robertknight rten issues

robertknight / rten

ONNX neural network inference engine

123 stars 9 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Add fast path for ArgMin / ArgMax when axis is contiguous

#411 robertknight closed 2 days ago
0
Optimize GatherND by avoiding redundant zeroing of output and adding fast path for contiguous inputs

#410 robertknight closed 2 days ago
0
Optimize copying of non-contiguous tensors with 5+ dimensions

#409 robertknight closed 2 days ago
0
Reserve KV cache capacity after the first model run

#408 robertknight closed 3 days ago
0
Support passing captures by-value to subgraphs

#407 robertknight closed 4 days ago
0
Prefer `Tensor::from` for creating vectors from array literals

#406 robertknight closed 5 days ago
0
Re-order execution plan to enable more operations to run in-place

#405 robertknight closed 5 days ago
0
Streamline NodeId -> usize conversions

#404 robertknight closed 1 week ago
0
Output transcription speed as a multiple of real-time in Whisper example

#403 robertknight closed 1 week ago
0
Support (de-)serialization of tensors using serde

#402 robertknight closed 1 week ago
0
Reuse existing output node when replacing operator with a fusion

#401 robertknight closed 1 week ago
0
Normalize and chunk audio inputs in wav2vec2 example

#400 robertknight closed 1 week ago
0
Support in-place `Concat` when input is a captured value

#399 robertknight closed 3 days ago
3
Support fusing Transpose + MatMul where both inputs are transposed

#398 robertknight opened 2 weeks ago
0
Add Whisper speech recognition example

#397 robertknight closed 3 weeks ago
0
Add context to token generation errors

#396 robertknight closed 3 weeks ago
0
Support `cache_position` inputs in Hugging Face models

#395 robertknight closed 3 weeks ago
0
Add `token_id_filter` function to create logits filter from predicate

#394 robertknight closed 3 weeks ago
0
Add API to rten-generate for filtering/processing logits

#393 robertknight closed 3 weeks ago
0
Support newer `merges` format in tokenizer.json files

#392 robertknight closed 3 weeks ago
0
Parsing of `merges` field in tokenizer.json fails

#391 Tracreed closed 3 weeks ago
1
Enclose linked identifiers in backticks in API docs

#390 robertknight closed 1 month ago
0
WASM relaxed SIMD support

#389 robertknight opened 1 month ago
1
Optimize Cast op when source and dest types are the same, add u8/i8 cast tests

#388 robertknight closed 1 month ago
0
int8 and uint8 support in Pad and other ops

#387 igor-yusupov closed 1 month ago
2
Generalize `IntoOpResult` impls

#386 robertknight closed 1 month ago
0
Use a macro to de-duplicate proxy layout impls

#385 robertknight closed 1 month ago
0
Replace remaining uses of `check_dims` with `static_dims`

#384 robertknight closed 1 month ago
0
Generalize `Output` => `Tensor`/`TensorView` convenience methods

#383 robertknight closed 1 month ago
0
VALID auto_pad value

#382 igor-yusupov opened 1 month ago
8
Convert `NodeId` from an alias for `usize` to u32-sized opaque type

#381 robertknight closed 1 month ago
0
Use "dep:" prefix to prevent implicit creation of features

#380 robertknight closed 1 month ago
0
Support setting `RunOptions` for model inference when using `Generator`

#379 robertknight closed 1 month ago
0
Remove self-dependency of crates in development

#378 robertknight closed 1 month ago
0
Improve debug formatting of tensors

#377 robertknight closed 1 month ago
0
Fix AVX-512 build of im2col.rs

#376 robertknight closed 1 month ago
0
Use a Python virtualenv in CI

#375 robertknight closed 1 month ago
0
Update YOLO example instructions to use YOLO v11

#374 robertknight closed 1 month ago
0
Use `f32::round_ties_even` instead of `libm::rintf`

#373 robertknight closed 1 month ago
0
Widen float16 constants to float32 in model converter

#372 robertknight closed 1 month ago
0
Fuse pointwise operations into matmul / convolution operations

#371 robertknight opened 1 month ago
0
Implement better depthwise convolution kernels

#370 robertknight opened 1 month ago
0
Apply fusions correctly when fused ops are chained

#369 robertknight closed 1 month ago
0
Add `--no-optimize` flag to rten CLI

#368 robertknight closed 1 month ago
0
Replace old tensor slicing methods with new API

#367 robertknight closed 1 month ago
0
Infer layout for `slice_copy` based on layout and range

#366 robertknight closed 1 month ago
0
Add fallible variants of `TensorBase::slice_with`

#365 robertknight closed 1 month ago
0
Use `zip` method instead of standalone function

#364 robertknight closed 1 month ago
0
Support ranges with unknown index count in `slice_with`

#363 robertknight closed 2 months ago
0
Replace most remaining uses of `slice` with `slice_with`

#362 robertknight closed 2 months ago
0