openvino
[NPUW] Support prefill-chunk for text-embedding model
#33076

Merged

[NPUW] Support prefill-chunk for text-embedding model #33076

AlexanderKalistratov merged 16 commits into openvinotoolkit:master from mengweiguo:qwen3-embedding-pr

github-actions added category: NPU

github-actions added category: NPUW

sys-openvino-ci added ExternalIntelPR

mengweiguo force pushed 199 days ago

mengweiguo changed the title ~~Support prefill-chunk for text-embedding model~~ NPUW] Support prefill-chunk for text-embedding model 198 days ago

mengweiguo changed the title ~~NPUW] Support prefill-chunk for text-embedding model~~ [NPUW] Support prefill-chunk for text-embedding model 198 days ago

mengweiguo marked this pull request as ready for review 198 days ago

mengweiguo requested a review 198 days ago

mengweiguo force pushed 196 days ago

mengweiguo force pushed 194 days ago

mengweiguo force pushed 186 days ago

dmatveev commented on 2025-12-16

AlexanderKalistratov commented on 2025-12-05

mengweiguo force pushed 180 days ago

mengweiguo force pushed 179 days ago

mengweiguo requested a review from

dmatveev 179 days ago

mengweiguo requested a review from

AlexanderKalistratov 179 days ago

AlexanderKalistratov commented on 2025-12-23

mengweiguo force pushed 178 days ago

AlexanderKalistratov commented on 2025-12-24

Support prefill-chunk for text-embedding model

dc8bb51a

code cleanup

e3908be9

Add option `normalize` support

25761266

Adjust padding side

23a3f644

Move data to left side

cf78e19d

Fix CPU fallbakc issue.

a883d8a0

Remove post model and cache chunk output

92d3403e

Put Lora out

6ac98b50

Fix conflict

8d11919a

Rafactor compiled-model and infer-request

29b8ec77

Add `LLMInferBaseRequest` as base request for LLM and Embedding

fd1f7895

Update serialization version `0.16->0.17`

d1cc2c8a

Rebuild model pass and add model check

f9eab5de

Remove `input_token_ids`

c975df36

Code cleanup

a459661b

dmatveev added this to the 2026.0 milestone 168 days ago

dmatveev commented on 2026-01-02

Move `pad_position_ids` to `infer_request_utils.hpp`

9764a0b3

mengweiguo force pushed to 9764a0b3 166 days ago

mengweiguo requested a review from

dmatveev 162 days ago

mengweiguo requested a review from

AlexanderKalistratov 162 days ago

AlexanderKalistratov approved these changes on 2026-01-08

AlexanderKalistratov enabled auto-merge 162 days ago

AlexanderKalistratov merged e5c24ae9 into master 156 days ago

Reviewers

AlexanderKalistratov

dmatveev

Assignees

No one assigned

Labels

ExternalIntelPR category: NPU category: NPUW

Milestone

2026.0

openvino [NPUW] Support prefill-chunk for text-embedding model #33076 Merged

[NPUW] Support prefill-chunk for text-embedding model #33076

openvino
[NPUW] Support prefill-chunk for text-embedding model
#33076

Merged