Skip to content

Update ROCm to 7.1 (supersedes #8867) - #9297

Merged
lstein merged 11 commits into
mainfrom
rocm-7.1-update
Jun 24, 2026
Merged

Update ROCm to 7.1 (supersedes #8867)#9297
lstein merged 11 commits into
mainfrom
rocm-7.1-update

Conversation

@lstein

@lstein lstein commented Jun 23, 2026

Copy link
Copy Markdown
Collaborator

Supersedes #8867 (prepared as a new PR — no write access to the author's fork). Part of #8655, resolves #9130.

  • torch 2.10.0+rocm7.1 / torchvision 0.25.0+rocm7.1 / triton-rocm 3.6.0
  • Removed redundant pytorch-triton-rocm (torch 2.10 pulls triton-rocm itself)
  • Removed defunct docker/Dockerfile-rocm-full (CI builds docker/Dockerfile)
  • ARM-Linux explicitly unsupported via tool.uv.environments

heathen711 and others added 6 commits June 23, 2026 15:43
@heathen711 @lstein
Upgrade ROCm support from 6.3 to 7.1
418ae3f 
- Update PyTorch to 2.10.0+rocm7.1 (from 2.7.1+rocm6.3)
- Update TorchVision to 0.25.0+rocm7.1 (from 0.22.1+rocm6.3)
- Update PyTorch ROCm index URL to rocm7.1
- Update Docker ROCm runtime to 7.1.1 (from 6.3.4)

ROCm 7.1 adds support for 3 new GPU architectures:
- gfx950: CDNA3+ (Instinct MI350, MI325X)
- gfx1150: RDNA 3.5 (Strix Point APU)
- gfx1151: RDNA 3.5 (Strix Halo APU)

All 11 previously supported architectures remain compatible.
Total supported architectures: 14 (11 existing + 3 new)

Relates to #8655
@heathen711 @claude @lstein
Fix uv sync for ROCm 7.1 by explicitly declaring triton-rocm
33d2ebc 
The uv package manager couldn't resolve triton-rocm as a transitive
dependency of torch from the PyTorch custom index. This commit adds
triton-rocm==3.6.0 explicitly to the rocm extras with a Linux-only
platform marker.

Changes:
- Added triton-rocm==3.6.0; sys_platform == 'linux' to rocm extras
- Added triton-rocm source configuration pointing to torch-rocm index
- Already committed: torch version constraint relaxation (>=2.7.0,<3.0)
- Already committed: platform environment restrictions for ROCm

Verified with uv sync --extra rocm:
- torch==2.10.0+rocm7.1 installed correctly
- torchvision==0.25.0+rocm7.1 installed correctly
- triton-rocm==3.6.0 resolved successfully
- ROCm 7.1.25424 detected on AMD Radeon PRO V620
- All 8 model architectures accessible
- GPU compute operations verified

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
@heathen711 @lstein
Update uv lock and remove old dockerfile
4a91902
heathen711 @qwencoder @lstein
docs: update ROCm torch backend from 6.3 to 7.1
5944bd1 
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
heathen711 @qwencoder @lstein
fix: regenerate uv.lock with ROCm 7.1 dependencies
2f9e87d 
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
@lstein
remove redundant inclusion of torch-triton wheel && confirm no aarch6…
a3ad058 
…4-linux support
@lstein
lstein requested a review from blessedcoolant as a code owner June 23, 2026 19:49
@github-actions github-actions Bot added docker Root docs PRs that change docs python-deps PRs that change python dependencies labels Jun 23, 2026
@lstein lstein self-assigned this Jun 23, 2026
@lstein lstein added the 6.13.5 Library Updates label Jun 23, 2026
@lstein lstein moved this to 6.13.5 LIBRARY UPDATES in Invoke - Community Roadmap Jun 23, 2026
@github-actions github-actions Bot added the frontend PRs that change frontend files label Jun 23, 2026
lstein and others added 4 commits June 23, 2026 20:26
@lstein @claude
chore: drop spurious typegen changes from ROCm PR
b01fe8d 
The ROCm 7.1 update is dependency-only and does not alter any backend
route or model, so schema.ts/openapi.json should match main. The earlier
regeneration introduced spurious diffs (removed @example blocks,
reformatted JSDoc) from a local tooling/backend mismatch, failing the
typegen-checks "compare files" step. Restore both files to main.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@lstein @claude
fix(deps): keep macOS on torch 2.7.x; gate rocm extra to linux
16c3f77 
Loosening the base torch pin to <3.0 (so the rocm extra could use
2.10.0+rocm7.1) let macOS resolve torch 2.12.0, whose MPS backend fails
with OOM on GitHub's hosted macOS runners (no usable Metal GPU),
breaking the py3.x macos-default pytest jobs.

Split the base pin by platform so macOS stays on 2.7.x (matching main)
while linux/win still allow >=2.10, and gate the rocm extra's
torch/torchvision to sys_platform == 'linux' (ROCm is x86_64-linux-only)
so non-linux resolution is unaffected.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@lstein @claude
chore(frontend): regenerate schema for pydantic 2.13 docstring descri…
6edf84b 
…ptions

The dependency update bumps pydantic 2.11.7 -> 2.13.4, which now emits a
plain class docstring as the JSON-schema description. CacheStats gains
"Collect statistics on cache performance." in both openapi.json and
schema.ts. This is the sole schema delta from the bump and is what the
openapi-checks/typegen-checks jobs regenerate.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@lstein
Merge branch 'main' into rocm-7.1-update
b17cba0
@lstein
lstein merged commit f6e8460 into main Jun 24, 2026
17 checks passed
@lstein
lstein deleted the rocm-7.1-update branch June 24, 2026 01:16
dunkeroni pushed a commit to dunkeroni/InvokeAI that referenced this pull request Jun 29, 2026
@lstein @heathen711
@claude heathen711 @qwencoder @dunkeroni
Update ROCm to 7.1 (supersedes invoke-ai#8867) (invoke-ai#9297)
177e7df 
* Upgrade ROCm support from 6.3 to 7.1

- Update PyTorch to 2.10.0+rocm7.1 (from 2.7.1+rocm6.3)
- Update TorchVision to 0.25.0+rocm7.1 (from 0.22.1+rocm6.3)
- Update PyTorch ROCm index URL to rocm7.1
- Update Docker ROCm runtime to 7.1.1 (from 6.3.4)

ROCm 7.1 adds support for 3 new GPU architectures:
- gfx950: CDNA3+ (Instinct MI350, MI325X)
- gfx1150: RDNA 3.5 (Strix Point APU)
- gfx1151: RDNA 3.5 (Strix Halo APU)

All 11 previously supported architectures remain compatible.
Total supported architectures: 14 (11 existing + 3 new)

Relates to invoke-ai#8655

* Fix uv sync for ROCm 7.1 by explicitly declaring triton-rocm

The uv package manager couldn't resolve triton-rocm as a transitive
dependency of torch from the PyTorch custom index. This commit adds
triton-rocm==3.6.0 explicitly to the rocm extras with a Linux-only
platform marker.

Changes:
- Added triton-rocm==3.6.0; sys_platform == 'linux' to rocm extras
- Added triton-rocm source configuration pointing to torch-rocm index
- Already committed: torch version constraint relaxation (>=2.7.0,<3.0)
- Already committed: platform environment restrictions for ROCm

Verified with uv sync --extra rocm:
- torch==2.10.0+rocm7.1 installed correctly
- torchvision==0.25.0+rocm7.1 installed correctly
- triton-rocm==3.6.0 resolved successfully
- ROCm 7.1.25424 detected on AMD Radeon PRO V620
- All 8 model architectures accessible
- GPU compute operations verified

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

* Update uv lock and remove old dockerfile

* docs: update ROCm torch backend from 6.3 to 7.1

Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>

* fix: regenerate uv.lock with ROCm 7.1 dependencies

Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>

* remove redundant inclusion of torch-triton wheel && confirm no aarch64-linux support

* chore(frontend): openapi and typegen

* chore: drop spurious typegen changes from ROCm PR

The ROCm 7.1 update is dependency-only and does not alter any backend
route or model, so schema.ts/openapi.json should match main. The earlier
regeneration introduced spurious diffs (removed @example blocks,
reformatted JSDoc) from a local tooling/backend mismatch, failing the
typegen-checks "compare files" step. Restore both files to main.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(deps): keep macOS on torch 2.7.x; gate rocm extra to linux

Loosening the base torch pin to <3.0 (so the rocm extra could use
2.10.0+rocm7.1) let macOS resolve torch 2.12.0, whose MPS backend fails
with OOM on GitHub's hosted macOS runners (no usable Metal GPU),
breaking the py3.x macos-default pytest jobs.

Split the base pin by platform so macOS stays on 2.7.x (matching main)
while linux/win still allow >=2.10, and gate the rocm extra's
torch/torchvision to sys_platform == 'linux' (ROCm is x86_64-linux-only)
so non-linux resolution is unaffected.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* chore(frontend): regenerate schema for pydantic 2.13 docstring descriptions

The dependency update bumps pydantic 2.11.7 -> 2.13.4, which now emits a
plain class docstring as the JSON-schema description. CacheStats gains
"Collect statistics on cache performance." in both openapi.json and
schema.ts. This is the sole schema delta from the bump and is what the
openapi-checks/typegen-checks jobs regenerate.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Heathen711 <Heathen711@me.com>
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
Co-authored-by: heathen711 <heathen711@example.com>
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
lstein added a commit to stellarfeline/InvokeAI that referenced this pull request Jul 15, 2026
@lstein @claude
Merge origin/main into feat/docker-arm64
724766e 
Resolve pyproject.toml/uv.lock conflicts with the ROCm 7.1 (invoke-ai#9297) and
Transformers 5.5.4 (invoke-ai#9248) updates on main:

- Reapply the linux_aarch64 exclusion markers to the new rocm extra
  (torch 2.10.0+rocm7.1 / torchvision 0.25.0+rocm7.1 / triton-rocm 3.6.0).
- Extend tool.uv.environments (added on main) to include linux/aarch64;
  the previous value excluded it from resolution entirely, which would
  have made this PR a no-op.
- Add explicit PyPI-pinned aarch64 fallback entries inside the cpu/cuda/
  rocm extras. Without them, uv partitions the base torch/torchvision
  declarations into the no-extra conflict universe, so 'uv sync --extra
  cpu' on aarch64 would install no torch at all.
- Add an explicit pypi index source for torch/torchvision on linux
  aarch64; uv does not fall back to the default index inside an extra's
  resolution split when the extra-scoped source does not match.
- Regenerate uv.lock (uv 0.11.23, revision 3 like main); verified with
  CI's uv 0.6.10 via 'uv lock --locked'.

Verified from the lock: every extra on linux/aarch64 resolves
torch==2.7.1 + torchvision==0.22.1 from PyPI; x86_64/darwin/win32
resolution is unchanged from main.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@blessedcoolant blessedcoolant Awaiting requested review from blessedcoolant blessedcoolant is a code owner
@dunkeroni dunkeroni Awaiting requested review from dunkeroni dunkeroni is a code owner
@Pfannkuchensack Pfannkuchensack Awaiting requested review from Pfannkuchensack Pfannkuchensack is a code owner

Assignees

@lstein lstein

Labels

6.13.5 Library Updates docker docs PRs that change docs frontend PRs that change frontend files python-deps PRs that change python dependencies Root

Projects

Invoke - Community Roadmap
Status: 6.13.5 LIBRARY UPDATES

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

[enhancement]: Rocm 7.2

2 participants

@lstein @heathen711