What is ACE-Step-1.5 built with?

Q: Which repos have a stack like ACE-Step-1.5's?

The ones that share most of its stack: mudler/LocalAI, ray-project/ray, and hpcaitech/ColossalAI.

ACE-Step-1.5 is a Python repo built with FastAPI, Lightning, and uvicorn. It stores its data in diskcache, uses AI through accelerate, torchvision, triton-windows, torch, transformers, and 12 more, makes its content with torchcodec, and deploys with loguru. It has 126,537 written lines, with 190 test files.

Read on November 9, 2026, at commit ca1e85f. Read it as markdown

Frontend

Interface 4

Gradio59 files
Rich8 files
matplotlib1 file
Textual

Backend

The base 1

FastAPI28 files

Data 1

diskcache2 files

AI 17

torch125 files
mlx13 files
peft5 files
safetensors4 files
torchaudio4 files
torchao3 files
diffusers2 files
ModelScope2 files
accelerate
torchvision
transformers18 files
einops6 files

5 more

vector-quantize-pytorch6 files
flash-attn4 files
triton3 files
mlx-lm1 file
triton-windows

Support

The base 2

Lightning5 files
uvicorn2 files

Content 1

torchcodec1 file

Deploy 1

loguru98 files

Tools 2

Numba1 file
tensorboard

Other 5

SciPy
soundfile5 files
toml1 file
xxhash1 file
python-multipart

Other repos with a similar stack

Ordered by how much of its stack they share with ACE-Step-1.5, then by stars.

mudlerLocalAI
50.3k
LocalAI is the open-source AI engine. Run any model - LLMs, vision, voice, image, video - on any hardware. No GPU required.
22 of 34 in common · GoCompare →
ray-projectray
44.2k
Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.
19 of 34 in commonCompare →
hpcaitechColossalAI
41.4k
Making large AI models cheaper, faster and more accessible
17 of 34 in commonCompare →
hiyougaLlamaFactory
76k
Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)
16 of 34 in commonCompare →
OpenBMBMiniCPM-V
27.1k
See, hear, and interact on your local devices.
14 of 34 in commonCompare →
debpalashVoiceStudio
105.3k
VoiceStudio is the open-source, fully-local ElevenLabs alternative — voice cloning, voice design, video dubbing, dictation, transcription & audiobook creation in 646 languages.
13 of 34 in commonCompare →

Frequently asked questions

What language is ACE-Step-1.5 written in?

ACE-Step-1.5 is written in Python, with 126,537 written lines.

What framework does ACE-Step-1.5 use?

ACE-Step-1.5 is built with FastAPI, Lightning, and uvicorn.

What database does ACE-Step-1.5 use?

ACE-Step-1.5 stores its data in diskcache.

Does ACE-Step-1.5 use AI?

Yes. ACE-Step-1.5 uses AI through accelerate, torchvision, triton-windows, torch, transformers, mlx, einops, vector-quantize-pytorch, peft, flash-attn, safetensors, torchaudio, torchao, triton, diffusers, ModelScope, and mlx-lm.

Does ACE-Step-1.5 have tests?

Yes. ACE-Step-1.5 has 190 test files.

Which repos have a stack like ACE-Step-1.5's?

The ones that share most of its stack:mudler/LocalAI ray-project/ray hpcaitech/ColossalAI