wasi-nn: only support components#8530
Closed
abrown wants to merge 3 commits into
Closed
Conversation
abrown
force-pushed
the
wasi-nn-only-wit
branch
2 times, most recently
from
51bc48a to
bc567a5
Compare
Member
|
From a technical perspective this all looks fine to me, thanks! For a "is this the time to do this or not" I'd defer to the wasi-nn subgroup, which I know you're intimately tied into as well and likely have a consensus before making this PR |
The wasi-nn specification has moved forward, adopting component model features such as resources. These cannot be specified and implemented using WITX, so the WITX implementation is removed, leaving the WIT implementation intact. In the future, preview1-ABI support could be re-added with a component adapter, as other WASI proposals have done.
With WITX support removed, these files no longer are needed in-tree.
The `build.rs` script was only telling Wiggle where the WITX files
lived. We do need to retain it, though, for `env!("OUT_DIR")` to still
work.
prtest:full
abrown
added a commit
to abrown/bytecodealliance-meetings
that referenced
this pull request
May 3, 2024
This PR removes preview1 support for wasi-nn in Wasmtime; let's discuss the ramifications. [#8530]: bytecodealliance/wasmtime#8530
abrown
added a commit
to bytecodealliance/meetings
that referenced
this pull request
May 3, 2024
This PR removes preview1 support for wasi-nn in Wasmtime; let's discuss the ramifications. [#8530]: bytecodealliance/wasmtime#8530
abrown
added a commit
that referenced
this pull request
Jun 25, 2024
Recent discussion in the wasi-nn proposal (see [wasi-nn#59], e.g.) has concluded that the right approach for representing wasi-nn "things" (tensors, graph, etc.) is with a component model _resource_. This sweeping change brings Wasmtime's implementation in line with that decision. Initially I had structured this PR to remove all of the WITX-based implementation (#8530). But, after consulting in a Zulip [thread] on what other WASI proposals aim to do, this PR pivoted to support _both_` the WITX-based and WIT-based ABIs (e.g., preview1 era versus preview2, component model era). What is clear is that the WITX-based specification will remain "frozen in time" while the WIT-based implementation moves forward. What that means for this PR is a "split world" paradigm. In many places, we have to distinguish between the `wit` and `witx` versions of the same thing. This change isn't the end state yet: it's a big step forward towards bringing Wasmtime back in line with the WIT spec but, despite my best efforts, doesn't fully fix all the TODOs left behind over several years of development. I have, however, taken the liberty to refactor and fix various parts as I came across them (e.g., the ONNX backend). I plan to continue working on this in future PRs to figure out a good error paradigm (the current one is too wordy) and device residence. [wasi-nn#59]: WebAssembly/wasi-nn#59 [thread]: https://bytecodealliance.zulipchat.com/#narrow/stream/219900-wasi/topic/wasi-nn's.20preview1.20vs.20preview2.20timeline
abrown
added a commit
that referenced
this pull request
Jun 25, 2024
Recent discussion in the wasi-nn proposal (see [wasi-nn#59], e.g.) has concluded that the right approach for representing wasi-nn "things" (tensors, graph, etc.) is with a component model _resource_. This sweeping change brings Wasmtime's implementation in line with that decision. Initially I had structured this PR to remove all of the WITX-based implementation (#8530). But, after consulting in a Zulip [thread] on what other WASI proposals aim to do, this PR pivoted to support _both_` the WITX-based and WIT-based ABIs (e.g., preview1 era versus preview2, component model era). What is clear is that the WITX-based specification will remain "frozen in time" while the WIT-based implementation moves forward. What that means for this PR is a "split world" paradigm. In many places, we have to distinguish between the `wit` and `witx` versions of the same thing. This change isn't the end state yet: it's a big step forward towards bringing Wasmtime back in line with the WIT spec but, despite my best efforts, doesn't fully fix all the TODOs left behind over several years of development. I have, however, taken the liberty to refactor and fix various parts as I came across them (e.g., the ONNX backend). I plan to continue working on this in future PRs to figure out a good error paradigm (the current one is too wordy) and device residence. [wasi-nn#59]: WebAssembly/wasi-nn#59 [thread]: https://bytecodealliance.zulipchat.com/#narrow/stream/219900-wasi/topic/wasi-nn's.20preview1.20vs.20preview2.20timeline prtest:full
abrown
added a commit
to abrown/wasmtime
that referenced
this pull request
Jun 25, 2024
Recent discussion in the wasi-nn proposal (see [wasi-nn#59], e.g.) has concluded that the right approach for representing wasi-nn "things" (tensors, graph, etc.) is with a component model _resource_. This sweeping change brings Wasmtime's implementation in line with that decision. Initially I had structured this PR to remove all of the WITX-based implementation (bytecodealliance#8530). But, after consulting in a Zulip [thread] on what other WASI proposals aim to do, this PR pivoted to support _both_` the WITX-based and WIT-based ABIs (e.g., preview1 era versus preview2, component model era). What is clear is that the WITX-based specification will remain "frozen in time" while the WIT-based implementation moves forward. What that means for this PR is a "split world" paradigm. In many places, we have to distinguish between the `wit` and `witx` versions of the same thing. This change isn't the end state yet: it's a big step forward towards bringing Wasmtime back in line with the WIT spec but, despite my best efforts, doesn't fully fix all the TODOs left behind over several years of development. I have, however, taken the liberty to refactor and fix various parts as I came across them (e.g., the ONNX backend). I plan to continue working on this in future PRs to figure out a good error paradigm (the current one is too wordy) and device residence. [wasi-nn#59]: WebAssembly/wasi-nn#59 [thread]: https://bytecodealliance.zulipchat.com/#narrow/stream/219900-wasi/topic/wasi-nn's.20preview1.20vs.20preview2.20timeline prtest:full
Merged
abrown
added a commit
to abrown/wasmtime
that referenced
this pull request
Jun 26, 2024
Recent discussion in the wasi-nn proposal (see [wasi-nn#59], e.g.) has concluded that the right approach for representing wasi-nn "things" (tensors, graph, etc.) is with a component model _resource_. This sweeping change brings Wasmtime's implementation in line with that decision. Initially I had structured this PR to remove all of the WITX-based implementation (bytecodealliance#8530). But, after consulting in a Zulip [thread] on what other WASI proposals aim to do, this PR pivoted to support _both_` the WITX-based and WIT-based ABIs (e.g., preview1 era versus preview2, component model era). What is clear is that the WITX-based specification will remain "frozen in time" while the WIT-based implementation moves forward. What that means for this PR is a "split world" paradigm. In many places, we have to distinguish between the `wit` and `witx` versions of the same thing. This change isn't the end state yet: it's a big step forward towards bringing Wasmtime back in line with the WIT spec but, despite my best efforts, doesn't fully fix all the TODOs left behind over several years of development. I have, however, taken the liberty to refactor and fix various parts as I came across them (e.g., the ONNX backend). I plan to continue working on this in future PRs to figure out a good error paradigm (the current one is too wordy) and device residence. [wasi-nn#59]: WebAssembly/wasi-nn#59 [thread]: https://bytecodealliance.zulipchat.com/#narrow/stream/219900-wasi/topic/wasi-nn's.20preview1.20vs.20preview2.20timeline prtest:full
abrown
added a commit
to abrown/wasmtime
that referenced
this pull request
Jun 26, 2024
Recent discussion in the wasi-nn proposal (see [wasi-nn#59], e.g.) has concluded that the right approach for representing wasi-nn "things" (tensors, graph, etc.) is with a component model _resource_. This sweeping change brings Wasmtime's implementation in line with that decision. Initially I had structured this PR to remove all of the WITX-based implementation (bytecodealliance#8530). But, after consulting in a Zulip [thread] on what other WASI proposals aim to do, this PR pivoted to support _both_` the WITX-based and WIT-based ABIs (e.g., preview1 era versus preview2, component model era). What is clear is that the WITX-based specification will remain "frozen in time" while the WIT-based implementation moves forward. What that means for this PR is a "split world" paradigm. In many places, we have to distinguish between the `wit` and `witx` versions of the same thing. This change isn't the end state yet: it's a big step forward towards bringing Wasmtime back in line with the WIT spec but, despite my best efforts, doesn't fully fix all the TODOs left behind over several years of development. I have, however, taken the liberty to refactor and fix various parts as I came across them (e.g., the ONNX backend). I plan to continue working on this in future PRs to figure out a good error paradigm (the current one is too wordy) and device residence. [wasi-nn#59]: WebAssembly/wasi-nn#59 [thread]: https://bytecodealliance.zulipchat.com/#narrow/stream/219900-wasi/topic/wasi-nn's.20preview1.20vs.20preview2.20timeline prtest:full
abrown
added a commit
to abrown/wasmtime
that referenced
this pull request
Jun 26, 2024
Recent discussion in the wasi-nn proposal (see [wasi-nn#59], e.g.) has concluded that the right approach for representing wasi-nn "things" (tensors, graph, etc.) is with a component model _resource_. This sweeping change brings Wasmtime's implementation in line with that decision. Initially I had structured this PR to remove all of the WITX-based implementation (bytecodealliance#8530). But, after consulting in a Zulip [thread] on what other WASI proposals aim to do, this PR pivoted to support _both_` the WITX-based and WIT-based ABIs (e.g., preview1 era versus preview2, component model era). What is clear is that the WITX-based specification will remain "frozen in time" while the WIT-based implementation moves forward. What that means for this PR is a "split world" paradigm. In many places, we have to distinguish between the `wit` and `witx` versions of the same thing. This change isn't the end state yet: it's a big step forward towards bringing Wasmtime back in line with the WIT spec but, despite my best efforts, doesn't fully fix all the TODOs left behind over several years of development. I have, however, taken the liberty to refactor and fix various parts as I came across them (e.g., the ONNX backend). I plan to continue working on this in future PRs to figure out a good error paradigm (the current one is too wordy) and device residence. [wasi-nn#59]: WebAssembly/wasi-nn#59 [thread]: https://bytecodealliance.zulipchat.com/#narrow/stream/219900-wasi/topic/wasi-nn's.20preview1.20vs.20preview2.20timeline prtest:full
abrown
added a commit
to abrown/wasmtime
that referenced
this pull request
Jun 26, 2024
Recent discussion in the wasi-nn proposal (see [wasi-nn#59], e.g.) has concluded that the right approach for representing wasi-nn "things" (tensors, graph, etc.) is with a component model _resource_. This sweeping change brings Wasmtime's implementation in line with that decision. Initially I had structured this PR to remove all of the WITX-based implementation (bytecodealliance#8530). But, after consulting in a Zulip [thread] on what other WASI proposals aim to do, this PR pivoted to support _both_` the WITX-based and WIT-based ABIs (e.g., preview1 era versus preview2, component model era). What is clear is that the WITX-based specification will remain "frozen in time" while the WIT-based implementation moves forward. What that means for this PR is a "split world" paradigm. In many places, we have to distinguish between the `wit` and `witx` versions of the same thing. This change isn't the end state yet: it's a big step forward towards bringing Wasmtime back in line with the WIT spec but, despite my best efforts, doesn't fully fix all the TODOs left behind over several years of development. I have, however, taken the liberty to refactor and fix various parts as I came across them (e.g., the ONNX backend). I plan to continue working on this in future PRs to figure out a good error paradigm (the current one is too wordy) and device residence. [wasi-nn#59]: WebAssembly/wasi-nn#59 [thread]: https://bytecodealliance.zulipchat.com/#narrow/stream/219900-wasi/topic/wasi-nn's.20preview1.20vs.20preview2.20timeline prtest:full
github-merge-queue Bot
pushed a commit
that referenced
this pull request
Jun 27, 2024
* wasi-nn: use resources Recent discussion in the wasi-nn proposal (see [wasi-nn#59], e.g.) has concluded that the right approach for representing wasi-nn "things" (tensors, graph, etc.) is with a component model _resource_. This sweeping change brings Wasmtime's implementation in line with that decision. Initially I had structured this PR to remove all of the WITX-based implementation (#8530). But, after consulting in a Zulip [thread] on what other WASI proposals aim to do, this PR pivoted to support _both_` the WITX-based and WIT-based ABIs (e.g., preview1 era versus preview2, component model era). What is clear is that the WITX-based specification will remain "frozen in time" while the WIT-based implementation moves forward. What that means for this PR is a "split world" paradigm. In many places, we have to distinguish between the `wit` and `witx` versions of the same thing. This change isn't the end state yet: it's a big step forward towards bringing Wasmtime back in line with the WIT spec but, despite my best efforts, doesn't fully fix all the TODOs left behind over several years of development. I have, however, taken the liberty to refactor and fix various parts as I came across them (e.g., the ONNX backend). I plan to continue working on this in future PRs to figure out a good error paradigm (the current one is too wordy) and device residence. [wasi-nn#59]: WebAssembly/wasi-nn#59 [thread]: https://bytecodealliance.zulipchat.com/#narrow/stream/219900-wasi/topic/wasi-nn's.20preview1.20vs.20preview2.20timeline prtest:full * vet: audit `ort`-related crate updates * Simplify `WasiNnView` With @alexcrichton's help, this change removes the `trait WasiNnView` and `struct WasiNnImpl` wrapping that the WIT-based implementation used for accessing the host context. Instead, `WasiNnView` is now a `struct` containing the mutable references it needs to make things work. This unwraps one complex layer of abstraction, though it does have the downside that it complicates CLI code to split borrows of `Host`. * Temporarily disable WIT check * Refactor errors to use `trappable_error_type` This change simplifies the return types of the host implementations of the WIT-based wasi-nn. There is more work to be done with errors, e.g., to catch up with the upstream decision to return errors as resources. But this is better than the previous mess.
Member
Author
|
Superseded by #8873. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This is a relatively drastic change to remove all the preview1, WITX-based support from Wasmtime's wasi-nn implementation. This has ramifications:
bench-apilearns how to handle componentswasi-nnbindings crate is no longer usable with Wasmtime (it only understands the preview1 ABI)examplesare left for a later refactor)Why this change, then? The WebAssembly specification has pushed ahead, defining new functionality that depends on component model features (e.g., resources). So far, the
wasmtime-wasi-nnhas had its preview1- and preview2-ABI implementations coexisting side by side,wit.rsandwitx.rs. This was only possible because the WIT and WITX definitions were roughly similar. But I plan to update the Wasmtime implementation to support the new spec changes, which means the older preview1-ABI WITX code would immediately be out of date. And the differences between the updated WIT code and the old WITX code would only grow over time.