rlx-models 0.2.11

Model loading for RLX — config parsing, safetensors weights, graph builders
Documentation

rlx-models

There is very little structured metadata to build this page from currently. You should check the main library docs, readme, or Cargo.toml in case the author documented the features in them.

This version has 82 feature flags, 6 of them enabled by default.

default

embed (default)

flux2-image (default)

flux2-tokenizer (default)

qwen35-tokenizer (default)

flux2 (default)

qwen35 (default)

aec

all-backends

all-models

amd-gpu

apple-silicon

asr

bert

bioclip2

bonsai

clinicalbert

cohere

coreml

cuda

diamond

diarize

dinov2

encoders

florence2

gemma

gpu

granite

hf-download

kittentts

kyutai-tts

lfm

llada2

llama-base

llama32

llm

locateanything

metal

mimi

minicpm5

mistral

mlx

moshi

neutts

nomic

nvidia-gpu

ocr

omnicoder

orpheus

parity-candle

parity-gemm

parity-llama

parity-pytorch

This feature flag does not enable additional features.

phi

portable-gpu

qwen25-vl

qwen25-vl-vision

qwen3

qwen3-asr

qwen3-tts

qwen35-vlm

rocm

runner

sam

sam-ir

sam2

sam3

tinyllama

tokenizers

tpu

tsac

tts

vad

vision

vision-models

vjepa2

voxtral

voxtral-tts

vulkan

wav2vec2-asr

wav2vec2-bert

whisper

whisper-subtitles