Spaces:
Runtime error
Runtime error
Commit History
-
fix: lower ZeroGPU duration to 120s for free-tier users
All @spaces.GPU(duration=...) decorators exceeded the free-tier ZeroGPU maximum, causing instant rejection before generation starts. Lowered all durations from 180-300s to 120s to fix the error: 'The requested GPU duration (300s) is larger than the maximum allowed' Fixes: community reported issue on HuggingFace Spaced32ba2d
-
fix: move @spaces.GPU decorators to local wrappers to fix ZeroGPU pickling error
On ZeroGPU, Gradio serializes (pickles) event handler functions to send them to GPU worker processes. When fn= was a lambda capturing dit_handler, Python tried to pickle the entire handler including AceStepDiTModel which contains unpicklable local objects in __init__, causing: 'Can't pickle local object AceStepDiTModel.__init__' Fix: Remove @_get_spaces_gpu_decorator from module-level functions and instead apply it to local wrapper functions defined inside setup_event_handlers(), passed directly as fn= (same pattern as the working generation_wrapper). Affected handlers: - process_source_audio (Analyze button) - handle_create_sample (Create Sample button) - handle_format_sample (Format button) - calculate_score_handler_with_selection (Score buttons) - generate_lrc_handler (LRC buttons)6813c41
-
fix is_flash_attention_available
b7a9cc9 -
Merge branch 'main' of https://huggingface.co/spaces/ACE-Step/Ace-Step-v1.5 into main
93cbbd1 -
switch to zero gpu
c60d64b
-
Link this Space to the model so it's visible and accessible from the model page (#5)
- Link this Space to the model so it's visible and accessible from the model page (aeff64b6fe8fa697440758cc7a85696fe4c7cace) Co-authored-by: Victor Mustar <victor@users.noreply.huggingface.co>7396526
-
resolve intermittent CUDA assertion error in concurrent serving scenarios
6d3b89f
-
fix max audio code id
033008e -
reverse fix
048f54f -
fix max audio code id
f2036a7
-
update tech report link
fe01169
-
add acemusic link
71cbaca -
fix cfg kv block allocate
9ff5c6a
-
enforce upload audio duration limits
5f3faee -
fix duration
76de6b9
-
set default_concurrency_limit & set bs=2 no interactive
e19bc36
-
bigger sample btn
ed57c14
-
fix ffmpeg issue
6fafd1b -
fix ffmpeg full install
5a6536d -
load audio fallback
288b94f -
fix torchcodec ffmpeg
96b5f27 -
fix transformers version
d5b1200 -
fix CUDA Graph mismatch in nanovllm
469cce0 -
fix cover load src audio using ffmpeg
ac82e04 -
hide files list
ec6b572 -
support shift3 download
fb234f6
-
update project page url
42e0725
-
refact ui (fix repaint/cover score bug)
9fe9970 -
refact ui
0b990cd
-
fix audio play position reset
Gong Junminfa05c34 -
fix build-essential
Gong Junmin2d1c19e -
fix Qwen3ForCausalLM has no attribute
Gong Junmin13bcd3e -
fix Qwen3ForCausalLM has no attribute
Gong Junmin259f26f -
fix transformer version issue
Gong Junminda29ff1 -
fix data permission
Gong Junminbdee37a -
support nano vllm docker
Gong Junmind85b31b -
support nano vllm docker
Gong Junminf527e35 -
support nano vllm
Gong Junminb8a3728 -
fix: model id
2ac0444 -
fix service mode
Gong Junmin22101a6
-
fix: auto download model
b04b635 -
feat: gradio version
7334110 -
feat: update gradio version
e38053e -
feat: change flash-atten
6ed2902 -
fix: requirements
ccf578a -
fix: update model repo_id
ef09b17 -
feat: update readme
a657594 -
fix: change python version
294dd0f -
feat: update git lfs
9359f63 -
feat: add packages
5e9098a