Use the soundfile alias the file already imports

Return both outputs at the format they were given

Size the rotary embeddings to the attention's head dimension

Pass rotary embeddings into attention, as transformers 5 requires

Declare the more restrictive of the code and weights licenses

Describe what each control does to the sound

Credit the original authors and link the repository and paper

Record the model's tags, as reported by /controls

Bump transformers to a release that permits huggingface-hub 1.x, which gradio 6 requires

Fix the launch keywords and the taxonomy import

Bring the Space up to the pyharp example standard (taxonomy tags, __main__ guard, show_error, v0.4.0)

Correct the license and give the Space an apt emoji and gradient

Bring Space metadata up to HARP deployment guidelines

updated gitignore and comments in app.py

updated .gitignore

hardcode CFG=3.0 acc to paper

simplify app.py, remove debug files

restoring pyharp version

test: restore 49e9b94 - lazy loading, no pyharp

isolating error: 60s blocking before launch

simplifying launch args

restore build_endpoint for HARP

replace build_endpoint to isolate cause

remove .queue() to isolate error cause

remove show_error=False to isolate error cause

remove pwa=True to isolate error cause

using pyharp from github, revert torch

switching to pyharp template

testing with gradio 6.19.0

lazy loading so server startup is not delayed

disabling SSR

force unbuffered stdout, skip model download if present

sdk_version: 5.28.0

binding to 0.0.0.0

writing CPU code

testing org anyaccomp implementation

disable SSR

remove @spaces.GPU, fix deprecation and vocab_size warning

use keyword args for LlamaConfig

fix: restore emoji in README frontmatter

no torchcodec needed

fix: pin torch to ZeroGPU-compatible version, add spaces + @spaces.GPU

initial commit

initial commit