[None][feat] Add FlashInfer MLA attention backend support by Tracin · Pull Request #13428 · NVIDIA/TensorRT-LLM

Tracin · 2026-04-24T10:06:06Z

Implement FlashInfer-based Multi-head Latent Attention (MLA) for
DeepSeek-style models, enabling use of BatchPrefillWithRaggedKVCacheWrapper
for context and BatchMLAPagedAttentionWrapper for generation.

Key design decisions:

KV cache: zero-copy ckv/kpe views split from the existing paged KV pool buffer (kv_lora_rank + qk_rope_head_dim), avoiding extra allocation
Context phase: ragged prefill with expanded Q/K/V after appending latent to paged MLA caches via append_paged_mla_kv_cache
Generation phase: paged MLA decode with q_nope/q_pe split from fused_q; latent cache append handled before each decode step
RoPE applied externally in MLA.forward before latent_cache construction; FlashInfer kernels receive pre-RoPE'd inputs
Plans called unconditionally each forward pass (no stale caching)
backend=auto on BatchMLAPagedAttentionWrapper for FA3 on Hopper
mla_rope_generation stub copies RoPE'd q_pe into fused_q to satisfy the forward_absorption_generation call chain

Also fix metadata type assertion in attention.py to allow
FlashInferAttentionMetadata for MLA, and replace hasattr duck-typing
with isinstance check for kv_lens_cuda_runtime.

Validated with 84/84 unit tests and end-to-end DeepSeek-V3-Lite NVFP4
inference with --attention_backend FLASHINFER.

Support CUDA graph for flashinfer MLA.

Drop redundant contiguous/reshape copies in flashinfer MLA path.

[None][test] Fix FlashInfer MLA unit test and add B200/B300 smoke coverage

test_attention_mla_flashinfer: restrict to SM100 and set max_num_tokens to sum(context_sequence_lengths) so FlashInfer's batch_indices/positions buffers are sized for the full batch (was max(...), causing OOB copy).
Add TestDeepSeekV3Lite::test_bfloat16_flashinfer accuracy smoke test and register it in l0_b200.yml and l0_b300.yml.

Summary by CodeRabbit

Release Notes

New Features
- Added Multi-head Latent Attention (MLA) support to the FlashInfer attention backend, enabling optimized attention computation for MLA-based models.
Tests
- Added new bfloat16 accuracy tests with FlashInfer backend for enhanced model evaluation coverage and validation.

Description

Test Coverage

PR Checklist

Please review the following before submitting your PR:

PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.
PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.
Test cases are provided for new code paths (see test instructions)
Any new dependencies have been scanned for license and vulnerabilities
CODEOWNERS updated if ownership changes
Documentation updated as needed
Update tava architecture diagram if there is a significant design change in PR.
The reviewers assigned automatically/manually are appropriate for the PR.
Please check this after reviewing the above items as appropriate for this PR.

GitHub Bot Help

To see a list of available CI bot commands, please comment /bot help.

Tracin · 2026-04-24T10:06:16Z

/bot run

tensorrt-cicd · 2026-04-24T10:13:47Z

PR_Github #45385 [ run ] triggered by Bot. Commit: 91e5197 Link to invocation

coderabbitai · 2026-04-24T10:15:06Z

📝 Walkthrough

Walkthrough

This changeset adds MLA (Multi-head Latent Attention) support to the FlashInfer attention backend through dedicated planning dataclasses and state management. Core logic extends to handle both context and generation phases with phase-specific RoPE handling and output tensor allocation. Integration and unit tests validate the new functionality across GPU targets.

Changes

Cohort / File(s)	Summary
FlashInfer MLA Backend Implementation `tensorrt_llm/_torch/attention_backend/flashinfer.py`	Introduces `RaggedPlanParams` and `MLADecodePlanParams` dataclasses for MLA planning. Adds `support_mla()` class method and `mla_params` initialization. Extends `forward()` with `latent_cache` parameter. Implements context path (ragged prefill with planned decode) and generation path (latent append plus paged decode). Manages stable device buffers for CUDA graph compatibility and explicit synchronization for `plan()` execution outside stream capture.
MLA Module RoPE Specialization `tensorrt_llm/_torch/modules/attention.py`	Refines `extract_extra_attrs` metadata validation with unified `isinstance` check. Specializes RoPE application in `MLA.forward_impl()` for `FlashInferAttentionMetadata`: context and generation phases now slice `position_ids` before RoPE application and reconstruct `latent_cache_*` from RoPE'd `k_pe`. Updates `forward_absorption_generation` to compute `num_seqs` from `attn_metadata.num_generations` for FlashInfer metadata.
MLA FlashInfer Unit Tests `tests/unittest/_torch/attention/test_attention_mla.py`	Adds `apply_mla_rope` utility for MLA-style RoPE computation. Introduces `test_attention_mla_flashinfer` parametrized test (SM100 GPU with BF16 constraint). Implements backend-specific RoPE pre-rotation for context and generation phases, bypassing prior fused RoPE paths for FlashInfer execution. Retains non-FlashInfer generation path with prior cuSeqlen allocation logic.
Integration Test Cases `tests/integration/defs/accuracy/test_llm_api_pytorch.py`, `tests/integration/test_lists/test-db/l0_b200.yml`, `tests/integration/test_lists/test-db/l0_b300.yml`	Adds `test_bfloat16_flashinfer()` method to accuracy test suite with BF16-only variant running `GSM8K` evaluation. Registers test in `l0_b200` and `l0_b300` integration test lists. Uses low-memory skip marker for gating.

Sequence Diagram

sequenceDiagram
    participant Test as Test/Application
    participant FA as FlashInferAttention
    participant MLAModule as MLA Module
    participant FFI as FlashInfer Library
    participant CUDA as CUDA/GPU

    Test->>FA: forward(q, k, v, attn_metadata, latent_cache, phase)
    
    alt Context Phase (Prefill)
        FA->>FA: plan_ragged() for KV cache layout
        FA->>CUDA: allocate stable device buffers
        FA->>CUDA: sync (explicit plan execution)
        
        FA->>MLAModule: extract RoPE positions from attn_metadata
        MLAModule->>MLAModule: apply_mla_rope(k_pe, positions)
        MLAModule->>FA: return RoPE'd k_pe
        
        FA->>FFI: append_paged_mla_kv_cache(RoPE'd k, v, latent_cache)
        FFI->>CUDA: update MLA latent cache
        
        FA->>FFI: forward_ragged_mla(q, latent_cache)
        FFI->>CUDA: compute attention with v_head_dim output
        FA->>Test: return context output (v_head_dim per head)
    else Generation Phase (Decode)
        MLAModule->>MLAModule: apply_mla_rope(q_pe, k_pe, positions)
        MLAModule->>FA: return RoPE'd tensors
        
        FA->>FFI: append_paged_mla_kv_cache(RoPE'd k, v, latent_cache)
        FFI->>CUDA: update MLA latent cache
        
        FA->>FA: execute cached BatchMLAPagedAttentionWrapper plan
        FA->>FFI: forward_paged_mla(q, planned_decode_wrapper)
        FFI->>CUDA: compute attention with kv_lora_rank output
        FA->>Test: return generation output (kv_lora_rank per head)
    end

Estimated code review effort

🎯 4 (Complex) | ⏱️ ~55 minutes

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 42.86% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly summarizes the main change: adding FlashInfer MLA attention backend support.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description check	✅ Passed	PR description includes detailed technical implementation details, design decisions, test coverage information, and a completed PR checklist confirming guideline compliance.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

Create PR with unit tests

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

coderabbitai

Actionable comments posted: 5

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@tensorrt_llm/_torch/attention_backend/flashinfer.py`:
- Around line 158-185: plan_ragged currently calls
BatchPrefillWithRaggedKVCacheWrapper.plan() during forward execution which can
violate CUDA graph capture; remove the call from plan_ragged so it only
initializes/returns self._ragged_prefill_wrapper, and move the actual planning
call into prepare (mirroring MLA decode): in prepare, ensure you create
self._ragged_prefill_wrapper if None and invoke
self._ragged_prefill_wrapper.plan(...) with qo_indptr, kv_indptr and the same
fields from plan_params (num_heads, num_kv_heads, head_dim, head_dim_vo,
use_fp16_qk_reduction=False, causal=True, q_data_type, kv_data_type, sm_scale)
so planning happens outside the forward path.
- Around line 927-956: This path runs ragged prefill only on the new chunk's
q_ctx/k_ctx/v_ctx and ignores any cached paged MLA KV produced by
append_paged_mla_kv_cache(), causing incorrect attention when a cached prefix
exists; update the logic around metadata.plan_ragged and wrapper.run to detect a
cached paged MLA KV (e.g., check a flag like metadata.has_paged_mla_kv_cache or
metadata.paged_mla_kv_present) and if present, short-circuit this branch (either
skip ragged prefill and use the paged-KV-aware path or fall back to a full
prefill that reads the cached KV) so that q_ctx/k_ctx/v_ctx are combined with
the cached prefix rather than used in isolation. Ensure the guard is placed
before creating RaggedPlanParams/wrapper and before wrapper.run so the cached
prefix is never ignored.

In `@tests/integration/test_lists/test-db/l0_b200.yml`:
- Line 39: The QA scheduled functional lists are missing the integration test
node; add the test identifier
accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16_flashinfer
to the QA functional list named llm_function_full.txt (and also append it to
llm_function_core.txt if you want core single-node FlashInfer coverage),
ensuring the entry follows the existing README format for functional test cases
so it propagates from the pre-merge list into the QA scheduled lists.

In `@tests/unittest/_torch/attention/test_attention_mla.py`:
- Around line 776-785: gen_metadata_kwargs currently sets max_num_tokens to
sum(context_sequence_lengths) which under-allocates for multi-token generation;
change max_num_tokens to reflect the tokens in the current decode batch (i.e.,
use generation_seq_len_q * len(context_sequence_lengths) or otherwise sum the
per-request generation lengths) so FlashInferAttentionMetadata is sized for the
generation_seq_len_q across the decode batch; update the assignment to the
max_num_tokens key in gen_metadata_kwargs accordingly (referencing
gen_metadata_kwargs, max_num_tokens, context_sequence_lengths, and
generation_seq_len_q).
- Around line 795-798: The assignment to gen_metadata_kwargs['enable_flash_mla']
is incorrectly indented causing a Flake8 E123; locate the conditional that
checks backend_name == "TRTLLM" and reformat the assignment so the key and its
value expression (torch.cuda.get_device_capability() == (9, 0)) are on the same
logical line or properly aligned (e.g., gen_metadata_kwargs['enable_flash_mla']
= torch.cuda.get_device_capability() == (9, 0)) to satisfy linting while
preserving the existing condition and semantics.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

Push a commit to this branch (recommended)
Create a new PR with the fixes

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: a7c2d5b9-66d6-49b5-aeac-4a5c2ec68275

📥 Commits

Reviewing files that changed from the base of the PR and between c4b8e8e and 91e5197.

📒 Files selected for processing (6)

tensorrt_llm/_torch/attention_backend/flashinfer.py
tensorrt_llm/_torch/modules/attention.py
tests/integration/defs/accuracy/test_llm_api_pytorch.py
tests/integration/test_lists/test-db/l0_b200.yml
tests/integration/test_lists/test-db/l0_b300.yml
tests/unittest/_torch/attention/test_attention_mla.py

tensorrt-cicd · 2026-04-24T11:20:35Z

PR_Github #45385 [ run ] completed with state FAILURE. Commit: 91e5197
/LLM/main/L0_MergeRequest_PR pipeline #35626 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

Link to invocation

Tracin · 2026-04-27T02:43:20Z

/bot run

tensorrt-cicd · 2026-04-27T02:49:46Z

PR_Github #45628 [ run ] triggered by Bot. Commit: d253811 Link to invocation

tensorrt-cicd · 2026-04-27T06:07:30Z

PR_Github #45628 [ run ] completed with state SUCCESS. Commit: d253811
/LLM/main/L0_MergeRequest_PR pipeline #35841 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

Link to invocation

Implement FlashInfer-based Multi-head Latent Attention (MLA) for DeepSeek-style models, enabling use of BatchPrefillWithRaggedKVCacheWrapper for context and BatchMLAPagedAttentionWrapper for generation. Key design decisions: - KV cache: zero-copy ckv/kpe views split from the existing paged KV pool buffer (kv_lora_rank + qk_rope_head_dim), avoiding extra allocation - Context phase: ragged prefill with expanded Q/K/V after appending latent to paged MLA caches via append_paged_mla_kv_cache - Generation phase: paged MLA decode with q_nope/q_pe split from fused_q; latent cache append handled before each decode step - RoPE applied externally in MLA.forward before latent_cache construction; FlashInfer kernels receive pre-RoPE'd inputs - Plans called unconditionally each forward pass (no stale caching) - backend=auto on BatchMLAPagedAttentionWrapper for FA3 on Hopper - mla_rope_generation stub copies RoPE'd q_pe into fused_q to satisfy the forward_absorption_generation call chain Also fix metadata type assertion in attention.py to allow FlashInferAttentionMetadata for MLA, and replace hasattr duck-typing with isinstance check for kv_lens_cuda_runtime. Validated with 84/84 unit tests and end-to-end DeepSeek-V3-Lite NVFP4 inference with --attention_backend FLASHINFER. Support CUDA graph for flashinfer MLA. Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com> Drop redundant contiguous/reshape copies in flashinfer MLA path. Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com> [None][test] Fix FlashInfer MLA unit test and add B200/B300 smoke coverage - test_attention_mla_flashinfer: restrict to SM100 and set max_num_tokens to sum(context_sequence_lengths) so FlashInfer's batch_indices/positions buffers are sized for the full batch (was max(...), causing OOB copy). - Add TestDeepSeekV3Lite::test_bfloat16_flashinfer accuracy smoke test and register it in l0_b200.yml and l0_b300.yml. Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Tracin · 2026-05-12T03:15:44Z

/bot run

tensorrt-cicd · 2026-05-12T03:24:26Z

PR_Github #47852 [ run ] triggered by Bot. Commit: dee75ec Link to invocation

tensorrt-cicd · 2026-05-12T08:26:36Z

PR_Github #47852 [ run ] completed with state SUCCESS. Commit: dee75ec
/LLM/main/L0_MergeRequest_PR pipeline #37723 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Tracin · 2026-05-12T09:49:58Z

/bot run

tensorrt-cicd · 2026-05-12T09:55:23Z

PR_Github #47950 [ run ] triggered by Bot. Commit: a00d56b Link to invocation

tensorrt-cicd · 2026-05-12T14:53:40Z

PR_Github #47950 [ run ] completed with state SUCCESS. Commit: a00d56b
/LLM/main/L0_MergeRequest_PR pipeline #37792 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Tracin · 2026-05-13T02:18:10Z

/bot run

tensorrt-cicd · 2026-05-13T02:23:38Z

PR_Github #48068 [ run ] triggered by Bot. Commit: 92bda1c Link to invocation

tensorrt-cicd · 2026-05-13T06:08:55Z

PR_Github #48068 [ run ] completed with state SUCCESS. Commit: 92bda1c
/LLM/main/L0_MergeRequest_PR pipeline #37901 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Tracin · 2026-05-20T05:01:32Z

Hi @yuxianq, new comments are also addressed. Please take a look. Thanks!

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Tracin · 2026-05-20T05:06:09Z

/bot run

tensorrt-cicd · 2026-05-20T05:12:23Z

PR_Github #49340 [ run ] triggered by Bot. Commit: 0d71ba3 Link to invocation

tensorrt-cicd · 2026-05-20T09:36:44Z

PR_Github #49340 [ run ] completed with state SUCCESS. Commit: 0d71ba3
/LLM/main/L0_MergeRequest_PR pipeline #38997 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

Tracin · 2026-05-22T01:45:08Z

/bot run

Signed-off-by: Qi Zhang (qizh) <10434017+Tracin@users.noreply.github.com>

Tracin · 2026-05-22T14:02:22Z

/bot run

tensorrt-cicd · 2026-05-22T14:08:54Z

PR_Github #49949 [ run ] triggered by Bot. Commit: 491421e Link to invocation

tensorrt-cicd · 2026-05-22T18:16:53Z

PR_Github #49949 [ run ] completed with state SUCCESS. Commit: 491421e
/LLM/main/L0_MergeRequest_PR pipeline #39520 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

Tracin · 2026-05-23T11:50:33Z

/bot run

tensorrt-cicd · 2026-05-23T11:56:20Z

PR_Github #50048 [ run ] triggered by Bot. Commit: 491421e Link to invocation

tensorrt-cicd · 2026-05-23T16:37:24Z

PR_Github #50048 [ run ] completed with state SUCCESS. Commit: 491421e
/LLM/main/L0_MergeRequest_PR pipeline #39609 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

Tracin · 2026-05-24T02:38:20Z

/bot run

tensorrt-cicd · 2026-05-24T02:44:21Z

PR_Github #50073 [ run ] triggered by Bot. Commit: 491421e Link to invocation

tensorrt-cicd · 2026-05-24T06:52:31Z

PR_Github #50073 [ run ] completed with state SUCCESS. Commit: 491421e
/LLM/main/L0_MergeRequest_PR pipeline #39629 completed with status: 'SUCCESS'

CI Report

Link to invocation

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Tracin requested review from a team as code owners April 24, 2026 10:06

Tracin requested a review from QiJune April 24, 2026 10:06

github-actions Bot assigned Tracin Apr 24, 2026

coderabbitai Bot reviewed Apr 24, 2026

View reviewed changes

StanleySun639 approved these changes Apr 27, 2026

View reviewed changes

Tracin added 2 commits May 11, 2026 20:09

Rebase branch.

dee75ec

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Tracin force-pushed the flashinfer_mla branch from d253811 to dee75ec Compare May 12, 2026 03:14

Tracin requested a review from a team as a code owner May 12, 2026 03:14

[None][feat] Support FlashInfer MLA chunked prefill

a00d56b

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Add DeepSeek V3 Lite FlashInfer QA tests

e8e6bb9

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Tracin requested a review from a team as a code owner May 13, 2026 02:17

Merge branch 'main' into flashinfer_mla

92bda1c

yuxianq reviewed May 20, 2026

View reviewed changes

Comment thread tensorrt_llm/_torch/attention_backend/flashinfer.py

yuxianq reviewed May 20, 2026

View reviewed changes

Comment thread tensorrt_llm/_torch/attention_backend/flashinfer.py

yuxianq reviewed May 20, 2026

View reviewed changes

Comment thread tensorrt_llm/_torch/attention_backend/flashinfer.py Outdated

Tracin added 4 commits May 19, 2026 20:07

Route FlashInfer MLA normal prefill through default path

033420f

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Fix FlashInfer MLA cached context dispatch

bcc19e5

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Address FlashInfer MLA review comments

7dcf29a

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Route first MLA context chunk through normal prefill

1545fbc

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Fix FlashInfer MLA error formatting

0d71ba3

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Merge branch 'main' into flashinfer_mla

491421e

Signed-off-by: Qi Zhang (qizh) <10434017+Tracin@users.noreply.github.com>

yuxianq approved these changes May 25, 2026

View reviewed changes

Tracin merged commit e45a8e3 into NVIDIA:main May 25, 2026
7 checks passed

KleinBlueC pushed a commit to KleinBlueC/TensorRT-LLM that referenced this pull request May 26, 2026

[None][feat] Add FlashInfer MLA attention backend support (NVIDIA#13428)

42ecf33

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

coderabbitai Bot mentioned this pull request May 27, 2026

[TRTLLM-12982][perf] remove sync after FlashInfer attention plan() #14634

Merged

1 task

bmarimuthu-nv pushed a commit to nv-auto-deploy/TensorRT-LLM that referenced this pull request May 28, 2026

[None][feat] Add FlashInfer MLA attention backend support (NVIDIA#13428)

18b5325

Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>

Tracin deleted the flashinfer_mla branch June 9, 2026 09:56

coderabbitai Bot mentioned this pull request Aug 1, 2026

feat(collector): upgrade the TRT-LLM collector to 1.3.0rc20 with full-sweep perf data for 8 systems ai-dynamo/aiconfigurator#1356

Open

Uh oh!

Conversation

Tracin commented Apr 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary by CodeRabbit

Release Notes

Description

Test Coverage

PR Checklist

GitHub Bot Help

Uh oh!

Tracin commented Apr 24, 2026

Uh oh!

tensorrt-cicd commented Apr 24, 2026

Uh oh!

coderabbitai Bot commented Apr 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Changes

Sequence Diagram

Estimated code review effort

❌ Failed checks (1 warning)

Uh oh!

coderabbitai Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tensorrt-cicd commented Apr 24, 2026

Uh oh!

Tracin commented Apr 27, 2026

Uh oh!

tensorrt-cicd commented Apr 27, 2026

Uh oh!

tensorrt-cicd commented Apr 27, 2026

Uh oh!

Tracin commented May 12, 2026

Uh oh!

tensorrt-cicd commented May 12, 2026

Uh oh!

tensorrt-cicd commented May 12, 2026

Uh oh!

Tracin commented May 12, 2026

Uh oh!

tensorrt-cicd commented May 12, 2026

Uh oh!

tensorrt-cicd commented May 12, 2026

Uh oh!

Tracin commented May 13, 2026

Uh oh!

tensorrt-cicd commented May 13, 2026

Uh oh!

tensorrt-cicd commented May 13, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Tracin commented May 20, 2026

Uh oh!

Tracin commented May 20, 2026

Uh oh!

tensorrt-cicd commented May 20, 2026

Uh oh!

tensorrt-cicd commented May 20, 2026

Uh oh!

Tracin commented May 22, 2026

Uh oh!

Tracin commented May 22, 2026

Uh oh!

tensorrt-cicd commented May 22, 2026

Uh oh!

tensorrt-cicd commented May 22, 2026

Uh oh!

Tracin commented May 23, 2026

Uh oh!

tensorrt-cicd commented May 23, 2026

Uh oh!

Tracin commented Apr 24, 2026 •

edited

Loading

coderabbitai Bot commented Apr 24, 2026 •

edited

Loading