Enable VLM lookup. by xipingyan · Pull Request #2707 · openvinotoolkit/openvino.genai

xipingyan · 2025-09-05T00:55:28Z

1: Move m_pipeline->generate_candidates(); fromContinuousBatchingPipeline::PromptLookupImpl::step() to ContinuousBatchingPipeline::ContinuousBatchingImpl::step()
2: Encode original prompt to org_prompt_ids as lookup table.

Tickets: CVS-172889, CVS-176586

Copilot

Pull Request Overview

This PR enables VLM (Vision Language Model) lookup functionality by enhancing the continuous batching pipeline to support prompt lookup for embedding-based models. The changes introduce a new interface for providing prompt token IDs when using embedding inputs, enabling better integration between visual language models and prompt lookup decoding.

Move candidate generation from PromptLookupImpl::step() to ContinuousBatchingImpl::step()
Add new interface to pass prompt token IDs for embedding input models through get_inputs_embeds_with_token_type_ids method
Update all VLM model implementations to support the new token_type_ids interface

Reviewed Changes

Copilot reviewed 25 out of 25 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
src/cpp/src/visual_language/qwen2vl/classes.hpp	Add get_inputs_embeds_with_token_type_ids method declaration
src/cpp/src/visual_language/qwen2vl/classes.cpp	Implement new method and refactor existing get_inputs_embeds to use it
src/cpp/src/visual_language/phi4mm/classes.hpp	Add get_inputs_embeds_with_token_type_ids method declaration
src/cpp/src/visual_language/phi4mm/classes.cpp	Implement new method and refactor existing get_inputs_embeds to use it
src/cpp/src/visual_language/phi3_vision/classes.hpp	Add get_inputs_embeds_with_token_type_ids method declaration
src/cpp/src/visual_language/phi3_vision/classes.cpp	Implement new method and refactor existing get_inputs_embeds to use it
src/cpp/src/visual_language/minicpm/classes.hpp	Add get_inputs_embeds_with_token_type_ids method declaration
src/cpp/src/visual_language/minicpm/classes.cpp	Implement new method and refactor existing get_inputs_embeds to use it
src/cpp/src/visual_language/llava_next/classes.hpp	Add get_inputs_embeds_with_token_type_ids method declaration
src/cpp/src/visual_language/llava_next/classes.cpp	Implement new method and refactor existing get_inputs_embeds to use it
src/cpp/src/visual_language/llava/classes.hpp	Add get_inputs_embeds_with_token_type_ids method declaration
src/cpp/src/visual_language/llava/classes.cpp	Implement new method and refactor existing get_inputs_embeds to use it
src/cpp/src/visual_language/internvl_chat/classes.hpp	Add get_inputs_embeds_with_token_type_ids method declaration
src/cpp/src/visual_language/internvl_chat/classes.cpp	Implement new method and refactor existing get_inputs_embeds to use it
src/cpp/src/visual_language/inputs_embedder.hpp	Update constructors to accept prompt_lookup parameter and add prompt_lookup support
src/cpp/src/visual_language/inputs_embedder.cpp	Update has_token_type_ids method and constructors for prompt lookup support
src/cpp/src/sequence_group.hpp	Add handling for embeddings in remove_last_tokens method
src/cpp/src/prompt_lookup/prompt_lookup_impl.hpp	Add constructor for embedding-based models
src/cpp/src/prompt_lookup/prompt_lookup_impl.cpp	Support token_type_ids and remove candidate generation from step method
src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.hpp	Add constructor for embedding models and make generate_candidates virtual
src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp	Fix loop variable type and add candidate padding logic
src/cpp/src/continuous_batching/pipeline_impl.hpp	Add virtual generate_candidates method declaration
src/cpp/src/continuous_batching/pipeline_impl.cpp	Move candidate generation to main pipeline step and add empty default implementation
src/cpp/src/continuous_batching/pipeline.cpp	Pass prompt_lookup flag to InputsEmbedder constructor
src/cpp/src/continuous_batching/model_runner.hpp	Add proper token_type_ids tensor existence check

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp

xipingyan · 2025-09-05T13:36:29Z

@yangsu2022 , could you please help review firstly?

1: global variable pass prompts_ids. Signed-off-by: xipingya <xiping.yan@intel.com>

Don't need to add new interface, just reuse "token_type_ids". Signed-off-by: xipingya <xiping.yan@intel.com>

Signed-off-by: xipingya <xiping.yan@intel.com>

2: fix match bug, for example: input_ids={2, 3, 1, 1, 2, 3, 4, 5, 6, 9, 2, 3, 1, 2, 3} num_pred_tokens=3 max_ngram_size=3 return candidate: 2,3,1 Signed-off-by: xipingya <xiping.yan@intel.com>

Signed-off-by: xipingya <xiping.yan@intel.com>

Copilot

Pull Request Overview

Copilot reviewed 25 out of 25 changed files in this pull request and generated 3 comments.

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp

src/cpp/src/continuous_batching/model_runner.hpp

Signed-off-by: xipingya <xiping.yan@intel.com>

Copilot

Pull Request Overview

Copilot reviewed 25 out of 25 changed files in this pull request and generated 4 comments.

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp

src/cpp/src/continuous_batching/model_runner.hpp

Signed-off-by: xipingya <xiping.yan@intel.com>

Copilot

Pull Request Overview

Copilot reviewed 25 out of 25 changed files in this pull request and generated 1 comment.

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp

Signed-off-by: xipingya <xiping.yan@intel.com>

Copilot

Pull request overview

Copilot reviewed 25 out of 25 changed files in this pull request and generated 10 comments.

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.hpp

src/cpp/src/continuous_batching/pipeline_base.cpp

samples/python/visual_language_chat/visual_language_chat.py

src/cpp/src/visual_language/inputs_embedder.hpp

tests/python_tests/test_vlm_pipeline.py

samples/python/visual_language_chat/visual_language_chat.py

Copilot

Pull request overview

Copilot reviewed 25 out of 25 changed files in this pull request and generated 4 comments.

Comments suppressed due to low confidence (1)

src/cpp/src/prompt_lookup/prompt_lookup_impl.cpp:50

After moving candidate generation into ContinuousBatchingImpl::step(), PromptLookupImpl::step() no longer updates m_sd_metrics.draft_duration, so draft vs main timing breakdown in SpeculativeDecodingMetrics becomes misleading (draft stays 0 and candidate-generation time is effectively attributed to main). Consider reintroducing draft timing around generate_candidates_for_prompt_lookup() for the prompt-lookup pipeline, or updating the metrics naming/logic accordingly.

    ManualTimer main_timer("prompt_lookup_decoding: pipeline: step()");
    main_timer.start();
    m_pipeline->step();
    main_timer.end();
    m_sd_metrics.main_duration += main_timer.get_duration();
    m_pipeline_metrics = m_pipeline->get_metrics();

tests/python_tests/test_vlm_pipeline.py

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.hpp

src/cpp/src/sequence_group.hpp

Copilot · 2026-02-13T02:09:28Z

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp

+            for (const auto& candidate : candidates) {
+                running_sequence->append_token(candidate, 0);
+            }
+            max_validation_len = std::max(max_validation_len, candidates.size());
        }
        request->set_num_validated_tokens(max_validation_len);
    }


max_validation_len is computed after padding candidates up to sampling_params.num_assistant_tokens, and then stored via set_num_validated_tokens(). This makes the pipeline try to validate padded tokens as if they were real candidates (and can also exceed min_num_assistant_tokens near max_new_tokens). Validation length should reflect only the number of real candidate tokens generated, not the padded shape length.

Copilot

Pull request overview

Copilot reviewed 25 out of 25 changed files in this pull request and generated 5 comments.

Comments suppressed due to low confidence (1)

src/cpp/src/prompt_lookup/prompt_lookup_impl.cpp:46

After moving candidate generation out of PromptLookupImpl::step(), the draft/candidate generation duration is no longer accounted for in prompt-lookup-specific metrics (previously accumulated into m_sd_metrics.draft_duration). If these metrics are consumed externally, this is a regression. Consider timing generate_candidates_for_prompt_lookup() in the prompt-lookup path and adding it back into the same metrics field (e.g., via an override that measures and accumulates duration, or by returning duration from the hook).

    ManualTimer step_timer("prompt_lookup_decoding: step()");
    step_timer.start();

    auto generated_len_before = m_pipeline->get_generated_request_len();

    ManualTimer main_timer("prompt_lookup_decoding: pipeline: step()");
    main_timer.start();

src/cpp/src/utils.cpp

tests/python_tests/test_vlm_pipeline.py

Copilot

Pull request overview

Copilot reviewed 25 out of 25 changed files in this pull request and generated no new comments.

…penvino.genai into xp/enable_vlm_lookup

Copilot

Pull request overview

Copilot reviewed 26 out of 26 changed files in this pull request and generated 1 comment.

Comments suppressed due to low confidence (3)

src/cpp/src/continuous_batching/pipeline_impl.cpp:529

ContinuousBatchingImpl::generate(...) now accepts prompt_ids, but the parameter is never used (and is not forwarded into add_request(...)). This can cause -Wunused-parameter build failures and prevents prompt IDs from reaching SequenceGroup when using this implementation. Either propagate prompt_ids per request (with the same bounds checks used for token_type_ids) or explicitly mark the parameter unused if it is intentionally unsupported here.

std::vector<EncodedGenerationResult>
ContinuousBatchingPipeline::ContinuousBatchingImpl::generate(const std::vector<ov::Tensor>& input_ids,
                                                             const std::vector<GenerationConfig>& sampling_params,
                                                             const StreamerVariant& streamer,
                                                             const std::optional<std::vector<ov::Tensor>>& token_type_ids,
                                                             const std::optional<std::vector<std::pair<ov::Tensor, std::optional<int64_t>>>>& position_ids_list,
                                                             const std::optional<std::vector<ov::Tensor>>& prompt_ids) {

    _reset_cache_usage_statistics();
    ManualTimer generate_timer("generate()");
    generate_timer.start();

    OPENVINO_ASSERT(!has_non_finished_requests(), "Generate cannot be called while ContinuousBatchingPipeline is already in running state. Use ContinuousBatchingPipeline::add_request");
    OPENVINO_ASSERT(input_ids.size() == sampling_params.size());

    if (position_ids_list.has_value()) {
        OPENVINO_ASSERT((*position_ids_list).size() == input_ids.size());
    }

    auto start_time =  std::chrono::steady_clock::now();
    PerfMetrics perf_metrics;
    auto& raw_perf_counters = perf_metrics.raw_metrics;
    raw_perf_counters.m_inference_durations =  {{ MicroSeconds(0.0f) }};

    // checks that all requests has the same LoRA adapters property value
    for (size_t i = 1; i < sampling_params.size(); ++i) {
        OPENVINO_ASSERT(sampling_params[i - 1].adapters == sampling_params[i].adapters,
            "LoRA adapters value must be the same for all requests");
    }
    set_adapters(sampling_params[0].adapters);

    const auto streamer_ptr = std::make_shared<ThreadedStreamerWrapper>(streamer, m_tokenizer);

    OPENVINO_ASSERT(!streamer_ptr->has_callback() || input_ids.size() == 1 && sampling_params[0].num_return_sequences == 1 &&
        (sampling_params[0].is_greedy_decoding() || sampling_params[0].is_multinomial()),
        "Currently streaming is possible only with batch size=1 and only for greedy or multinomial decoding");

    std::vector<GenerationHandle> generations;
    for (size_t request_id = 0; request_id < input_ids.size(); ++request_id) {
        OPENVINO_ASSERT(1 == input_ids[request_id].get_shape().at(0), "Use multiple tensors to pass a batch.");
        if (position_ids_list.has_value()) {
            const auto [position_ids, rope_delta] = (*position_ids_list)[request_id];
            m_inputs_embedder->set_position_ids(position_ids);
            if (rope_delta.has_value()) {
                m_inputs_embedder->set_rope_delta(*rope_delta);
            }
        }
        bool has_valid_token = token_type_ids.has_value() && request_id < token_type_ids->size();
        generations.push_back(
            add_request(request_id, input_ids[request_id], sampling_params[request_id], has_valid_token ? std::make_optional((*token_type_ids)[request_id]) : std::nullopt)
        );

src/cpp/src/prompt_lookup/prompt_lookup_impl.cpp:50

PromptLookupImpl::step() no longer measures/updates m_sd_metrics.draft_duration, so prompt-lookup candidate generation time is now effectively counted as main_duration (or not tracked separately at all). If consumers rely on SpeculativeDecodingMetrics fields (e.g., debug printing/percentages), this will misreport performance. Consider timing generate_candidates_for_prompt_lookup() separately again (at its new location) and updating draft_duration accordingly.

void ContinuousBatchingPipeline::PromptLookupImpl::step() {
    auto& raw_perf_counters = m_perf_metrics.raw_metrics;

    ManualTimer step_timer("prompt_lookup_decoding: step()");
    step_timer.start();

    auto generated_len_before = m_pipeline->get_generated_request_len();

    ManualTimer main_timer("prompt_lookup_decoding: pipeline: step()");
    main_timer.start();
    m_pipeline->step();
    main_timer.end();
    m_sd_metrics.main_duration += main_timer.get_duration();
    m_pipeline_metrics = m_pipeline->get_metrics();

src/cpp/src/utils.cpp:139

is_paged_attention_available() now returns true on ARM64 only when HAVE_SVE is defined, but the exception messages in explicitly_requires_paged_attention() still state that support is available on “ARM64 platforms” in general. To avoid misleading users, please update the associated error messages (or add a brief comment here) to reflect the new ARM64+SVE requirement.

inline bool is_paged_attention_available() {
#if defined(OPENVINO_ARCH_X86_64) || (defined(OPENVINO_ARCH_ARM64) && defined(HAVE_SVE))
    return true;
#else
    return false;
#endif

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp

Copilot

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

…ForPromptLookupImpl()

…penvino.genai into xp/enable_vlm_lookup

Copilot

Pull request overview

Copilot reviewed 26 out of 26 changed files in this pull request and generated 2 comments.

Comments suppressed due to low confidence (2)

src/cpp/src/utils.cpp:139

is_paged_attention_available() now returns true on ARM64 only when HAVE_SVE is defined. However, the exception messages in explicitly_requires_paged_attention() still say PagedAttention is available on "x86_64 or ARM64". That becomes misleading on ARM64 builds without SVE.

Suggestion: update the related error messages to reflect the new requirement (e.g., ARM64 with SVE), so users get actionable guidance.

inline bool is_paged_attention_available() {
#if defined(OPENVINO_ARCH_X86_64) || (defined(OPENVINO_ARCH_ARM64) && defined(HAVE_SVE))
    return true;
#else
    return false;
#endif

src/cpp/src/prompt_lookup/prompt_lookup_impl.cpp:50

After moving candidate generation into ContinuousBatchingImpl::step(), PromptLookupImpl::step() no longer accounts for candidate generation time in m_sd_metrics.draft_duration. As a result, SpeculativeDecodingMetrics for prompt-lookup will report draft_duration == 0 and shift all time into main_duration, making the printed/queried metrics incorrect.

Suggestion: plumb candidate-generation duration back (e.g., store it in m_pipeline_metrics or return it from generate_candidates_for_prompt_lookup()), and update m_sd_metrics.draft_duration in PromptLookupImpl::step() similarly to the previous implementation.

    ManualTimer main_timer("prompt_lookup_decoding: pipeline: step()");
    main_timer.start();
    m_pipeline->step();
    main_timer.end();
    m_sd_metrics.main_duration += main_timer.get_duration();
    m_pipeline_metrics = m_pipeline->get_metrics();

src/cpp/src/continuous_batching/pipeline_impl.cpp

samples/cpp/visual_language_chat/visual_language_chat.cpp

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 26 out of 26 changed files in this pull request and generated 1 comment.

Comments suppressed due to low confidence (1)

src/cpp/src/utils.cpp:139

is_paged_attention_available() now gates ARM64 support on HAVE_SVE, but this macro is not defined anywhere in this repo and is not a standard compiler macro (the common one is __ARM_FEATURE_SVE). If HAVE_SVE isn’t injected by the build, this change will silently disable ContinuousBatching/PromptLookup on all ARM64 builds and start throwing at runtime. Consider using a standard feature macro (or a repo-defined CMake option) and/or add a build-time definition to guarantee consistent behavior.

inline bool is_paged_attention_available() {
#if defined(OPENVINO_ARCH_X86_64) || (defined(OPENVINO_ARCH_ARM64) && defined(HAVE_SVE))
    return true;
#else
    return false;
#endif

src/cpp/src/continuous_batching/pipeline_base.cpp

github-actions bot added category: visual language Visual language pipeline category: continuous batching Continuous batching no-match-files category: prompt lookup Prompt look-up decoding labels Sep 5, 2025

xipingyan added do_not_merge do_not_review and removed do_not_merge do_not_review labels Sep 5, 2025

xipingyan marked this pull request as ready for review September 5, 2025 13:29

Copilot AI review requested due to automatic review settings September 5, 2025 13:29

Copilot AI reviewed Sep 5, 2025

View reviewed changes

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp Show resolved Hide resolved

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp Show resolved Hide resolved

xipingyan requested review from peterchen-intel, wangleis and yangsu2022 September 5, 2025 13:35

xipingyan added 3 commits September 11, 2025 10:12

Draft enable VLM lookup.

db9da07

1: global variable pass prompts_ids. Signed-off-by: xipingya <xiping.yan@intel.com>

Remove global variable pass prompt ids.

4a8901c

Don't need to add new interface, just reuse "token_type_ids". Signed-off-by: xipingya <xiping.yan@intel.com>

Update some comments.

bbb9de3

Signed-off-by: xipingya <xiping.yan@intel.com>

xipingyan force-pushed the xp/enable_vlm_lookup branch from 4c2fe5b to bbb9de3 Compare September 11, 2025 02:12

xipingyan added 2 commits September 11, 2025 12:39

1: fix potential issue: max_ngram_size < input_length;

eec1fa7

2: fix match bug, for example: input_ids={2, 3, 1, 1, 2, 3, 4, 5, 6, 9, 2, 3, 1, 2, 3} num_pred_tokens=3 max_ngram_size=3 return candidate: 2,3,1 Signed-off-by: xipingya <xiping.yan@intel.com>

avoiding potential signed/unsigned comparison issues

e96d490

Signed-off-by: xipingya <xiping.yan@intel.com>

xipingyan requested a review from Copilot September 11, 2025 05:38

Copilot AI reviewed Sep 11, 2025

View reviewed changes

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp Outdated Show resolved Hide resolved

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp Outdated Show resolved Hide resolved

src/cpp/src/continuous_batching/model_runner.hpp Outdated Show resolved Hide resolved

move to loop before.

3a7b3a5

Signed-off-by: xipingya <xiping.yan@intel.com>

xipingyan requested a review from Copilot September 11, 2025 06:43

Copilot AI reviewed Sep 11, 2025

View reviewed changes

don't update param variable.

f2fc501

Signed-off-by: xipingya <xiping.yan@intel.com>

xipingyan requested a review from Copilot September 11, 2025 06:56

Copilot AI reviewed Sep 11, 2025

View reviewed changes

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp Outdated Show resolved Hide resolved

static_cast for type convert.

5ee3df6

Signed-off-by: xipingya <xiping.yan@intel.com>

remove args.enable_lookup

8479454

Copilot AI reviewed Feb 11, 2026

View reviewed changes

Wovchena requested changes Feb 11, 2026

View reviewed changes

samples/python/visual_language_chat/visual_language_chat.py Outdated Show resolved Hide resolved

simplify assignment

99b125b

Wovchena mentioned this pull request Feb 12, 2026

Extend copilot-instructions.md #3332

Merged

1 task

rename parameter org_prompt in encode_prompt

59d5b57

Copilot AI reviewed Feb 13, 2026

View reviewed changes

sunxiaoxia2022 added 2 commits February 13, 2026 10:21

clang

59ce498

pass prompt_lookup into VLMPipeline() in test

2d25b5b

Copilot AI reviewed Feb 13, 2026

View reviewed changes

sunxiaoxia2022 added 2 commits February 13, 2026 11:41

add limitation with PROMPT_LOOKUP in test case

2872b9e

clang

5796a27

Copilot AI reviewed Feb 13, 2026

View reviewed changes

Wovchena approved these changes Feb 16, 2026

View reviewed changes

Wovchena and others added 3 commits February 20, 2026 13:46

Merge branch 'master' into xp/enable_vlm_lookup

33765a2

change prompt_lookup to parameter in sample

e094876

Merge branch 'xp/enable_vlm_lookup' of https://github.com/xipingyan/o…

b871180

…penvino.genai into xp/enable_vlm_lookup

Copilot AI reviewed Feb 25, 2026

View reviewed changes

src/cpp/src/prompt_lookup/continuous_batching_for_prompt_lookup.cpp Outdated Show resolved Hide resolved

sunxiaoxia2022 and others added 4 commits February 25, 2026 20:44

lint issue

3ae64d0

rm unused include

eb787a0

Merge branch 'master' into xp/enable_vlm_lookup

24bc943

Merge branch 'master' into xp/enable_vlm_lookup

5bd9a8a

Copilot AI reviewed Feb 27, 2026

View reviewed changes

sunxiaoxia2022 and others added 3 commits February 27, 2026 16:47

rm input parameter is_validation_mode_enabled from ContinuousBatching…

3778f5a

…ForPromptLookupImpl()

Merge branch 'xp/enable_vlm_lookup' of https://github.com/xipingyan/o…

0077763

…penvino.genai into xp/enable_vlm_lookup

Merge branch 'master' into xp/enable_vlm_lookup

a694712

Copilot AI reviewed Mar 2, 2026

View reviewed changes

src/cpp/src/continuous_batching/pipeline_impl.cpp Show resolved Hide resolved

samples/cpp/visual_language_chat/visual_language_chat.cpp Outdated Show resolved Hide resolved

DEVICE and PROMPT_LOOKUP are optional

0f55024

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Copilot AI reviewed Mar 3, 2026

View reviewed changes

src/cpp/src/continuous_batching/pipeline_base.cpp Outdated Show resolved Hide resolved

change org_prompt_ids_list to original_prompt_ids_list

dfdc600

Conversation

xipingyan commented Sep 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Uh oh!

Uh oh!

xipingyan commented Sep 5, 2025

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI Feb 13, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

xipingyan commented Sep 5, 2025 •

edited

Loading