	DPPS $7, X2, X11                        // 66440f3a40da07
	DPPS $7, X11, X11                       // 66450f3a40db07
	EMMS                                    // 0f77
	//TODO: ENTERQ $0x12, $0xf123           // c823f112
	EXTRACTPS $0, X2, (BX)                  // 660f3a171300
	EXTRACTPS $1, X11, (BX)                 // 66440f3a171b01
	EXTRACTPS $2, X2, (R11)                 // 66410f3a171302

carry // chains in the flags in parallel across the whole operation, and aggressively // unrolls loops. arm64 processes four words at a time. // // It's unclear why the assembly for all other architectures, as well as for // amd64 without ADX, perform better than the compiler output. // TODO(filippo): file cmd/compile performance issue. var supportADX = cpu.X86HasADX && cpu.X86HasBMI2 func init() { if cpu.AMD64 { impl.Register("aes", "ADX", &supportADX) } } //go:noescape func addMulVVW1024(z, x *uint,...

carry // chains in the flags in parallel across the whole operation, and aggressively // unrolls loops. arm64 processes four words at a time. // // It's unclear why the assembly for all other architectures, as well as for // amd64 without ADX, perform better than the compiler output. // TODO(filippo): file cmd/compile performance issue. var supportADX = cpu.X86HasADX && cpu.X86HasBMI2 func init() { if cpu.AMD64 { impl.Register("aes", "ADX", &supportADX) } } //go:noescape func addMulVVW1024(z, x *uint,...