Move parameter node creation from decoder to translate#252
Closed
zhaixuejun1993 wants to merge 39 commits into
Closed
Move parameter node creation from decoder to translate#252zhaixuejun1993 wants to merge 39 commits into
zhaixuejun1993 wants to merge 39 commits into
Conversation
…g_src to recorde the src ggml tensor for OpenVINO dynamic shape infer
…utput_info Clean node information in OpenVINO backend
…issue Enable zero-size copy for OpenVINO backend view
enable qwen35 Fix after rebase remove logging
…t reason: the backend test initializes unary op inputs over a wide range, [-150, 150]. For FP32, exp(x) overflows around x ~= 88.7, so this test can randomly generate values right in or beyond the overflow region
In stateful mode the NEOX RoPE branch fed rank-3 data ([S, n_heads, head_size]) into the Multiply against the rank-4 cos/sin tables ([1, S, 1, n_dims/2]). That mixed-rank broadcast is miscomputed by the OpenVINO GPU plugin, corrupting the rotated Q/K and producing garbage output (e.g. Phi-3-mini). Lift the data to rank-4 before the split/ Multiply so the operands are equal-rank, matching what the TYPE_NORMAL branch already does. CPU and stateless paths are unaffected. Phi-3-mini-Q4_K_M, wiki.test perplexity, GPU stateful: before: PPL = 27120.43 after: PPL = 6.2263 (CPU reference: 6.2251)
…ov name in ov bk; 3) fix issue in arch test & op test with latest code update
Update OpenVINO backend: remove unique name, add new OV name, fix tests
…s-in-llama OpenVINO Backenb: remove changes in llama.cpp
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
Doc change (use x64 Native Tools Command Prompt for VS)
Upgrade cache key to include all source names in OpenVINO backend
…h-test Enable Llama architecture tests in CI for OpenVINO backend
ravi9
force-pushed
the
dev_backend_openvino
branch
from
July 14, 2026 04:06
7a9b001 to
ab6cd7e
Compare
Merged
ravi9
added a commit
that referenced
this pull request
Jul 14, 2026
Owner
|
Merged PR#253 instead of this. |
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
5 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 pull request refactors how model inputs and extra inputs are represented and handled in the OpenVINO backend, replacing direct use of OpenVINO node objects with new structured info types. The changes improve abstraction, simplify input management, and make the codebase more maintainable and extensible. Key updates include the introduction of
ModelInputInfoandModelExtraInputInfostructs, updates to interfaces and internal data structures, and refactoring of related logic to use these new types.API and Data Structure Refactoring:
ModelInputInfoandModelExtraInputInfostructs to represent model inputs and extra inputs, respectively, encapsulating type, shape, value, and parameter status. (ggml/src/ggml-openvino/openvino/decoder.h)GgmlDecoderand its implementations to use these new structs instead ofstd::shared_ptr<ov::Node>. (ggml/src/ggml-openvino/openvino/decoder.h,ggml/src/ggml-openvino/ggml-decoder.h) [1] [2] [3]Logic and Implementation Updates:
GgmlOvDecoderto use the new structs, simplifying logic and removing redundant code. (ggml/src/ggml-openvino/ggml-decoder.cpp) [1] [2] [3]ggml/src/ggml-openvino/openvino/translate_session.cpp) [1] [2]Utility Function Update:
ModelExtraInputInfostruct, removing the need for a separate map of extra input values. (ggml/src/ggml-openvino/utils.cpp)Dependency Cleanup:
ggml/src/ggml-openvino/ggml-decoder.cpp)These changes collectively improve code clarity and set the stage for easier future modifications to input handling.## Overview
Additional information
Requirements