Summary: 106 instances, 45 unique

Text	Count
// TODO update import path once op is modularized.	51
// TODO (yassogba) after 3.0 assess whether we need to keep this gated	1
// TODO call broadcastTo kernel directly once backends implement broadcstTo	1
// TODO: raise exception when a `.computeMask()` call	1
_RE_PATTERN_TODO = re.compile(r'^//(\s*)TODO(\(.+?\))?:?(\s|$)?')	1
const upstreamSummaries = 'TODO: fill in upstream of chained summaries';	1
const upstreamSummaries = 'TODO: fill in upstream of zip summaries';	1
// TODO: handle cases when x is complex.	1
// TODO: Use tf.copyRegisteredKernels once synced to tfjs-core 2.5.0.	1
// TODO: Make this into its own kernel.	1
// TODO: WebGPU doesn't support read data synchronously from GPU to CPU.	1
// TODO remove once fenceSync and clientWaitSync is implemented upstream.	1
// TODO validate that all these paths exist.	1
// TODO: Use CPU implementation due to the precision problem in Safari.	1
// TODO: move this logic to forward function when the broadcastTo op is	1
// TODO look into adding support for checking these extensions in expo-gl	1
// TODO: Ensure vec4 previous does not result in redundant sample,	1
// TODO File issue to be able get this from expo.	1
// TODO: Remove once this is fixed:	1
// TODO: Encoding should happen on GPU once we no longer have to download	1
// TODO: Once reduction ops are packed, batchMatMul will always be packed	1
* `coloredLabels[i]` (TODO: to be supported here).	1
// TODO: add more fields as needed.	5
// TODO: @webgpu/types 0.1.6 version has a bug to support both old	1
# Checks for common mistakes in TODO comments.	1
// TODO: Remove the limitaion of rank <= 4.	1
// TODO: WebGPU limits the maximum allowed shared memory size as 16K. To make	1
// TODO: Support slice ops for complex type.	1
// TODO: To see if we need to relax the limitation. Currently, it's only for	2
// TODO: Use matmul instead once it supports complex64 type.	1
// TODO - what about ID in this case? Handle in write()??	1
// TODO: consider changing this type. Possibly lift into top level state	1
// TODO: support 5d and 6d	2
// TODO: Remove the condition of !x.isUniform.	1
// TODO: This could end up being a redundant load with another one in	1
() => 'TODO: NCHW is unimplemented');	6
// TODO followup on removing this from the the interface	1
* TODO (cais,yassogba) consider removing GraphDef as GraphDefs now	1
// TODO: Currently we need to make sure that innerShapeA and outerShapeB	1
// TODO: Import kernel function once zeros is modularized.	1
// TODO: add support for int32 input.	1
// TODO: Implement CPU forwarding for complex inputs.	1
// TODO use type only imports for these types once we are on a version of	1
// TODO Use Engine.runKernel when CPU/WebGL/WASM backends implement this.	1
// TODO: handle cases when x is complex. Once the cpu implementation	1