Harry Mellor 4b0da7b60e Enable hybrid attention models for Transformers backend (#18494) ...

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

2025-05-23 10:12:08 +08:00

..

adapter_commons

Update deprecated type hinting in vllm/adapter_commons (#18073)

2025-05-13 08:32:51 -07:00

assets

[Misc] Rename assets for testing (#17575)

2025-05-02 03:29:25 -07:00

attention

[ROCm][Kernel][V1] Enable AMD Radeon GPU Custom Paged Attention on v1 (#17004)

2025-05-21 08:35:00 -07:00

benchmarks

Re-submit: Fix: Proper RGBA -> RGB conversion for PIL images. (#18569)

2025-05-23 01:59:18 +00:00

compilation

[Compile][Platform] Make PiecewiseBackend pluggable and extendable (#18076)

2025-05-22 12:11:53 -07:00

core

[Core] [Bugfix] Add Input Embeddings (#15428)

2025-05-02 01:06:39 -07:00

device_allocator

Update deprecated type hinting in vllm/device_allocator and vllm/distributed (#18126)

2025-05-14 04:07:57 -07:00

distributed

[BugFix][CPU] Fix x86 SHM distributed module initialization (#18536)

2025-05-22 07:35:00 -07:00

engine

[Frontend] [Core] Add Tensorizer support for V1, LoRA adapter serialization and deserialization (#17926)

2025-05-22 18:44:18 -07:00

entrypoints

[Frontend][Bug Fix] Update llama4 pythonic jinja template and llama4_pythonic parser (#17917)

2025-05-22 16:44:08 -07:00

executor

[Feature] Support Pipeline Parallism in torchrun SPMD offline inference for V1 (#17827)

2025-05-15 22:28:27 -07:00

inputs

Feature/vllm/input embedding completion api (#17590)

2025-05-18 20:18:05 -07:00

logging_utils

[Core][Feature] Input metadata dump on crash (#13407)

2025-05-07 22:15:09 +00:00

lora

[Frontend] [Core] Add Tensorizer support for V1, LoRA adapter serialization and deserialization (#17926)

2025-05-22 18:44:18 -07:00

model_executor

Enable hybrid attention models for Transformers backend (#18494)

2025-05-23 10:12:08 +08:00

multimodal

Re-submit: Fix: Proper RGBA -> RGB conversion for PIL images. (#18569)

2025-05-23 01:59:18 +00:00

platforms

[Compile][Platform] Make PiecewiseBackend pluggable and extendable (#18076)

2025-05-22 12:11:53 -07:00

plugins

Update deprecated type hinting in platform, plugins, triton_utils, vllm_flash_attn (#18129)

2025-05-14 05:28:16 -07:00

profiler

Update deprecated type hinting in vllm/profiler (#18057)

2025-05-13 04:34:34 -07:00

prompt_adapter

[Misc] set single whitespace between log sentences (#13771)

2025-02-25 10:26:12 +08:00

reasoning

[V1] Structured Outputs + Thinking compatibility (#16577)

2025-05-14 15:45:24 -07:00

spec_decode

Add full API docs and improve the UX of navigating them (#17485)

2025-05-03 19:42:43 -07:00

third_party

[BugFix] fix some typos found by typos. (#16314)

2025-04-09 03:43:59 -07:00

transformers_utils

Re-submit: Fix: Proper RGBA -> RGB conversion for PIL images. (#18569)

2025-05-23 01:59:18 +00:00

triton_utils

[Bugfix] Fix triton import with local TritonPlaceholder (#17446)

2025-05-06 17:53:09 +08:00

usage

[MISC] fix typo (#18305)

2025-05-17 10:52:09 -07:00

v1

[V1][Spec Decoding] Use model_loader.get_model() to load models (#18273)

2025-05-23 02:05:44 +00:00

vllm_flash_attn

[BugFix] Fix vllm_flash_attn install issues (#17267)

2025-04-27 17:27:56 -07:00

worker

[Bugfix] Fix MRoPE Errors in the Qwen-VL Model When Processing Pure Text (#18526)

2025-05-22 05:22:53 -07:00

__init__.py

[misc] instruct pytorch to use nvml-based cuda check (#15951)

2025-04-03 01:02:58 +08:00

_custom_ops.py

[Kernel] Have rotary embeddings support tensors (#18046)

2025-05-14 15:43:55 -07:00

_ipex_ops.py

[BugFix] Fix mla cpu - missing 3 required positional arguments (#17494)

2025-05-01 14:36:52 +08:00

beam_search.py

[Core][Bugfix] Fix Offline MM Beam Search (#16390)

2025-04-15 10:33:02 +08:00

collect_env.py

[Misc] reformat the collect-env output (#18285)

2025-05-16 19:46:18 -07:00

config.py

Enable hybrid attention models for Transformers backend (#18494)

2025-05-23 10:12:08 +08:00

connections.py

Add full API docs and improve the UX of navigating them (#17485)

2025-05-03 19:42:43 -07:00

env_override.py

[misc] tune some env vars for GB200 (#16992)

2025-04-23 10:59:48 +08:00

envs.py

Throw better error for when running into k8s service discovery issue (#18209)

2025-05-15 21:07:28 -07:00

forward_context.py

Modularize fused experts and integrate PPLX kernels (#15956)

2025-05-14 13:11:54 -07:00

jsontree.py

[VLM] Limit multimodal input cache by memory (#14805)

2025-03-15 02:52:05 -07:00

logger.py

Add full API docs and improve the UX of navigating them (#17485)

2025-05-03 19:42:43 -07:00

logits_process.py

[Bugfix] Fix bad words for Mistral models (#17753)

2025-05-07 23:32:10 -07:00

outputs.py

[Bugfix] Add kwargs to RequestOutput __init__ to be forward compatible (#18513)

2025-05-22 05:24:43 -07:00

pooling_params.py

[Frontend] Using matryoshka_dimensions control the allowed output dimensions. (#16970)

2025-04-24 07:06:28 -07:00

py.typed

Add py.typed so consumers of vLLM can get type checking (#1509)

2023-10-30 14:50:47 -07:00

sampling_params.py

Change top_k to be disabled with 0 (still accept -1 for now) (#17773)

2025-05-09 10:01:49 +00:00

scalar_type.py

[Misc] Fix ScalarType float4 naming (#17690)

2025-05-06 01:07:15 -07:00

scripts.py

[Frontend] Move CLI code into vllm.cmd package (#12971)

2025-02-12 23:12:21 -08:00

sequence.py

[Core] [Bugfix]: tensor parallel with prompt embeds (#18171)

2025-05-19 20:21:27 -07:00

test_utils.py

[Misc] replace does not exist model (#18119)

2025-05-14 02:13:47 -07:00

tracing.py

Update deprecated Python 3.8 typing (#13971)

2025-03-02 17:34:51 -08:00

utils.py

[Platform] Move platform check to right place (#18470)

2025-05-22 12:11:28 -07:00

version.py

[Metrics] Hide deprecated metrics (#15458)

2025-04-02 07:37:19 -07:00