vllm-omni good first issues (Python)

Align and expand LTX support about 6 hours ago

help wanted

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC] Streaming Video Input for Omni-Modal Real-Time Interaction about 6 hours ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Token-level streaming text input for TTS: commitment policy, capacity-aware segmentation, and speech-state inheritance about 18 hours ago

help wanted good first issue tts RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Duplex Request Metrics 1 day ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Good First Issue]: Add vendor-organized community recipes for “run model X on hardware Y for task Z” in vLLM-Omni 1 day ago

help wanted good first issue high priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Slim L1 CPU coverage for diffusion comm primitives (test_comm.py) 1 day ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Cut L3 H100 entrypoint cold starts (Qwen realtime, BAGEL llm_engine) and stop 60s 5% VRAM waits 1 day ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Align vLLM-Omni error handling with vLLM 2 days ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Roadmap] MiniMax-H3 follow-up: accuracy/performance CI, quantization, and optimizations 3 days ago

help wanted ROCm NPU Hardware Plugin diffusion xpu quantization CI/CD Kernel optimization

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[New Model]: Add microsoft/Mage-VL with end-to-end full-duplex multimodal streaming inference 3 days ago

help wanted new model

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Finish the config refactor 3 days ago

help wanted medium priority RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Help Wanted][Examples] Ratchet Python examples toward shared task runners 4 days ago

documentation help wanted good first issue diffusion tts CI/CD refactor

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Feature]: MiniMax-H3 sliding-window generation for videos longer than 15 s (e.g. 30 s) 4 days ago

help wanted good first issue

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Performance][Diffusion] Boogu-Image roadmap: full model-series support, feature optimizations, and hardware-specific optimizations 4 days ago

help wanted good first issue diffusion

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug][A100][H100]: minicpm 4.5, tests/e2e/online_serving/test_minicpmo_4_5_duplex_expansion.py:90: AssertionError True is Flase 4 days ago

bug help wanted high priority ci-failure

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Support Scalable Frontend: Implementing --api-server-count for Multimodal Throughput 4 days ago

help wanted

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: vLLM-Omni NPU 2026 Q2 Roadmap 4 days ago

help wanted good first issue NPU high priority Hardware Plugin

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug][NPU] Minimax-H3 INT8 online quantization on Ascend 910B2, err: Tensor matB not implemented for DT_INT8, should be in dtype support list [DT_FLOAT,DT_FLOAT16,DT_BFLOAT16,] 4 days ago

bug help wanted NPU low priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: staged MiniMax-H3 SVDQuant support 5 days ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Nightly / CI failed - Nemotron VoiceChat native duplex times out waiting for model output 5 days ago

bug help wanted high priority ci-failure

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Performance][VoxCPM2]: TTFA starts queueing at 2 RPS on H100 5 days ago

enhancement help wanted high priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Nightly / CI failed - MiniCPM-o 4.5 image-to-text-audio async_chunk content mismatch 5 days ago

bug help wanted high priority ci-failure

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: NPU: stopping torch profiler mid-RECORD causes Segfault and crashes Stage-0 EngineCore 6 days ago

bug help wanted NPU low priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Nightly CI, openbmb/VoxCPM2, performance metrics regressed by more than 10% compared to the baseline in some scenarios 7 days ago

bug help wanted ci-failure low priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Dynamic Task-DiT Residency for Unified MiniMax-H3 Serving 7 days ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Nightly / CI failed - Qwen3-Omni ModelOpt NVFP4 stage 1 EngineCore initialization failed 7 days ago

bug help wanted high priority ci-failure

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Nightly CI, Qwen/Qwen3-Omni-30B-A3B-Instruct, performance metrics regressed by more than 10% compared to the baseline in some scenarios 7 days ago

bug help wanted ci-failure medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: there is a ERROR "RuntimeError: The expanded size of the tensor (74) must match the existing size (0) at non-singleton dimension 1. Target sizes: [3, 74]. Tensor sizes: [3, 0]", when call a realtime audio req to qwen3-omni server 7 days ago

bug help wanted NPU medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: MiniMax-H3 modulation Triton kernels crash on Ascend NPU due to coreDim exceeding 65535 limit 7 days ago

bug help wanted NPU medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Nightly CI, MiniMaxAI/MiniMax-H3, performance metrics regressed by more than 10% compared to the baseline in some scenarios 9 days ago

bug help wanted medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Feature]: Auto-syncing example/*/*/README.md to docs/user_guide/examples/*/*.md 10 days ago

help wanted good first issue

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[New Model]: [Meta] Tuna series 10 days ago

help wanted good first issue new model

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[New Model] F5-TTS (flow-matching DiT-based TTS) 10 days ago

help wanted

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Model-declared readiness warmup for diffusion pipelines 10 days ago

help wanted high priority RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[New Model]: Audio8/Audio8-TTS-Preview-0.1b (FalconH1 hybrid slow AR + DualAR fast branch) 10 days ago

help wanted new model tts

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Omni-DuplexEval Benchmark Support in vLLM-Omni 10 days ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Qwen3-Omni /v1/realtime response never finishes 11 days ago

bug help wanted medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: weekly CI, Diffusion · H100 · Multi-GPU over-reserves four H100s for mixed 2/4-GPU tests 12 days ago

bug help wanted diffusion low priority CI/CD

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug][ROCm]: AMD ready / merge CI, CPU-only Simple tests reserve MI300 GPUs 12 days ago

help wanted ROCm CI/CD RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[New Model]: Audio8/Audio8-TTS-Preview-0.6b 12 days ago

help wanted good first issue new model tts

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: PR nightly-test label reruns the full nightly pipeline on every push 12 days ago

bug help wanted low priority CI/CD

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug][NPU]: Nightly Qwen3-Omni perf jobs occupy 4 A3 NPUs for 117-141 minutes 12 days ago

help wanted NPU CI/CD RFC benchmark/profiler/metrics/logger

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug][NPU]: Nightly MiniCPM-o 4.5 accuracy job occupies 2 A3 NPUs for 109 minutes 12 days ago

help wanted NPU CI/CD RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug][NPU]: Nightly Wan2.2 I2V perf job occupies 4 A3 NPUs for 99 minutes 12 days ago

help wanted NPU diffusion CI/CD RFC benchmark/profiler/metrics/logger

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: weekly CI, TTS · L4 serializes all model files and times out at 90 minutes 12 days ago

bug help wanted ci-failure tts CI/CD

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Continuous Quantization Support 13 days ago

help wanted good first issue high priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Nightly / CI failed - Krea-2 README example - 403 GatedRepoError for krea/Krea-2-Raw 13 days ago

bug help wanted ci-failure diffusion CI/CD

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: CUDA ready / merge CI, CPU-only Simple tests reserve L4 GPUs on every build 13 days ago

bug help wanted CI/CD

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: FP8 Quantization for Key and Value Tensors in Diffusion Model Attention Layers 13 days ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: CUBLAS_STATUS_INVALID_VALUE when serving BAGEL-7B-MoT with --quantization fp8 in Docker 13 days ago

bug help wanted

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: NVFP4 Quantization Support for Diffusion Models 13 days ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: TurboQuant — Sub-4-bit KV Cache Quantization for Long-Context Omni Models 13 days ago

enhancement help wanted

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Clean up deploy-config surface for single-stage (esp. diffusion) models 13 days ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: trtllm diffusion attention — roadmap 13 days ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Decouple vllm-omni GGUF quantization from upstream vllm#39583 13 days ago

help wanted good first issue

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: NVIDIA ModelOpt Quantized Diffusion & Video Model Deployment in vLLM-OMNI 13 days ago

help wanted good first issue

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC] Full-Duplex Session Architecture for vLLM-OMNI 13 days ago

help wanted high priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: weekly CI, SenseNova U1 Cache-DiT test hangs until the 120-minute timeout 13 days ago

bug help wanted ci-failure diffusion CI/CD

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Qwen3-Omni seed-TTS — model speaks the task description / answers the prompt instead of reading it; ~2x worse on the vLLM 0.27 rebase 14 days ago

bug help wanted high priority ci-failure

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[New Model]: SenseNova-U1.5-8B-MoT 14 days ago

help wanted new model

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Nightly / CI failed - tests/e2e/online_serving/test_qwen3_omni_expansion.py audio-vs-text checks failing and worsening (1 case on first attempt, 4 on retry) - AssertionError: The audio content is not same as the text 15 days ago

bug help wanted high priority ci-failure

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC] Improve diffusion worker/engine control-plane reliability and output contracts 16 days ago

help wanted high priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Production VAE optimization for diffusion pipelines with MiniMax-H3 17 days ago

help wanted RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: [Cosmos3 optimization] Feature Support, Quantilization, Hardware Recipes, and Scenario Optimization 17 days ago

help wanted good first issue high priority Hardware Plugin diffusion critical

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Feature]: Make diffusion worker process titles and log prefixes topology-aware 17 days ago

enhancement help wanted diffusion

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC] Native MAGI-2 Preview architecture and distributed deployment 18 days ago

help wanted new model

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[New Model]: SenseNova-Vision-7B-MoT 18 days ago

help wanted new model

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: --allowed-local-media-path is ignored for pure diffusion speech 20 days ago

bug help wanted low priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Per-stage runtime.env is parsed but no longer applied during stage launch 20 days ago

bug help wanted high priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug] FP8 quantization and HSDP cannot be enabled simultaneously for Diffusion models (dynamic_scaled_fp8_quant called on CPU tensor) 22 days ago

bug help wanted medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Qwen3-Omni-30B-A3B-Instruct run with error `AssertionError: DP adjusted local rank 1 is out of bounds for 1 devices.` 22 days ago

bug help wanted medium priority omni

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Benchmark Layout Reorganization 23 days ago

help wanted high priority diffusion tts

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Shared CUDA Graph runner for model-local audio submodules 24 days ago

help wanted high priority core RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: CI improvement plan — E2E layering, Acc, pinned deps, job matrix 25 days ago

help wanted CI/CD

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: MOSS-TTS-Local-v1.5 Serving Optimization 26 days ago

help wanted tts RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Merge CI, test_bagel_img2img_shared_memory_connector, AssertionError: Pixel mismatch at (100, 100): expected (156, 172, 217), got (193, 161, 164) 28 days ago

bug help wanted high priority ci-failure

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[New Model]: MAGI-2-preview 30 days ago

help wanted new model

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

/v1/realtime has no tool/function-calling support about 1 month ago

enhancement help wanted good first issue

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: weekly CI, tests/e2e/online_serving/test_flux_kontext_expansion.py, OOM about 1 month ago

bug help wanted ci-failure low priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

vLLM-Omni Q3 Config / Entrypoints Roadmap about 1 month ago

help wanted

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Async Diffusion Output about 1 month ago

help wanted good first issue RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: TTS Development Roadmap - Q2 2026 about 1 month ago

help wanted good first issue high priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug] Qwen3-Omni talker garbles one-word answers (~13-27%/request, seed-deterministic, pre-existing on main); Whisper-small judge inflates test_one_word_prompt_001 failures ~2x about 1 month ago

bug help wanted medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Feature]: Support request-level batching for Wan2.2 pipelines about 1 month ago

enhancement help wanted diffusion

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug] SenseNova-U1 E2E pixel-golden tests fail after think-mode generation drift about 1 month ago

bug help wanted low priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: run tests/e2e/offline_inference/test_mammoth_moda2_expansion.py, ValueError: Tokenizer class MammothUTokenizer does not exist or is not currently imported. about 1 month ago

bug help wanted low priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[New Model]: ovedrive/Qwen-Image-2512-4bit about 1 month ago

help wanted new model

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Regression on dependency declaration about 1 month ago

bug help wanted medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Inaccurate x2v benchmark results about 1 month ago

bug help wanted medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Stage 0 deadlocks on first inference (Qwen3-Omni-30B-A3B-Instruct, 8×L4, v0.26.0rc1) about 1 month ago

bug help wanted medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Nightly CI, Qwen3-Omni-30B-A3B-Instruct, performance metrics regressed by more than 10% compared to the baseline in some scenarios about 1 month ago

bug help wanted high priority ci-failure

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: Talker EngineCore crashes with islice ValueError instead of rejecting completions API for thinker+talker models about 1 month ago

bug help wanted medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: MiniCPM-o 4.5 Daily-Omni accuracy ~66.75% vs paper/HF ~80.2% about 1 month ago

bug help wanted medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Bug]: MiniCPM-o 4.5: streaming chat TTS conditions Talker on 1 text token → unintelligible audio / Seed-TTS WER ≫ 1 about 1 month ago

bug help wanted medium priority

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: MoE performance optimizations about 2 months ago

help wanted diffusion RFC Kernel optimization

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Refactoring Qwen3-Omni and Qwen2.5-Omni thinker about 2 months ago

help wanted

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Ref-Target KV Cache for Video Diffusion Models about 2 months ago

help wanted high priority RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[Feature]: TeaCache support for Wan2.2 V2V / VACE pipeline about 2 months ago

enhancement help wanted diffusion RFC

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[New Model]: Omni-Diffusion about 2 months ago

help wanted good first issue new model

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

[RFC]: Support Deterministic Rollout for RL Training of Omni-Modal & Diffusion Models about 2 months ago

help wanted

vllm-project/vllm-omni

3.2K

A framework for efficient model inference with omni-modality models

Python

#audio-generation#diffusion#image-generation#inference#model-serving#multimodal#pytorch#transformer#video-generation

vllm-project/vllm-omni

Open Issues Need Help