-
Notifications
You must be signed in to change notification settings - Fork 165
Pull requests: openvinotoolkit/openvino.genai
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Preserve default config for LLM samples
category: samples
GenAI samples
category: sampling
Sampling / Decoding algorithms
Whisper pipeline: implement chunk streamer for long-form audio processing
category: GenAI C++ API
Changes in GenAI C++ public headers
category: Python API
Python API for GenAI
category: sampling
Sampling / Decoding algorithms
category: whisper
Whisper pipeline
[GHA] Split VLM job by models
category: GHA
CI based on Github actions
category: sampling
Sampling / Decoding algorithms
[Image generation] Non public API changes for image generation
category: continuous batching
Continuous batching
category: Python API
Python API for GenAI
category: samples
GenAI samples
category: sampling
Sampling / Decoding algorithms
category: text to image
Text 2 image pipeline
category: visual language
Visual language pipeline
#1146
opened Nov 5, 2024 by
ilya-lavrenov
Loading…
Return ref generation
category: cmake / build
Cmake scripts
category: GHA
CI based on Github actions
category: visual language
Visual language pipeline
Add GPU FP16 overflow fix
category: WWB
PR changes WWB
do_not_merge
#1120
opened Nov 1, 2024 by
AlexKoff88
•
Draft
Verify chatglm3 6b
category: GHA
CI based on Github actions
category: sampling
Sampling / Decoding algorithms
category: tokenizers
Tokenizer class or submodule update
#1119
opened Oct 31, 2024 by
Aniruddha521
Loading…
StaticLLMPipeline: Enable chat test
category: sampling
Sampling / Decoding algorithms
#1117
opened Oct 31, 2024 by
TolyaTalamanov
Loading…
StaticLLMPipeline: Use set_tensor for kvcache model
category: LLM
LLM pipeline (stateful, static)
#1106
opened Oct 30, 2024 by
TolyaTalamanov
•
Draft
text2image: Add pimp approach for unet_2d_condition & batch-size 1 implementation of unet for initial NPU support
category: GenAI C++ API
Changes in GenAI C++ public headers
category: text to image
Text 2 image pipeline
#1101
opened Oct 29, 2024 by
RyanMetcalfeInt8
Loading…
Corrected performance data when batch size is greater than 1
category: sampling
Sampling / Decoding algorithms
do_not_merge
Update onnx requirement from <=1.16.1 to <=1.17.0 in /tools/llm_bench
category: llm_bench
Label for tool/llm_bench folder
dependencies
Pull requests that update a dependency file
[WWB]: Add support for VisualLanguage models
category: sampling
Sampling / Decoding algorithms
category: WWB
PR changes WWB
#1086
opened Oct 28, 2024 by
AlexKoff88
•
Draft
Master optimum-intel for llava
category: GHA
CI based on Github actions
category: sampling
Sampling / Decoding algorithms
Update optimum-intel branch
category: GHA
CI based on Github actions
category: sampling
Sampling / Decoding algorithms
[CB] Speculative decoding benchmarking
category: LLM
LLM pipeline (stateful, static)
category: samples
GenAI samples
category: sampling
Sampling / Decoding algorithms
category: speculative decoding
Speculative decoding
Grouped source files in GenAI src folder
category: cmake / build
Cmake scripts
category: continuous batching
Continuous batching
category: GHA
CI based on Github actions
category: Python API
Python API for GenAI
category: sampling
Sampling / Decoding algorithms
category: speculative decoding
Speculative decoding
category: text to image
Text 2 image pipeline
category: visual language
Visual language pipeline
category: whisper
Whisper pipeline
#1016
opened Oct 18, 2024 by
ilya-lavrenov
Loading…
[test] Update Optimum to Use Symbols
do_not_merge
do_not_review
#1004
opened Oct 17, 2024 by
jane-intel
Loading…
Token rotation
category: cmake / build
Cmake scripts
category: continuous batching
Continuous batching
category: sampling
Sampling / Decoding algorithms
do_not_merge
do_not_review
Update chat_sample README.md
category: samples
GenAI samples
category: sampling
Sampling / Decoding algorithms
#978
opened Oct 15, 2024 by
raymondlo84
Loading…
Build GenAI via OpenVINO wheel
category: cmake / build
Cmake scripts
#958
opened Oct 14, 2024 by
ilya-lavrenov
•
Draft
fix the caculation of performance metric
category: llm_bench
Label for tool/llm_bench folder
do_not_merge
#940
opened Oct 10, 2024 by
QuPengfei
Loading…
Use continuous batching by default
category: continuous batching
Continuous batching
category: GHA
CI based on Github actions
category: LLM
LLM pipeline (stateful, static)
category: sampling
Sampling / Decoding algorithms
do_not_merge
do_not_review
#882
opened Sep 19, 2024 by
Wovchena
Loading…
Previous Next
ProTip!
Find all pull requests that aren't related to any open issues with -linked:issue.