GitLab

Signed-off-by: Tim Moon <tmoon@nvidia.com>
Co-authored-by: Przemyslaw Tredak <ptredak@nvidia.com>

* Use fused implementation of RoPE in MultiHeadAttention
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Fix freqs dtype
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

---------
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Add option to avoid updating transpose cache when possible
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Fix typo
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Use string kwarg for FP8 transpose caching
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Remove unused attr
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

---------
Signed-off-by: Tim Moon <tmoon@nvidia.com>
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

* Pass knobs for TP comm overlap instead of env vars
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>

* Comment out debugging print
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>

* Remove docstring
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>

* Remove debugging output
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>

---------
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>
Co-authored-by: Jaemin Choi <jaeminc@nvidia.com>

* Support GEMM-GELU fusion with split AG overlap
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>

* Fix linter complaints
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: Jaemin Choi <minitu77@gmail.com>

* Avoid code duplication
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>

* Fix issue with modifying tuple
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>

* Disable GEMM-GELU fusion when split AG overlap is not enabled
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>

* Add ub_split_ag parameter to LayerNormMLP unit test
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>

* Move knob into LayerNormMLP, auto-disable fusion when split AG overlap is not enabled
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>

* Revert changes to test_layernorm_mlp_accuracy
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: Jaemin Choi <minitu77@gmail.com>

---------
Signed-off-by: Jaemin Choi <jaeminc@nvidia.com>
Signed-off-by: Jaemin Choi <minitu77@gmail.com>
Co-authored-by: Jaemin Choi <jaeminc@nvidia.com>
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

* Implement fused kernel for FP8 scale update
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Add fused kernel for amax and scale update

Add unit test.
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Replace paddle.fluid imports with paddle.base
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Move fused kernel to core library
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Debug test
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Use FP8 update kernel in Paddle
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Debug FP8 scale update in Paddle
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Fix lint errors
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Debug Paddle test failures
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Make update kernel in-place for PyTorch
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Revert cudnn-frontend commit
Signed-off-by: Tim Moon <tmoon@nvidia.com>

---------
Signed-off-by: Tim Moon <tmoon@nvidia.com>
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

Use cloned scale_inv for fp8 cast
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Added new unfused softmax cuda kernel to support causal attention mask
Signed-off-by: Oleg Goncharov <ogoncharov@nvidia.com>

* Added test suite for unfused causal softmax kernel
Signed-off-by: Oleg Goncharov <ogoncharov@nvidia.com>

* Removed test cases with large matrices from the causal softmax test suite
Signed-off-by: Oleg Goncharov <ogoncharov@nvidia.com>

* Cleaned up the code per lint
Signed-off-by: Oleg Goncharov <ogoncharov@nvidia.com>

* Added a compute buffer to causal softmax testing suite to store intermediate results without casting
Signed-off-by: Oleg Goncharov <ogoncharov@nvidia.com>

* Added more tests cases
Signed-off-by: Oleg Goncharov <ogoncharov@nvidia.com>

* Relaxed absolute tolerance atol
Signed-off-by: Oleg Goncharov <ogoncharov@nvidia.com>

* Relaxed absolute tolerance for BF16
Signed-off-by: Oleg Goncharov <ogoncharov@nvidia.com>

---------
Signed-off-by: Oleg Goncharov <ogoncharov@nvidia.com>

* test alibi between fa and fu
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* move alibi slopes and bias to global to avoid repeating calculation
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* fix alibi slopes/bias generation
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* fix _is_flash_attention_supported to allow alibi type
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* disable padding mask when alibi is used for fused attn arbi backend
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* add support for custom [n_heads] alibi_slopes in flash, fused, unfused attention
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* clean up last commit
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* remove alibi_type=none tests as they are unnecessary
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* update cudnn-frontend to 1.0.2
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* change bias/dbias shape to allow b,1/1,h/b,h in arbi backend
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* tweak tests for arbi post_scale_bias [1,h,s,s] or alibi_slopes [n_heads]
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* change bias/dbias shape in max512 backend - incomplete
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* remove max512 changes from last commit and disable max512 (and arbi temporarily) for [b, h, s, s]; pending cuDNN backend support
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* clean up and tweak backend selection logic
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* replace || with () in docstring
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* fix bias shape for max512 backend
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* combine slopes/bias generation to one function get_alibi() and fix alibi tests
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* fix lint
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* fix PR557 bugs
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* Update transformer_engine/pytorch/attention.py
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: cyanguwa <8636796+cyanguwa@users.noreply.github.com>

* encapsulate global alibi tensors into a dict cache
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* reduce alibi slopes test size
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* update to cudnn-frontend 1.0.3
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* use dBias shape to define bias_b/bias_h because jax materializes dBias rather than Bias in bwd abstract
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

---------
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>
Signed-off-by: cyanguwa <8636796+cyanguwa@users.noreply.github.com>
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

Do not cache sequence lengths based on layer number
Signed-off-by: Tim Moon <tmoon@nvidia.com>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Add zero_centered_gamma option to RMSNorm
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Improving tests
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* More improvements to tests
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Tweaking the tolerances
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Fix LayerNormMLP test
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Update transformer_engine/common/rmsnorm/rmsnorm_api.cpp
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Update transformer_engine/common/rmsnorm/rmsnorm_api.cpp
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* docs suggestions
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Tweak tolerances with bfloat16
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

---------
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

* fixes for recomputation
Signed-off-by: Jimmy Zhang <jiemingz@nvidia.com>

* lint
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Fix onnx export [wip]
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* fix
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* register op; fixes
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

---------
Signed-off-by: Jimmy Zhang <jiemingz@nvidia.com>
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
Co-authored-by: Jimmy Zhang <jiemingz@nvidia.com>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Update cudnn frontend to 1.0.3 to fix cudnn v9 Nans
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* make d_out contiguous for bwd
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* remove cudnnDestroy to let torch handle it
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* Update transformer_engine/pytorch/attention.py
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: cyanguwa <8636796+cyanguwa@users.noreply.github.com>

* Update transformer_engine/pytorch/attention.py
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: cyanguwa <8636796+cyanguwa@users.noreply.github.com>

* Update transformer_engine/pytorch/attention.py
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: cyanguwa <8636796+cyanguwa@users.noreply.github.com>

---------
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>
Signed-off-by: cyanguwa <8636796+cyanguwa@users.noreply.github.com>
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

* Adding support of sequence parallelism
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Adding RoPE
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Fix wrong batch_logical_axes
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Rnaming FSDP outer env var
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Poring RoPE to Praxis layers.
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Porting GeLU + [FP8 Cast].
Signed-off-by: Ming Huang <mingh@nvidia.com>

* WAR to make XLA successfully match FP8 GEMM on FFN1 with GeLU.
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Allowing arbitrary dimension of NVShape for the workspace allocation
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Adding checkpoint_name to fused functions of mlp.py to get better perf with nn.scan.
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Modify with review feedback.
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Fix bugs
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Fix typo.
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Fixed for lint
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Follow review feedback to modify code.
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Fix typo.
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Port SP to Praxis
Signed-off-by: Ming-Xu Huang <mingh@nvidia.com>

* Fix an issue when enabling both GQA and RoPE.
Signed-off-by: Ming-Xu Huang <mingh@nvidia.com>

* Update docs
Signed-off-by: Ming-Xu Huang <mingh@nvidia.com>

---------
Signed-off-by: Ming Huang <mingh@nvidia.com>
Signed-off-by: Ming-Xu Huang <mingh@nvidia.com>

* Fix unfused GQA perf
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Remove WAR for Check failed: reduction_kind.has_value()
Signed-off-by: Reese Wang <rewang@nvidia.com>

---------
Signed-off-by: Reese Wang <rewang@nvidia.com>

* update cudnn cmake for v9
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* add back license information
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

---------
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

Fused rope computation in fp32
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

Do not allocate FP8 workspace buffers when params are FP8
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Replace paddle.fluid imports with paddle.base
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Remove paddle.fluid usage from tests
Signed-off-by: Tim Moon <tmoon@nvidia.com>

---------
Signed-off-by: Tim Moon <tmoon@nvidia.com>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

Fixed offloading for PyT version/ Added Attention activation offloading support/ Native FP8 support (#632)

* Fixed offloading for PyT version/ Added Attention activation offloading support/ Native FP8 support
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Removed activation offloading for fused attention
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Fixed the illegal memory access issue for activation offloading of attention
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Removed the version guard
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Pipeline failures fix
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Fixed lint erros
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Lint error fix
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

---------
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>
Co-authored-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Removed cudaMalloc/WorkspaceManager in JAX csrc. JAX custom ops now request buffers from XLA for their workspace tensors.
Signed-off-by: Alp Dener <adener@nvidia.com>

* removed unused GEMM C++ API in TE-JAX
Signed-off-by: Alp Dener <adener@nvidia.com>

* fixed typo in layernorm_geglu_fp8_mlp and removed unnecessary shape reductions in primitives
Signed-off-by: Alp Dener <adener@nvidia.com>

* fixed import order for linting
Signed-off-by: Alp Dener <adener@nvidia.com>

* fixed custom op errors due to incorrect static arg nums in JAX jit
Signed-off-by: Alp Dener <adener@nvidia.com>

* shifted cudnnSetStream further down the kernel to avoid error when executing dummy kernel call with nullptr stream
Signed-off-by: Alp Dener <adener@nvidia.com>

* fixed linting errors for blank lines
Signed-off-by: Alp Dener <adener@nvidia.com>

---------
Signed-off-by: Alp Dener <adener@nvidia.com>

* use separate qkv
Signed-off-by: jaywan <jaywan@nvidia.com>

* add support for GQA
Signed-off-by: jaywan <jaywan@nvidia.com>

* minor changes
Signed-off-by: Shijie Wang <jaywan@nvidia.com>

* change rtol
Signed-off-by: Shijie Wang <jaywan@nvidia.com>

* fix reshape issue
Signed-off-by: Shijie Wang <jaywan@nvidia.com>

---------
Signed-off-by: jaywan <jaywan@nvidia.com>
Signed-off-by: Shijie Wang <jaywan@nvidia.com>

Fix MHA docstring
Signed-off-by: Isaac Ong <isaacong.jw@gmail.com>

* fused apply rope
Signed-off-by: Xin Yao <xiny@nvidia.com>

* Apply suggestions from code review
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: Xin Yao <yaox12@outlook.com>

* resolve comments
Signed-off-by: Xin Yao <xiny@nvidia.com>

* make rotary_percent optional
Signed-off-by: Xin Yao <xiny@nvidia.com>

* fix ci
Signed-off-by: Xin Yao <xiny@nvidia.com>

* fix test
Signed-off-by: Xin Yao <xiny@nvidia.com>

* add rope test to qa
Signed-off-by: Xin Yao <xiny@nvidia.com>

* fix linting
Signed-off-by: Xin Yao <xiny@nvidia.com>

* sync apex: add transpose_output_memory
Signed-off-by: Xin Yao <xiny@nvidia.com>

* small fix
Signed-off-by: Xin Yao <xiny@nvidia.com>

* sync apex: fuse sin/cos
Signed-off-by: Xin Yao <xiny@nvidia.com>

* sync apex: fused rope for thd format
Signed-off-by: Xin Yao <xiny@nvidia.com>

* fix lint
Signed-off-by: Xin Yao <xiny@nvidia.com>

* Fix license headers
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* add support for bshd format
Signed-off-by: Xin Yao <xiny@nvidia.com>

* support different seq length
Signed-off-by: Xin Yao <xiny@nvidia.com>

* update
Signed-off-by: Xin Yao <xiny@nvidia.com>

* update copyright
Signed-off-by: Xin Yao <xiny@nvidia.com>

* remove transpose_output_memory
Signed-off-by: Xin Yao <xiny@nvidia.com>

* Make outputs contiguous in SBHD case
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

---------
Signed-off-by: Xin Yao <xiny@nvidia.com>
Signed-off-by: Xin Yao <yaox12@outlook.com>
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Co-authored-by: Przemyslaw Tredak <ptredak@nvidia.com>

Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

[PyTorch] fix forward attention_type in MultiheadAttention
Signed-off-by: Markus Schnoes <markus.schnoes@gmx.de>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

[PyTorch] Workaround for incorrect output from torch.cuda.is_bf16_compatible() on V100s and TU102s (#626)

* replaced torch.cuda.is_bf16_compatible() with explicit sm_80 check via torch.cuda.get_device_capability()
Signed-off-by: Alp Dener <adener@nvidia.com>

* implement te.utils.is_bf16_compatible() to replace torch.cuda counterpart
Signed-off-by: Alp Dener <adener@nvidia.com>

---------
Signed-off-by: Alp Dener <adener@nvidia.com>

* Support building using the manylinux docker image. libpython is only required for embedded python.
Signed-off-by: Luke Petre <lpetre@midjourney.com>

* Be explicit about which python to use in cmake
Signed-off-by: Luke Petre <lpetre@midjourney.com>

* Remove cmake version check
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: Luke Petre <lpetre@gmail.com>

---------
Signed-off-by: Luke Petre <lpetre@midjourney.com>
Signed-off-by: Luke Petre <lpetre@gmail.com>
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

* added missing parameter materialization on real device for LayerNorm and RMSNorm
Signed-off-by: Alp Dener <adener@nvidia.com>

* added new unittest for deferred initialization and modified parameter materialization to support standalone execution outside of FSDP
Signed-off-by: Alp Dener <adener@nvidia.com>

* restored tensor parallel attributes that were being wiped out by the parameter reset
Signed-off-by: Alp Dener <adener@nvidia.com>

* fixed incorrect order of fp8 metadata initialization
Signed-off-by: Alp Dener <adener@nvidia.com>

* added deferred init unittest to the QA script
Signed-off-by: Alp Dener <adener@nvidia.com>

---------
Signed-off-by: Alp Dener <adener@nvidia.com>

Signed-off-by: Markus Schnoes <markus.schnoes@gmx.de>

Activation offloading to CPU's for the Linear, Layernorm Linear and the Layernorm MLP modules (#571)

* Added support activation offloading to CPU's
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Moving CPU offloading library to TE
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Restructured code, added switch to choose between weight/activation offloading
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Removed arg during constructor
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Fix nit-pick errors
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Documentation fixes
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Fix to the code block in docs
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Added offloading unit test
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Fixed formatting
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* wgrad fusion fix, minor errors and lint
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Errors, test, lint
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* RM test file
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Fixed stray PyT tensors in LayernormMLP getting offloaded
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Fixed typi
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>

* Fix offloading for rmsnorm, rm test
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Fix errors
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Float8Tensor compatible offloading
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Cleanup
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

---------
Signed-off-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
Co-authored-by: Selvaraj Anandaraj <selvaraja@login-eos01.eos.clusters.nvidia.com>
Co-authored-by: Przemyslaw Tredak <ptredak@nvidia.com>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

fix failing tests due to PR #557
Signed-off-by: Sudhakar Singh <sudhakars@nvidia.com>
Co-authored-by: cyanguwa <8636796+cyanguwa@users.noreply.github.com>

Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* make TransformerLayer accept a `bshd` or `sbhd` tensor format
Signed-off-by: Sudhakar Singh <sudhakars@nvidia.com>

* Fixes from feedback
Signed-off-by: Sudhakar Singh <sudhakars@nvidia.com>

* more feedback fixes
Signed-off-by: Sudhakar Singh <sudhakars@nvidia.com>

* remove incorrect info from docstring
Signed-off-by: Sudhakar Singh <sudhakars@nvidia.com>

* fix from feedback
Signed-off-by: Sudhakar Singh <sudhakars@nvidia.com>

---------
Signed-off-by: Sudhakar Singh <sudhakars@nvidia.com>

* Implemented deferred initialization via `device='meta'` option for te.Linear and added new PyTorch example to demonstrate its use with FullyShardedDataParallel execution.
Signed-off-by: Alp Dener <adener@nvidia.com>

* correcting Float8Tensor initialization and fixing linting errors
Signed-off-by: Alp Dener <adener@nvidia.com>

* removed duplicate code from upstream rebase, local tests passing
Signed-off-by: Alp Dener <adener@nvidia.com>

* improved comments/documentation for FSDP example
Signed-off-by: Alp Dener <adener@nvidia.com>

* converted reset_parameters() into a base module function
Signed-off-by: Alp Dener <adener@nvidia.com>

* fixed Float8Tensor creation with deferred init, all tests passing locally
Signed-off-by: Alp Dener <adener@nvidia.com>

* extended deferred initialization to all TE modules
Signed-off-by: Alp Dener <adener@nvidia.com>

* fixed linting errors
Signed-off-by: Alp Dener <adener@nvidia.com>

* removed unnecessary reference to the parent module of parameter, added clarifying comments in parameter reset
Signed-off-by: Alp Dener <adener@nvidia.com>

---------
Signed-off-by: Alp Dener <adener@nvidia.com>

* Support num_gqa_groups arguments
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Add GQA support on the JAX bridge code
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Fix the kv stride of the arbitrary backend
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Complete rewrite fused attention tests and add GQA coverage
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Support unfused GQA
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Calculate seqlen before the primitive for the better perf
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Add GQA layer tests
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Apply code style checks for te_jax
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Apply code style checks for tests
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Add num_gqa_groups doc
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Refine the qkv_type
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Correct the variable naming
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Handle Max512 CAUSAL
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Add WAR for the latest jax image
Signed-off-by: Reese Wang <rewang@nvidia.com>

---------
Signed-off-by: Reese Wang <rewang@nvidia.com>

* Actively free tensor in bwd
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* - Add inplace support for fp8 casting
- Allow skipping weight update in fp8 meta update
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Support weight caching for Linear
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Add weight caching for LayernormLinear
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Add weight caching for LayerNormMLP
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Add weight caching for Transformer layer
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Add PP unittests
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Fix CI
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

---------
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Adding Cast custom call
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Applying cast to the kernel of layernorm_fp8_dot
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Applying native cast to the kernel of fp8_dot.
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Apply Cast and native cast to layernorm_geglu_fp8_dot
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Fix the bug to enable layernorm_geglu_fp8_dot in LayernormMlp
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Modifiied code with the review feedback.
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Adding 2xACC control to FP8 GEMMs.
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Set precision as an static arg
Signed-off-by: Ming Huang <mingh@nvidia.com>

---------
Signed-off-by: Ming Huang <mingh@nvidia.com>

* Add SP for linear
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Add SP for LayerNormLinear
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Add SP for LayerNormMLP
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Add SP API for transformer layer
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Add sequence_parallel attr
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Add SP unittests for Transformer and Attention
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Fix compatibility with PaddleNLP
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>

* Copyright
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

---------
Signed-off-by: Tian Zheng (Engrg-Hardware 1) <tizheng@nvidia.com>
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>