This is an automated email from the ASF dual-hosted git repository.
tqchen pushed a change to branch unity
in repository https://gitbox.apache.org/repos/asf/tvm.git
omit a425bc7a39 [Unity] Pattern-based rewriting for dataflow block (#14446)
omit 646d50dc27 [Unity][Graph matching] Clean up undo stack for parent and
child nodes properly (#14440)
omit cecc5c3ade [Unity][Op][Docs] Update comment for `call_tir_dyn` (#14441)
omit 41230a981f [Unity][Graph matching] Automatically add `used-by`
constraints for `is_op` pattern (#14439)
omit dc7ba6c46c [Unity] Remove non-deterministic behavior from graph
pattern matching (#14417)
omit 9efc5b83a7 [Unity] Minor updates to DataFlowBlockRewrite (#14431)
omit 784733a425 [Unity][Fix] Annotate TIR op pattern could have no stores.
(#14420)
omit ac90c7af01 [Unity] Include constant shapes in the profiler result
(#14428)
omit ab3299c054 [Unity] Handle extern func calls in static memory planning
(#14419)
omit c5335d96f9 [Unity][Fix] Copy over module attrs in FuseTIR (#14418)
omit d93eb5c091 [Unity][Hexagon] Enable Relax VM for Hexagon (#14415)
omit 30db3de0e7 [Unity][Op] Expose scale in `R.nn.attention` and add its
legalize op (#14412)
omit ef4057a433 [Unity] Fix getting shapes for cutlass BYOC kernels (#14411)
omit c69c75407f [Unity][Op] Conv1d (#14388)
omit f45f11a9e5 [Unity][QNN][Hexagon]Support Relax Constants in the QNN
TOPI operations (#14386)
omit 219ed08e12 [Unity][Transform] Common Subexpression Elimination (#14361)
omit 25608f40c6 [Unity][TVMScript] Fix Shape Var occurrence in Tensor
annotation (#14404)
omit 414514c1bf [Unity][Op] Add stop_lift_params (#14368)
omit f6919620c1 [Unity] Support simple dynamic-shape-aware fusion (#14396)
omit 95b6f680b7 [Unity][Transform] SplitCallTIRByPattern and CUTLASS
backend (#14274)
omit 634cfad0dc [Unity] Add missing #include <array> (#14383)
omit 5a2f1ba2c6 [Unity][VM] Add CUDA graph vm builtins (#14371)
omit 0908a43466 [Unity] Also include output dtype in simt MathInstruction
(#14372)
omit e32164a805 [Unity][Fix] Allow scalar layout initialization (#14370)
omit b5b8e206d6 [Unity][TVMScript] Update GlobalVar `checked_type_` when
`emit_te` (#14367)
omit 5afb3ea5c5 [Unity] Add More Ops For FX Translator (#14348)
omit bc391d3429 [Unity][Fix] Infer Layout must support negative axes
(#14365)
omit 77496c33f3 [Unity][Pass] Fix FuseOps error if there is no output of a
given group (#14354)
omit f38171b0cd [Unity][WEB] Support async pipeline creation (#14362)
omit 84dc90d76b [Unity] Add support to append relay op attrs in translator
(#14356)
omit fc8bbbd6b4 [Unity][Transform] Fix AMP tests (#14360)
omit 3d7af30df7 [Unity][Transform] Introduce data-dependent operation of
reshape and its constant folding (#14282)
omit 9b8e003d50 [Unity][Fix] Fix block memory plan to handle bool (#14357)
omit d108639bce [Unity][Transform] AMP out_dtype=float16 testcases (#14358)
omit a5d659099d [Unity][BYOC] Check leaked intermediate variables in
cutlass patterns (#14350)
omit d623140045 [Unity] Support model kwargs in dynamo_capture_subgraph
(#14349)
omit 602fd10694 [Unity][Frontend] FX exp and strided_slice fix (#14338)
omit 029a5e8793 [Unity][BYOC] Update testcases to follow recent changes
(#14339)
omit de8c12ab3c [Unity] Remove Python interface of RemoveUnusedFunction
(#14336)
omit ccb9074907 [Unity][Pass] Reuse prior infra to implement more complete
DCE (#14334)
omit dd742a826a [Unity][Op] Fix Strided Slice Shape Inference (#14324)
omit aa1932492b [Unity][Transform] DefaultSchedule pass (#14266)
omit 24e0fc7c69 [Unity][Lint] Fix cpplint casting (#14333)
omit 3497cca0b5 [Unity][Transform] Automatic Mixed Precision (#14242)
omit 3b731b2eee [Unity][Transform] Simple Dead Code Elimination (#14262)
omit cdb435ccff [Unity][Transform] Automatic Layout Conversion (#14257)
omit 30817d1aef [Unity][TOPI] fp16 LayerNorm & GroupNorm (#14264)
omit 3e66b205d2 [Unity][Contrib] Introduce several features of cutlass
profiler (#14275)
omit 9cba9bfd7a [Unity][Transform] Enhance RewriteDataflowReshape transform
(#14265)
omit 0145fe97a4 [Unity][BYOC] Improve expressiveness of the pattern check
function in FuseOpsByPattern (#14310)
omit db7fdfd5fa [Unity][BYOC] Support matmul + residual block fusion in
CUTLASS BYOC (#14317)
omit 1a7244135f [Unity] Support pattern-based rewriting (#14312)
omit e61576ba4b [Unity][Web] WebGPU explicit max buffer size (#14321)
omit 0f6463fccb [Unity][Op] Enable special dimension value 0 in reshape
(#14311)
omit a9ca0cf0ab [Unity][Pass] Add a pass to alter the TIR implementation of
an operator (#14215)
omit 1a582b9d79 [Unity][DEBUG] Add Instrument (#14302)
omit 0f49776de3 [Unity][Op] Cumsum (#14297)
omit d394b6a89f [Unity] Fix StructInfo Infer for `vm.alloc_tensor` (#14283)
omit df7f510da8 [Unity] Mark tests that need python3.8 compact.
omit 2ce4af3e0c [TVMScript][Unity] Improve PyLint Compatibility (#14276)
omit 97b429a256 [Unity][ci] Use CPU-SMALL instances (#14256)
omit d268b13cac [Unity] Introduce call_dps_packed (#14183)
omit 96cd5b5b4e [Unity] Consider target context for Relay to Relax
conversion (#14269)
omit 08e2a69efc [Unity][Frontend] Import `tanh` and fix `layer_norm`
(#14247)
omit 71899e5529 [Unity][BYOC] Add conv2d and residual block patterns for
Relax cutlass BYOC (#14252)
omit 77695deec6 [Unity] Allow user defined func attrs in emit_te (#14255)
omit 3cb9e263b9 [Unity][Op] Add repeat, tile, conv2d_transpose, avg_pool2d
(#14238)
omit ac82cf8b0c [Unity][Op][Tweak] Improve `StructInfo` inference for
`shape_of` (#14243)
omit 3bddee1524 [Unity][WEB] Improve ndarray cache (#14236)
omit 556b542611 [Unity][WEB] Update text prompts for syntactical
correctness (#14237)
omit fba4b6bc50 [Unity][TVMScript] Fix prim_func lost issue in
relax.emit_te (#14189)
omit 2a32d64ef1 [Unity][TVMScript] Enable Context-Aware Parsing (#14234)
omit 6ca3325a73 [Unity][Bugfix] Do not include `PrimFunc`s in the
dependency graph when checking for recursion (#14228)
omit 80fce8db81 [Unity][Transform] SimplifyNormInference (#14221)
omit 544b0821ae [Unity] Improve implementation of FuseOps (#14229)
omit a6d9601595 [Unity] ensure memory.alloc_tensor/storage roundtrippable
(#14226)
omit a3f40a7635 [Unity][WEB] Simplify WebGPU Codegen per spec (#14225)
omit 4c39c31767 [Unity][Transform] Memory plan across the IRModule (#14220)
omit 6de29c50a2 [Unity][BYOC] Add dynamic shape support to CUTLASS matmul
(#14216)
omit 7a4bdcde3c [Unity][Frontend] from_fx keeps parameters in order (#14214)
omit 45a54f3a38 [Unity][WEB] Improve webgpu codegen options to skip
readonly (#14213)
omit 58e224f8b1 [Unity][Frontend] FX translator supports unwrapping unit
return tuple (#14212)
omit 4920cd26df [Unity][Frontend] Attach imported model weights, deprecate
ImporterOutput (#14211)
omit f7ccc3bc59 [Unity] Introduce Default GPU Schedule Pass (#14182)
omit 03e413ae43 [Unity][Frontend] FX translator support torch.baddbmm
(#14202)
omit 1978e44971 [Unity][TIR][Pass] ForceNarrowIndexToInt32 (#14203)
omit 2c75602cb4 [Unity][Fix] FX translating dtype (#14201)
omit 1896823417 [Unity][Frontend] FX translator returning weights with
`keep_params_as_input` (#14197)
omit 5bafde482d [Unity][Frontend] FX translator supporting more ops (#14196)
omit 012dacec71 [Unity][Op] Legalize `round`, `floor`, `ceil`, `sign`
(#14198)
omit 694da73413 [Unity][Op] Argmax and argmin (#14195)
omit 32049d825b [Unity][Op] Group normalization (#14194)
omit d68bfb97ee [Unity][Transform] LiftTransformParams handling multiple
functions (#14192)
omit 9ade1be9f7 [Unity][WEBGPU] Codegen improvements and WebRuntime (#14187)
omit 6c3a97c71c [Unity][OP] Add an operator for fused multi head attention
(#14150)
omit 031e380c47 [Unity][Analysis] Restore Python bindings for var analyses
(#14180)
omit fb3e269c71 [Unity][Op] Full support of Relax op `power` (#14171)
omit d50be1cdf6 [Unity][BYOC] Add batch matmul support to Relax CUTLASS
BYOC (#14166)
omit 5f4a11a284 [Unity][Analysis] Analysis for detecting recursion in Relax
(#14149)
omit 70e925c8de [Unity] Add bind_constants option to FuseOpsByPattern
(#14151)
omit 96d85b2da5 [Unity][BYOC] Use Relax legalize + CPU build for reference
in tests (#14162)
omit 1d60a6a337 [Unity][Analysis] Checking function return struct info in
well-formed check (#14155)
omit 78af3acde3 [Unity][Pass] Support Symbolic Shape Deduction during
BindParam (#14154)
omit 832c1ba04c [Unity][Debugging] AST printer (#14152)
omit 016b2800a1 [Unity][Pass] Enhance constant folding to fold relax ops by
evaluating them. (#14146)
omit 993c37d3c2 [Unity][Legalize] Fix Scalar Constant Legalization (#14127)
omit 4892b763b9 [Unity] Add callback to FuseOpsByPattern to check match
result is accepted (#14109)
omit 5f5638c05a [Unity][BYOC] Assign group to unused bindings and ignroe
PrimFunc (#14139)
omit a5fbbd573f [Unity][TVMScript] emit_te sugar (#14123)
omit 8c1d87a46c [Unity][BYOC] Add transposed matmul support to Relax
CUTLASS BYOC (#14128)
omit e85a1909db [Unity] Add Global info (#14132)
omit 17d8625a73 [Unity][WEB] Relax vm on web runtime (#14131)
omit 631e483330 [Unity][BlockBuilder] Add `name_hint` argument for `emit`
and `emit_output` (#14126)
omit 81a6438bc7 [Unity][Fix] Fix bug in MergeCompositeFunctions (#14117)
omit de2e70778e [Unity] Update tests again to adapt to latest TVMScript
syntax (#14115)
omit c973eae56c [Unity][BYOC]Add relax backend pattern registry (#14106)
omit 7ac87251d0 [Unity] Remove attributes of relax.print, assert and unique
(#14101)
omit dd00671ae3 [Unity][Layout] Add layout transformation analysis for
PrimFunc (#14066)
omit 35331cdea2 [Unity] Relax Recursive function (#14092)
omit 1ea40509c9 [Unity] Lower `shape_of` to a builtin (#14093)
omit 3e139b0a93 [Unity] Fix typo in the comment (#14096)
omit 111dd1f6f5 [Unity][Relax] Set Shape Function to Be Host Function
(#14090)
omit c728978f51 [Unity] Refactor Relax Build JIT UX (#14088)
omit fa0f49a6a7 [Unity][Fix][Pass] FoldConstant with DCE in dataflow block
(#14087)
omit b5e6048361 [Unity][Analysis] TIR pattern kind analysis for
multi-buffer write block (#14075)
omit cb7e29f7de [Unity][Op] `log_softmax` and `cross_entropy_with_logits`
(#14083)
omit 394f1261a5 [Unity][BYOC] Add DNNL backend (#14082)
omit 1774d2229c [Unity][BYOC] Add CUTLASS backend (#14081)
omit 418eaf0b6b [Unity] Add testcases for `expr_args_converter` (#14080)
omit abdfe98d85 [Unity][Pass] Canonicalize Bindings (#14079)
omit 183e4e1d84 [Unity][BYOC][Pass] RunCodegen and TensorRT (#14078)
omit ac49e71881 [Unity][Transform] Add LiftTransformParams pass (#14069)
omit 575fee9bb3 [Unity][Frontend] Annotate number of non-static input of FX
function (#14067)
omit e6fdfc6075 [Unity][BYOC] Add pass to merge composite functions to
offload large subgraphs (#14062)
omit 5f15d3a5fb [Unity][Pass] Remove Unused Function (#14061)
omit daa3184b29 [Unity][Fix][Pass] Fix FuseOps for lack graph edges (#14058)
omit 3097f6648f [Unity] Relax op: collapse sum (#14059)
omit 9b1948d0ba [Unity][BYOC] Add pattern-based partitioning pass (#14054)
omit b23e18c228 [Unity][VM] Add per-op profiling support (#14053)
omit 8bad813c99 [Unity][TVMScript] Overload `__neg__` for relax expr
(#14045)
omit 80c474fbf1 [Unity][Pass] FuseOps FuseTIR fixes (#14044)
omit 969047780a [Unity] Statement rewriter for DataflowBlock (#14043)
omit 0525e05aaf [Unity] Relax dataflow pattern language (matching) (#14041)
omit c45b1a6990 [Unity] Update tests to adapt to latest TVMScript syntax
(#14039)
omit db1bf6b039 [Unity] Disallow inline prim_func in relax IR (#14040)
omit 9879fbbd0b [Unity][Pass] Block-level static memory planning (#14038)
omit 8d575f2a73 [Unity] Initial PyTorch Frontend (#14037)
omit 317634bc19 [Unity][Op] Add ShapeExpr Tests for Reshape Op (#14035)
omit db588383bf [Unity][Pass] Operator legalization (#14029)
omit bd8fb78ac4 [Unity][TVMScript] Move tir/relax import in script out of
__init__.py (#14033)
omit a50cdd06e3 [Unity][Pass] Wellformed Analysis (#14032)
omit 2cc122cd24 [Unity][BlockBuilder] CallTE convert PrimValue args
(#14028)
omit d8a6d1d826 [Unity][Pass] Normalize Pass (#14031)
omit 44b636f9be [Unity] Relay -> Relax translator (#14026)
omit 792d7c5eda [Unity][Pass][TuningAPI] Introduce TuningAPI and
MetaSchedule pass (#14014)
omit f81e198ed4 [Unity][Pass] BindParams pass, FoldConstant pass (#14016)
omit 5a6579e1b0 [Unity][VM] Supporting "compiled" exec mode. (#14015)
omit 7de9c82626 [Unity][Pass] LambdaLift pass (#14012)
omit 18ade5f8ba [Unity][Pass] Operator Fusion Passes (#14001)
omit 26b4439cf1 [Unity] NestedMsg Support utility (#13995)
omit f64e91c6da [Unity] Relax op: manipulation (#13989)
omit 044f3bbc41 [Unity] Relax op: search (#13992)
omit bf6e2a9ef6 [Unity] Relax op: linear algebra (#13988)
omit 444d420450 [Unity] Relax op: creation (#13984)
omit 5b3239ad4d [Unity] Relax op: neural networks (#13993)
omit ec110c6023 [Unity] Relax op: statistical (#13991)
omit c534c9c7b3 [Unity] Relax op: arithmetic, comparison (#13983)
omit a9a561b472 [Unity] Relax op: image (#13994)
omit 9a9e4a7823 [Unity] Relax op: set (#13990)
omit 4dfa36202b [Unity] Relax op: datatype (#13986)
omit caddedb418 [Unity] Relax op: index (#13987)
omit 4051a69cec [Unity][TVMScript] Use explicit `R.shape` in TVMScript
(#13979)
omit dbedbb25ba [Unity] e2e Relax minimum build flow (#13961)
omit afe71010ef [Unity] Relax VM shape lowering pass (#13956)
omit 7f1e1f5528 [Unity] Relax VM codegen (#13954)
omit a2d032494f [Unity] Relax TVMScript Printer (#13944)
omit 63de0dacbd [Unity] Relax TVMScript Parser. (#13932)
omit 23a7cd1a21 [Unity] Relax BlockBuilder and ExprMutator (#13926)
omit a7086616d7 [Unity] Basic StructInfo Analysis and Expr construction
(#13916)
omit f6b68ab7fd [Unity][CI] Unity specific jenkins setup (do not upstream
to main) (#13910)
omit 2bb2e4bf75 [Unity][IR] First-class StructInfo (#13907)
omit 0117a28d22 [Unity] Relax expressions and types (#13901)
omit a27451755f [Unity] Relax VM (#13878)
add 70d80cc126 [Unity] Relax VM (#13878)
add 25e6174770 [Unity] Relax expressions and types (#13901)
add 51cf110c0c [Unity][IR] First-class StructInfo (#13907)
add 40c43108c2 [Unity][CI] Unity specific jenkins setup (do not upstream
to main) (#13910)
add e68ef58c84 [Unity] Basic StructInfo Analysis and Expr construction
(#13916)
add aa59d86ac7 [Unity] Relax BlockBuilder and ExprMutator (#13926)
add 540ba28f5c [Unity] Relax TVMScript Parser. (#13932)
add 334fd8be51 [Unity] Relax TVMScript Printer (#13944)
add 574084d340 [Unity] Relax VM codegen (#13954)
add 518a7bdd96 [Unity] Relax VM shape lowering pass (#13956)
add b3d913a737 [Unity] e2e Relax minimum build flow (#13961)
add caad7cdf48 [Unity][TVMScript] Use explicit `R.shape` in TVMScript
(#13979)
add 903628f106 [Unity] Relax op: index (#13987)
add 0319939e47 [Unity] Relax op: datatype (#13986)
add 11e5e1d6d4 [Unity] Relax op: set (#13990)
add 128042204c [Unity] Relax op: image (#13994)
add 64f281749a [Unity] Relax op: arithmetic, comparison (#13983)
add be4373a1d2 [Unity] Relax op: statistical (#13991)
add fa3cfbd60e [Unity] Relax op: neural networks (#13993)
add 45e6d0bade [Unity] Relax op: creation (#13984)
add 206762f0e6 [Unity] Relax op: linear algebra (#13988)
add 088afc5d54 [Unity] Relax op: search (#13992)
add 4430d26394 [Unity] Relax op: manipulation (#13989)
add 471ff84280 [Unity] NestedMsg Support utility (#13995)
add 604754a6c8 [Unity][Pass] Operator Fusion Passes (#14001)
add 50f616c95c [Unity][Pass] LambdaLift pass (#14012)
add 0a0166e81f [Unity][VM] Supporting "compiled" exec mode. (#14015)
add 660a1f361c [Unity][Pass] BindParams pass, FoldConstant pass (#14016)
add 30a5669b8a [Unity][Pass][TuningAPI] Introduce TuningAPI and
MetaSchedule pass (#14014)
add b792751aaa [Unity] Relay -> Relax translator (#14026)
add 554e609aca [Unity][Pass] Normalize Pass (#14031)
add 477d011093 [Unity][BlockBuilder] CallTE convert PrimValue args
(#14028)
add f61c330ec6 [Unity][Pass] Wellformed Analysis (#14032)
add 4df587aba6 [Unity][TVMScript] Move tir/relax import in script out of
__init__.py (#14033)
add d2b8ea330d [Unity][Pass] Operator legalization (#14029)
add e0c6ca0383 [Unity][Op] Add ShapeExpr Tests for Reshape Op (#14035)
add e15f59bcf0 [Unity] Initial PyTorch Frontend (#14037)
add e8ebd68e92 [Unity][Pass] Block-level static memory planning (#14038)
add 2e1d85149b [Unity] Disallow inline prim_func in relax IR (#14040)
add 60d3796e20 [Unity] Update tests to adapt to latest TVMScript syntax
(#14039)
add 7d70af3723 [Unity] Relax dataflow pattern language (matching) (#14041)
add 44b05d68bb [Unity] Statement rewriter for DataflowBlock (#14043)
add 7a71e53b69 [Unity][Pass] FuseOps FuseTIR fixes (#14044)
add e7cc2876ec [Unity][TVMScript] Overload `__neg__` for relax expr
(#14045)
add 5239022d97 [Unity][VM] Add per-op profiling support (#14053)
add 38178a3298 [Unity][BYOC] Add pattern-based partitioning pass (#14054)
add a7e936fca1 [Unity] Relax op: collapse sum (#14059)
add 7ce0a519e7 [Unity][Fix][Pass] Fix FuseOps for lack graph edges (#14058)
add 34d5944135 [Unity][Pass] Remove Unused Function (#14061)
add 2c5ccfa7a6 [Unity][BYOC] Add pass to merge composite functions to
offload large subgraphs (#14062)
add 7f5f816084 [Unity][Frontend] Annotate number of non-static input of FX
function (#14067)
add 73fe88d04a [Unity][Transform] Add LiftTransformParams pass (#14069)
add b137d22ed4 [Unity][BYOC][Pass] RunCodegen and TensorRT (#14078)
add 38f920d480 [Unity][Pass] Canonicalize Bindings (#14079)
add 2a51dcae46 [Unity] Add testcases for `expr_args_converter` (#14080)
add 026217d998 [Unity][BYOC] Add CUTLASS backend (#14081)
add 4365d1e29d [Unity][BYOC] Add DNNL backend (#14082)
add 2409be1934 [Unity][Op] `log_softmax` and `cross_entropy_with_logits`
(#14083)
add 9caab4d9d8 [Unity][Analysis] TIR pattern kind analysis for
multi-buffer write block (#14075)
add 872c9f23c6 [Unity][Fix][Pass] FoldConstant with DCE in dataflow block
(#14087)
add 6b87e35617 [Unity] Refactor Relax Build JIT UX (#14088)
add 3959707f19 [Unity][Relax] Set Shape Function to Be Host Function
(#14090)
add 79fe0a22aa [Unity] Fix typo in the comment (#14096)
add 779c54d9ad [Unity] Lower `shape_of` to a builtin (#14093)
add 3173f2855d [Unity] Relax Recursive function (#14092)
add 62a4c735fb [Unity][Layout] Add layout transformation analysis for
PrimFunc (#14066)
add 1bf8b63648 [Unity] Remove attributes of relax.print, assert and unique
(#14101)
add 30d204bb4a [Unity][BYOC]Add relax backend pattern registry (#14106)
add a1f8d72178 [Unity] Update tests again to adapt to latest TVMScript
syntax (#14115)
add 6a9caf40e7 [Unity][Fix] Fix bug in MergeCompositeFunctions (#14117)
add 3515f95ba0 [Unity][BlockBuilder] Add `name_hint` argument for `emit`
and `emit_output` (#14126)
add c1f304497e [Unity][WEB] Relax vm on web runtime (#14131)
add aaa457d304 [Unity] Add Global info (#14132)
add 837dd6c228 [Unity][BYOC] Add transposed matmul support to Relax
CUTLASS BYOC (#14128)
add a4b1ec097c [Unity][TVMScript] emit_te sugar (#14123)
add be170bb3d2 [Unity][BYOC] Assign group to unused bindings and ignroe
PrimFunc (#14139)
add 77f4c719f3 [Unity] Add callback to FuseOpsByPattern to check match
result is accepted (#14109)
add 1b6d104f68 [Unity][Legalize] Fix Scalar Constant Legalization (#14127)
add 788c12f84f [Unity][Pass] Enhance constant folding to fold relax ops by
evaluating them. (#14146)
add f17a7285bb [Unity][Debugging] AST printer (#14152)
add c78e68bf4d [Unity][Pass] Support Symbolic Shape Deduction during
BindParam (#14154)
add ab5aaa0d1b [Unity][Analysis] Checking function return struct info in
well-formed check (#14155)
add 89cff055d3 [Unity][BYOC] Use Relax legalize + CPU build for reference
in tests (#14162)
add 2ca340a30b [Unity] Add bind_constants option to FuseOpsByPattern
(#14151)
add 4fe3a6d1f4 [Unity][Analysis] Analysis for detecting recursion in Relax
(#14149)
add 2de852f203 [Unity][BYOC] Add batch matmul support to Relax CUTLASS
BYOC (#14166)
add 8de0ff78f5 [Unity][Op] Full support of Relax op `power` (#14171)
add 9797924d54 [Unity][Analysis] Restore Python bindings for var analyses
(#14180)
add 30dc2365e7 [Unity][OP] Add an operator for fused multi head attention
(#14150)
add 279317d4cd [Unity][WEBGPU] Codegen improvements and WebRuntime (#14187)
add 3f66edc8ed [Unity][Transform] LiftTransformParams handling multiple
functions (#14192)
add 283a3db675 [Unity][Op] Group normalization (#14194)
add c25659d88f [Unity][Op] Argmax and argmin (#14195)
add 2f114d28b3 [Unity][Op] Legalize `round`, `floor`, `ceil`, `sign`
(#14198)
add 012923572a [Unity][Frontend] FX translator supporting more ops (#14196)
add a115d03e76 [Unity][Frontend] FX translator returning weights with
`keep_params_as_input` (#14197)
add 27b6fa7cb4 [Unity][Fix] FX translating dtype (#14201)
add 6ab0abc08d [Unity][TIR][Pass] ForceNarrowIndexToInt32 (#14203)
add 3f4417c6db [Unity][Frontend] FX translator support torch.baddbmm
(#14202)
add 2d562f8926 [Unity] Introduce Default GPU Schedule Pass (#14182)
add 97dc05ee18 [Unity][Frontend] Attach imported model weights, deprecate
ImporterOutput (#14211)
add 033fc0e34b [Unity][Frontend] FX translator supports unwrapping unit
return tuple (#14212)
add 9f9cf2967c [Unity][WEB] Improve webgpu codegen options to skip
readonly (#14213)
add 8f2153cfb4 [Unity][Frontend] from_fx keeps parameters in order (#14214)
add 30848742cb [Unity][BYOC] Add dynamic shape support to CUTLASS matmul
(#14216)
add 73b03dadb9 [Unity][Transform] Memory plan across the IRModule (#14220)
add 780a579c72 [Unity][WEB] Simplify WebGPU Codegen per spec (#14225)
add 3345aeda4a [Unity] ensure memory.alloc_tensor/storage roundtrippable
(#14226)
add 30a9403729 [Unity] Improve implementation of FuseOps (#14229)
add 9bc1ecc672 [Unity][Transform] SimplifyNormInference (#14221)
add a6b37b80ce [Unity][Bugfix] Do not include `PrimFunc`s in the
dependency graph when checking for recursion (#14228)
add 3e03ca5abe [Unity][TVMScript] Enable Context-Aware Parsing (#14234)
add cea0161bcb [Unity][TVMScript] Fix prim_func lost issue in
relax.emit_te (#14189)
add 14a7de57b3 [Unity][WEB] Update text prompts for syntactical
correctness (#14237)
add 663c8211a4 [Unity][WEB] Improve ndarray cache (#14236)
add c7c4dee36b [Unity][Op][Tweak] Improve `StructInfo` inference for
`shape_of` (#14243)
add 633c9002e3 [Unity][Op] Add repeat, tile, conv2d_transpose, avg_pool2d
(#14238)
add 07f7fc0a9c [Unity] Allow user defined func attrs in emit_te (#14255)
add 04d6d15984 [Unity][BYOC] Add conv2d and residual block patterns for
Relax cutlass BYOC (#14252)
add df9d084553 [Unity][Frontend] Import `tanh` and fix `layer_norm`
(#14247)
add 71c23b7a11 [Unity] Consider target context for Relay to Relax
conversion (#14269)
add 70386ecc91 [Unity] Introduce call_dps_packed (#14183)
add c676d1392f [Unity][ci] Use CPU-SMALL instances (#14256)
add 121e1e7a03 [TVMScript][Unity] Improve PyLint Compatibility (#14276)
add 6be5c19c9c [Unity] Mark tests that need python3.8 compact.
add 68827fa1d6 [Unity] Fix StructInfo Infer for `vm.alloc_tensor` (#14283)
add e3f975543f [Unity][Op] Cumsum (#14297)
add 67e183c33f [Unity][DEBUG] Add Instrument (#14302)
add f96e09a9ce [Unity][Pass] Add a pass to alter the TIR implementation of
an operator (#14215)
add a46fd90478 [Unity][Op] Enable special dimension value 0 in reshape
(#14311)
add e0c4456f50 [Unity][Web] WebGPU explicit max buffer size (#14321)
add 4bfbd3a703 [Unity] Support pattern-based rewriting (#14312)
add 2f1642f128 [Unity][BYOC] Support matmul + residual block fusion in
CUTLASS BYOC (#14317)
add 778b343ed9 [Unity][BYOC] Improve expressiveness of the pattern check
function in FuseOpsByPattern (#14310)
add 7c5a570dc1 [Unity][Transform] Enhance RewriteDataflowReshape transform
(#14265)
add b10e2a2a31 [Unity][Contrib] Introduce several features of cutlass
profiler (#14275)
add f67657fe09 [Unity][TOPI] fp16 LayerNorm & GroupNorm (#14264)
add 5cd30b00ef [Unity][Transform] Automatic Layout Conversion (#14257)
add 59746ea9b0 [Unity][Transform] Simple Dead Code Elimination (#14262)
add 371c3a2f2f [Unity][Transform] Automatic Mixed Precision (#14242)
add 691983096c [Unity][Lint] Fix cpplint casting (#14333)
add 5229827888 [Unity][Transform] DefaultSchedule pass (#14266)
add 375a2921ea [Unity][Op] Fix Strided Slice Shape Inference (#14324)
add 920f98d8f9 [Unity][Pass] Reuse prior infra to implement more complete
DCE (#14334)
add f4a451d273 [Unity] Remove Python interface of RemoveUnusedFunction
(#14336)
add 050562bfe9 [Unity][BYOC] Update testcases to follow recent changes
(#14339)
add 7534531a6e [Unity][Frontend] FX exp and strided_slice fix (#14338)
add 021ce24d0c [Unity] Support model kwargs in dynamo_capture_subgraph
(#14349)
add fe34a55bee [Unity][BYOC] Check leaked intermediate variables in
cutlass patterns (#14350)
add 9ae5895aae [Unity][Transform] AMP out_dtype=float16 testcases (#14358)
add cd48ba9e33 [Unity][Fix] Fix block memory plan to handle bool (#14357)
add 57b90700d2 [Unity][Transform] Introduce data-dependent operation of
reshape and its constant folding (#14282)
add 27f0ad3230 [Unity][Transform] Fix AMP tests (#14360)
add 20cc2ff300 [Unity] Add support to append relay op attrs in translator
(#14356)
add 79492b81d9 [Unity][WEB] Support async pipeline creation (#14362)
add 7974e0f816 [Unity][Pass] Fix FuseOps error if there is no output of a
given group (#14354)
add b5cdcec5d6 [Unity][Fix] Infer Layout must support negative axes
(#14365)
add eb7b020c22 [Unity] Add More Ops For FX Translator (#14348)
add 14f7b5af61 [Unity][TVMScript] Update GlobalVar `checked_type_` when
`emit_te` (#14367)
add a0b28c486e [Unity][Fix] Allow scalar layout initialization (#14370)
add 9253f0fef1 [Unity] Also include output dtype in simt MathInstruction
(#14372)
add 8d8ced72a1 [Unity][VM] Add CUDA graph vm builtins (#14371)
add 983b569325 [Unity] Add missing #include <array> (#14383)
add d6efe9f042 [Unity][Transform] SplitCallTIRByPattern and CUTLASS
backend (#14274)
add f7f45dcc8e [Unity] Support simple dynamic-shape-aware fusion (#14396)
add 34b32cdf68 [Unity][Op] Add stop_lift_params (#14368)
add 6eb12ff85a [Unity][TVMScript] Fix Shape Var occurrence in Tensor
annotation (#14404)
add ae3cf5f5f1 [Unity][Transform] Common Subexpression Elimination (#14361)
add 6818799e74 [Unity][QNN][Hexagon]Support Relax Constants in the QNN
TOPI operations (#14386)
add a8205ed0c1 [Unity][Op] Conv1d (#14388)
add 906f69056d [Unity] Fix getting shapes for cutlass BYOC kernels (#14411)
add 5426085988 [Unity][Op] Expose scale in `R.nn.attention` and add its
legalize op (#14412)
add 7f44914317 [Unity][Hexagon] Enable Relax VM for Hexagon (#14415)
add 94b9bbc6bc [Unity][Fix] Copy over module attrs in FuseTIR (#14418)
add 4d29a7cff8 [Unity] Handle extern func calls in static memory planning
(#14419)
add e424eea3dd [Unity] Include constant shapes in the profiler result
(#14428)
add b1346b97a9 [Unity][Fix] Annotate TIR op pattern could have no stores.
(#14420)
add beb25a701c [Unity] Minor updates to DataFlowBlockRewrite (#14431)
add 22c363a72a [Unity] Remove non-deterministic behavior from graph
pattern matching (#14417)
add 61f7edc1af [Unity][Graph matching] Automatically add `used-by`
constraints for `is_op` pattern (#14439)
add ad4bfbc5f9 [Unity][Op][Docs] Update comment for `call_tir_dyn` (#14441)
add 0a4ce4dfc3 [Unity][Graph matching] Clean up undo stack for parent and
child nodes properly (#14440)
add 52f6e639de [Unity] Pattern-based rewriting for dataflow block (#14446)
add 5c8b7af59c [Unity] Fix lint during for upstream black
This update added new revisions after undoing existing revisions.
That is to say, some revisions that were in the old version of the
branch are not in the new version. This situation occurs
when a user --force pushes a change and generates a repository
containing something like this:
* -- * -- B -- O -- O -- O (a425bc7a39)
\
N -- N -- N refs/heads/unity (5c8b7af59c)
You should already have received notification emails for all of the O
revisions, and so the following emails describe only the N revisions
from the common base, B.
Any revisions marked "omit" are not gone; other references still
refer to them. Any revisions marked "discard" are gone forever.
No new revisions were added by this update.
Summary of changes:
include/tvm/runtime/relax_vm/executable.h | 4 ++++
src/relax/transform/fuse_tir.cc | 24 ----------------------
src/runtime/hexagon/hexagon_module.h | 1 -
...meta_schedule_schedule_cuda_layout_transform.py | 1 +
web/emcc/webgpu_runtime.cc | 2 ++
5 files changed, 7 insertions(+), 25 deletions(-)
