}

.docs-navigation a:focus {
    outline: none;
}

.docs-navigation a:hover {
    color: #35c1e4;
    text-decoration: none;
}

.docs-navigation a code {
    color: #02303A;
    overflow-wrap: break-word;
    padding: 0;
    word-break: break-all;
}

.docs-navigation a.active {
    text-decoration: underline;
    outline: 0;
    border: none;
    outline-style: none;

// and NumericVerify ops to compare output values from the quantized and float
// ops.
//
// When `legacy_float_scale` is true, the quantizer will use float scale instead
// of double, and call TOCO's quantization routines to maintain bit-exactness of
// the values with the TOCO quantizer.
TfLiteStatus QuantizeModel(
    absl::string_view model_buffer, const tflite::TensorType &input_type,
    const tflite::TensorType &output_type,

// non-nil error is returned.
func NewWithRetry[T any](f func() (T, error)) Lazy[T] {
	return &lazyImpl[T]{getter: f, retry: true}
}

func (l *lazyImpl[T]) Get() (T, error) {
	if atomic.LoadUint32(&l.done) == 0 {
		// Outlined slow-path to allow inlining of the fast-path.
		return l.doSlow()
	}
	return l.res, l.err
}

func (l *lazyImpl[T]) doSlow() (T, error) {
	l.m.Lock()
	defer l.m.Unlock()
	if l.done == 0 {

// Test to verify loop outlining.

// RUN: tf-opt --split-input-file --tfl-while-loop-outline %s | FileCheck %s
// Check that while loop outlining is nop if re-ran.
// RUN: tf-opt --tfl-while-loop-outline %s -o %t1
// RUN: tf-opt --tfl-while-loop-outline %t1 -o %t2
// RUN: diff %t1 %t2

// CHECK-LABEL: func @while
func.func @while() -> tensor<1xf32>
    attributes {tf.entry_function = {outputs = "result"}} {

and you'll receive a more detailed response to your email within 72 hours indicating the next steps
in handling your report.

Please, provide a detailed explanation of the issue. In particular, outline the type of the security
issue (DoS, authentication bypass, information disclose, ...) and the assumptions you're making (e.g. do
you need access credentials for a successful exploit).

	// waiting for the first's call to f to complete.
	// This is why the slow path falls back to a mutex, and why
	// the o.done.Store must be delayed until after f returns.

	if o.done.Load() == 0 {
		// Outlined slow-path to allow inlining of the fast-path.
		o.doSlow(f)
	}
}

func (o *Once) doSlow(f func()) {
	o.m.Lock()
	defer o.m.Unlock()
	if o.done.Load() == 0 {
		defer o.done.Store(1)
		f()
	}

	// Worker routine calls Add before sending a batch and
	// then spawns a routine that calls Done after batch was processed by the delegate backend.
	// This WaitGroup is used to wait for all sending routines to finish before shutting down audit backend.
	wg sync.WaitGroup

	// Limits the number of batches sent to the delegate backend per second.
	throttle flowcontrol.RateLimiter
}

	}
	if goexperiment.CgoCheck2 {
		cgoCheckPtrWrite(ptr, new)
	}
	return atomic.Casp1(ptr, old, new)
}

// Like above, but implement in terms of sync/atomic's uintptr operations.
// We cannot just call the runtime routines, because the race detector expects
// to be able to intercept the sync/atomic forms but not the runtime forms.

//go:linkname sync_atomic_StoreUintptr sync/atomic.StoreUintptr

// Test to verify translation & export work as intended with runtime.

// RUN: tf-opt --mlir-print-debuginfo --canonicalize --tfl-while-loop-outline %s | mlir-tflite-runner --dump-interpreter-state 2>&1 | FileCheck %s

// Verify value computed:
// ----------------------
// CHECK: result: Tensor<type: FLOAT32, shape: 1, values: 96>
// CHECK: pconst: Tensor<type: INT32, shape: , values: 1>

// Verify tensors in interpreter state:

							<div class="alert alert-info">${msg}</div>
						</la:info>
						<la:errors />
					</div>
					<%-- Message: END --%>
					<div class="col-md-12">
						<div class="card card-outline card-primary">
							<div class="card-header">
								<h3 class="card-title">
									<la:message key="labels.upgrade_data_migration" />
								</h3>
							</div>
							<div class="card-body">