NVIDIA/TensorRT-Incubator

Issues

• Implement eager mode caching

  #394 opened 23 days ago by pranavm-nvidia
  2

• Allow `DimensionSize`s in `InputInfo`

  #410 opened 9 days ago by pranavm-nvidia
  1

• Remove Parameter

  #416 opened 9 days ago by pranavm-nvidia
  1

• Profile eager mode

  #393 opened 23 days ago by pranavm-nvidia
  0

• Enable function registry for methods

  #338 opened 15 days ago by pranavm-nvidia
  6

• Stride validation fails for empty tensor input

  #398 opened 22 days ago by parthchadha
  0

• Fix or remove `skip_num_stack_entries` in `convert_to_tensors`

  #310 opened 16 days ago by pranavm-nvidia
  3

• Function Registry should do type checking on variadic arguments

  #341 opened 16 days ago by pranavm-nvidia
  1

• Improve efficiency of getting shapes of constants

  #360 opened 16 days ago by pranavm-nvidia
  1

• Improve behavior of evaluation during compile

  #409 opened 16 days ago by pranavm-nvidia
  1

• Enable data type constraints for modules

  #272 opened 2 months ago by pranavm-nvidia
  0

• Add few ops as methods of tensors

  #242 opened 2 months ago by akhilg-nv
  0

• Figure out how to convey the behavior of ops when receiving different Tensor subclasses

  #226 opened 16 days ago by pranavm-nvidia
  5

• Parameters in Module Containers Not Retained in state_dict after modification

  #386 opened 23 days ago by farazkh80
  1

• Merge dtype constraints and convert_to_tensors decorators

  #408 opened 16 days ago by pranavm-nvidia
  0

• Bug: mlir-tensorrt dynamic shapes

  #402 opened 17 days ago by parthchadha
  1

• Segfault in prompt_encoder

  #378 opened 17 days ago by yizhuoz004
  3

• Incorrect quantize output with fp16 scales

  #392 opened 23 days ago by parthchadha
  1

• `tripy.Module` does not support list of layers/ops of mixed types

  #302 opened 24 days ago by farazkh80
  1

• `stablehlo-to-tensorrt` conversion pass doesn't support `stablehlo.reduce` with multiple reduction dims

  #279 opened 2 months ago by farazkh80
  5

• 'stablehlo.dynamic_gather' op not a TensorRTEncodingOpInterface operation

  #381 opened a month ago by parthchadha
  1

• Dynamic shapes bug with gather op

  #326 opened a month ago by parthchadha
  5

• nanoGPT FP8 compilation failed

  #353 opened a month ago by yizhuoz004
  1

• infinite loop in TransposeEliminationPass with dynamic input

  #368 opened a month ago by zezhang
  2

• stablehlo.select : failed to determine expected shape

  #372 opened a month ago by parthchadha
  1

• segfault in SimplifyExtractOfReshape

  #377 opened a month ago by parthchadha
  1

• stablehlo.reduce_window dynamic shape issue

  #359 opened a month ago by yizhuoz004
  1

• Issue with the `tensorrt.slice` mlir canonicalizer

  #366 opened a month ago by matthewfl
  2

• [mlir-tensorrt] TensorRT Dialect `ReifyRankedShapedTypeOpInterface` implementation

  #369 opened a month ago by christopherbate
  0

• Sequential module support

  #295 opened a month ago by farazkh80
  2

• Make `preprocess_args` in `convert_to_tensors` work for variadic arguments

  #311 opened a month ago by pranavm-nvidia
  2

• Support collections of tensors as inputs to compiled functions

  #343 opened a month ago by pranavm-nvidia
  0

• dlpack non-linear stride

  #229 opened a month ago by parthchadha
  0

• Add strict flag to `model.load_state_dict`

  #296 opened a month ago by farazkh80
  1

• Parameters disappear from `state_dict` after `load_state_dict`

  #317 opened a month ago by farazkh80
  3

• Enable support for plugins with multiple inputs/outputs

  #342 opened a month ago by pranavm-nvidia
  0

• adaptive average pooling support with dynamic dimensions for `kernel_dims` and`stride`

  #335 opened a month ago by farazkh80
  1

• `stablehlo-to-tensorrt` does not support converting `stablehlo.dynamic_gather`

  #264 opened 2 months ago by parthchadha
  0

• [Bug] failed to derive upper bound

  #269 opened a month ago by yizhuoz004
  2

• LinspaceOp won't work properly on trt10.5

  #320 opened a month ago by zezhang
  4

• Disallow mixing Tensor/Shape types in operations

  #233 opened a month ago by pranavm-nvidia
  3

• Enable perf tests to work for forks

  #331 opened a month ago by pranavm-nvidia
  0

• Progress bar/indicator for `tp.compile`

  #294 opened 2 months ago by farazkh80
  0

• `tp.mean` failure when `dim` is multi dimensional with skipped dimensions

  #297 opened 2 months ago by farazkh80
  0

• Add support for multiple optimization profiles

  #315 opened a month ago by parthchadha
  0

• dtype constraints does not work for overloaded functions

  #300 opened 2 months ago by pranavm-nvidia
  0

• Use special fake tensor when tracing for compilation

  #250 opened 2 months ago by pranavm-nvidia
  1

• Compile time failure of trivial dynamic shape program

  #246 opened 2 months ago by parthchadha
  0

• Accuracy bug : avgpool with padding != 0

  #241 opened 2 months ago by parthchadha
  0

• reduce_window function invalid post clustering

  #237 opened 2 months ago by parthchadha
  1