Skip to content

[None] Add the torch source implementation for new params. - #89

Merged
nvchenghaoz merged 3 commits into
feat/ad-2025-07-07from
chenghao/torch-ref-0715
Jul 17, 2025
Merged

[None] Add the torch source implementation for new params.#89
nvchenghaoz merged 3 commits into
feat/ad-2025-07-07from
chenghao/torch-ref-0715

Conversation

@nvchenghaoz

Copy link
Copy Markdown

Since we added logit cap / attention sinks / sliding window for the attention backend operators, update the torch source implementation for new these parameters.

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>
@nvchenghaoz nvchenghaoz self-assigned this Jul 16, 2025
Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>
@nvchenghaoz
nvchenghaoz enabled auto-merge (squash) July 16, 2025 18:33

@lucaslie lucaslie left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks great. One follow-up task that could be interesting is to unify the bsnd and bnsd implementation and just use an argument for that.

Let's wait for #88 to land and then we can discuss further

@nvchenghaoz
nvchenghaoz merged commit 72b9ee9 into feat/ad-2025-07-07 Jul 17, 2025
1 check passed
lucaslie pushed a commit that referenced this pull request Jul 18, 2025
* Add the torch ref implementation for new params.

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>

* Remove comment

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>

---------

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>
lucaslie added a commit that referenced this pull request Jul 18, 2025
lucaslie pushed a commit that referenced this pull request Jul 21, 2025
* Add the torch ref implementation for new params.

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>

* Remove comment

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>

---------

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>
lucaslie added a commit that referenced this pull request Jul 21, 2025
nvchenghaoz added a commit that referenced this pull request Jul 26, 2025
* attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>

* Update the torch ref op

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>

* Revert "attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests"

This reverts commit 5743fb3.

---------

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>
Co-authored-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
MrGeva pushed a commit that referenced this pull request Jul 29, 2025
* attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>

* Update the torch ref op

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>

* Revert "attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests"

This reverts commit 5743fb3.

---------

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>
Co-authored-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
h-guo18 pushed a commit that referenced this pull request Jul 30, 2025
* attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>

* Update the torch ref op

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>

* Revert "attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests"

This reverts commit 5743fb3.

---------

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>
Co-authored-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
lucaslie pushed a commit that referenced this pull request Jul 31, 2025
* attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>

* Update the torch ref op

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>

* Revert "attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests"

This reverts commit 5743fb3.

---------

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>
Co-authored-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
Fridah-nv added a commit that referenced this pull request Jul 31, 2025
* attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>

* Update the torch ref op

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>

* Revert "attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests"

This reverts commit 5743fb3.

---------

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>
Co-authored-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
Fridah-nv added a commit that referenced this pull request Jul 31, 2025
* attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>

* Update the torch ref op

Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>

* Revert "attention matcher with torch._inductor pattern matcher,matching repeat kv, sdpa and group attention, update unit tests"

This reverts commit 5743fb3.

---------

Signed-off-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
Signed-off-by: nvchenghaoz <211069071+nvchenghaoz@users.noreply.github.com>
Co-authored-by: Frida Hou <201670829+Fridah-nv@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants