starred/ace-step-ui
1
0
Fork 0
mirror of https://github.com/fspecii/ace-step-ui.git synced 2026-04-25 06:05:47 +03:00
Code Issues 45 Projects Releases Packages Wiki Activity

[GH-ISSUE #50] Lyrics (and possibly other metadata) fail to generate using Simple mode #45

New issue
Open
opened 2026-02-26 21:31:00 +03:00 by kerem · 0 comments
kerem commented 2026-02-26 21:31:00 +03:00
Owner
Copy link

Originally created by @JoshuaKimsey on GitHub (Feb 12, 2026).
Original GitHub issue: https://github.com/fspecii/ace-step-ui/issues/50

Note: This is a re-post of this issue from https://github.com/cocktailpeanut/ace-step-ui.pinokio/issues/5, since I wanted this report to also go back to the main project and not just the Pinokio wrapper.

In Simple Mode, lyrics simply fail to generate when clicking generate after entering a prompt. Even when the prompt talks about what kind of vocals it should have, lyrics fail to generate. Based on the output in the Pinokio terminal (see below), it seems as though there may be some other missing metadata as well, as songs always seem to generate at 60 seconds in length, and no musical style is ever associated with the song after generation.

I encountered this issue on Linux (Bazzite 43), even after the newest February 10th updates adding back in the Gradio backend. There is an error seen in the output in Pinokio, but I'm unsure if it relates to lyric generation. It also seems like it may not be filling in some of the meta data correctly either? Here's the output in the Pinokio terminal when I use Simple mode:

Job job_1770857610244_hnwbp1j: Queued at position 1
Job job_1770857610244_hnwbp1j: Using Gradio /generation_wrapper {
  prompt: 'Soulful southern rock ballad with slide guitar, Ha',
  duration: undefined,
  batchSize: 1
}
Traceback (most recent call last):
  File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/queueing.py", line 766, in process_events
    response = await route_utils.call_process_api(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/route_utils.py", line 355, in call_process_api
    output = await app.get_blocks().process_api(
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/blocks.py", line 2143, in process_api
    inputs = await self.preprocess_data(
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/blocks.py", line 1790, in preprocess_data
    processed_value = block.preprocess(inputs_cached)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/components/slider.py", line 155, in preprocess
    Number.raise_if_out_of_bounds(payload, self.minimum, self.maximum)
  File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/components/number.py", line 134, in raise_if_out_of_bounds
    raise Error(f"Value {num} is less than minimum value {minimum}.")
gradio.exceptions.Error: 'Value False is less than minimum value 0.01.'
Job job_1770857610244_hnwbp1j: Gradio generation failed, trying Python spawn fallback {
  type: 'status',
  endpoint: '/generation_wrapper',
  fn_index: 73,
  time: 2026-02-12T00:53:30.284Z,
  original_msg: undefined,
  queue: true,
  title: 'Error',
  message: 'Value False is less than minimum value 0.01.',
  visible: true,
  duration: 10,
  stage: 'error',
  code: undefined,
  success: false
}
Job job_1770857610244_hnwbp1j: Using Python spawn (Gradio not available) {
  prompt: 'Soulful southern rock ballad with slide guitar, Ha',
  lyricsPreview: '',
  duration: undefined,
  batchSize: 1
}
[ACE-Step] Skipping import of cpp extensions due to incompatible torch version 2.10.0+cu128 for torchao version 0.15.0             Please see https://github.com/pytorch/ao/issues/2919 for more info
[ACE-Step] 2026-02-11 19:53:34.495 | INFO     | acestep.handler:initialize_service:712 - [initialize_service] Attempting to load model with attention implementation: sdpa
[ACE-Step] `torch_dtype` is deprecated! Use `dtype` instead!
[ACE-Step] 2026-02-11 19:53:34.680 | INFO     | acestep.handler:initialize_service:738 - [initialize_service] Keeping main model on cuda (persistent)
[ACE-Step] 2026-02-11 19:53:36.331 | INFO     | acestep.inference:generate_music:403 - [generate_music] LLM usage decision: thinking=True, use_cot_caption=True, use_cot_language=True, use_cot_metas=True, need_lm_for_cot=True, llm_initialized=False, use_lm=False
[ACE-Step] 2026-02-11 19:53:36.332 | INFO     | acestep.handler:generate_music:3490 - [generate_music] Starting generation...
[ACE-Step] 2026-02-11 19:53:36.332 | INFO     | acestep.handler:generate_music:3493 - [generate_music] Preparing inputs...
[ACE-Step] 2026-02-11 19:53:36.337 | WARNING  | acestep.handler:service_generate:2665 - [service_generate] dmd_gan version: infer_steps 12 exceeds maximum 8, clamping to 8
[ACE-Step] 2026-02-11 19:53:36.343 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:370 - [_load_model_context] Loading vae to cuda
[ACE-Step] 2026-02-11 19:53:36.495 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:383 - [_load_model_context] Loaded vae to cuda in 0.1520s
[ACE-Step] 2026-02-11 19:53:36.505 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:389 - [_load_model_context] Offloading vae to CPU
[ACE-Step] 2026-02-11 19:53:36.718 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:403 - [_load_model_context] Offloaded vae to CPU in 0.2122s
[ACE-Step] 2026-02-11 19:53:36.727 | INFO     | acestep.handler:_prepare_batch:2335 - 
[ACE-Step] ======================================================================
[ACE-Step] 2026-02-11 19:53:36.727 | INFO     | acestep.handler:_prepare_batch:2336 - 🔍 [DEBUG] DiT TEXT ENCODER INPUT (Inference)
[ACE-Step] 2026-02-11 19:53:36.727 | INFO     | acestep.handler:_prepare_batch:2337 - ======================================================================
[ACE-Step] 2026-02-11 19:53:36.727 | INFO     | acestep.handler:_prepare_batch:2338 - text_prompt:
[ACE-Step] # Instruction
[ACE-Step] Fill the audio semantic mask based on the given conditions:
[ACE-Step] # Caption
[ACE-Step] Soulful southern rock ballad with slide guitar, Hammond organ, and heartfelt vocals telling stories of small-town life and growing up in the American South
[ACE-Step] # Metas
[ACE-Step] - bpm: N/A
[ACE-Step] - timesignature: N/A
[ACE-Step] - keyscale: N/A
[ACE-Step] - duration: 60 seconds
[ACE-Step] <|endoftext|>
[ACE-Step] 2026-02-11 19:53:36.727 | INFO     | acestep.handler:_prepare_batch:2339 - ======================================================================
[ACE-Step] 2026-02-11 19:53:36.731 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:370 - [_load_model_context] Loading vae to cuda
[ACE-Step] 2026-02-11 19:53:36.884 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:383 - [_load_model_context] Loaded vae to cuda in 0.1530s
[ACE-Step] 2026-02-11 19:53:36.888 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:389 - [_load_model_context] Offloading vae to CPU
[ACE-Step] 2026-02-11 19:53:37.102 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:403 - [_load_model_context] Offloaded vae to CPU in 0.2136s
[ACE-Step] 2026-02-11 19:53:37.102 | INFO     | acestep.handler:preprocess_batch:2558 - [preprocess_batch] Inferring prompt embeddings...
[ACE-Step] 2026-02-11 19:53:37.102 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:370 - [_load_model_context] Loading text_encoder to cuda
[ACE-Step] `loss_type=None` was set in the config but it is unrecognized. Using the default loss: `ForCausalLMLoss`.
[ACE-Step] 2026-02-11 19:53:37.303 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:383 - [_load_model_context] Loaded text_encoder to cuda in 0.2004s
[ACE-Step] 2026-02-11 19:53:37.428 | INFO     | acestep.handler:preprocess_batch:2561 - [preprocess_batch] Inferring lyric embeddings...
[ACE-Step] 2026-02-11 19:53:37.442 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:389 - [_load_model_context] Offloading text_encoder to CPU
[ACE-Step] 2026-02-11 19:53:37.723 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:403 - [_load_model_context] Offloaded text_encoder to CPU in 0.2808s
[ACE-Step] 2026-02-11 19:53:37.723 | INFO     | acestep.handler:service_generate:2802 - [service_generate] Generating audio... (DiT backend: PyTorch (cuda))
[ACE-Step] 2026-02-11 19:53:37.828 | INFO     | acestep.handler:service_generate:2878 - [service_generate] DiT diffusion via PyTorch (%s)...
[ACE-Step] 2026-02-11 19:53:38.461 | INFO     | acestep.handler:generate_music:3628 - [generate_music] Model generation completed. Decoding latents...
[ACE-Step] 2026-02-11 19:53:38.462 | DEBUG    | acestep.handler:generate_music:3647 - [generate_music] pred_latents: torch.Size([1, 1500, 64]), dtype=torch.bfloat16
[ACE-Step] 2026-02-11 19:53:38.462 | DEBUG    | acestep.handler:generate_music:3648 - [generate_music] time_costs: {'encoder_time_cost': 0.02520442008972168, 'diffusion_time_cost': 0.6076416969299316, 'diffusion_per_step_time_cost': 0.07595521211624146, 'total_time_cost': 0.6328461170196533, 'offload_time_cost': 1.2119483947753906}
[ACE-Step] 2026-02-11 19:53:38.492 | INFO     | acestep.handler:generate_music:3665 - [generate_music] Decoding latents with VAE...
[ACE-Step] 2026-02-11 19:53:38.492 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:370 - [_load_model_context] Loading vae to cuda
[ACE-Step] 2026-02-11 19:53:38.635 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:383 - [_load_model_context] Loaded vae to cuda in 0.1432s
[ACE-Step] 2026-02-11 19:53:38.642 | DEBUG    | acestep.handler:generate_music:3692 - [generate_music] Before VAE decode: allocated=4.80GB, max=5.61GB
[ACE-Step] 2026-02-11 19:53:38.642 | INFO     | acestep.handler:generate_music:3711 - [generate_music] Effective free VRAM before VAE decode: 2.65 GB
[ACE-Step] 2026-02-11 19:53:38.642 | INFO     | acestep.handler:generate_music:3724 - [generate_music] Using tiled VAE decode to reduce VRAM usage...
[ACE-Step] 2026-02-11 19:53:38.642 | DEBUG    | acestep.handler:_get_auto_decode_chunk_size:1315 - [_get_auto_decode_chunk_size] Effective free VRAM: 2.65 GB
[ACE-Step] 2026-02-11 19:53:38.642 | DEBUG    | acestep.handler:_should_offload_wav_to_cpu:1353 - [_should_offload_wav_to_cpu] Effective free VRAM: 2.65 GB
[ACE-Step] 2026-02-11 19:53:38.642 | INFO     | acestep.handler:tiled_decode:2928 - [tiled_decode] chunk_size=512, offload_wav_to_cpu=True, latents_shape=torch.Size([1, 64, 1500])
[ACE-Step] 2026-02-11 19:53:39.139 | DEBUG    | acestep.handler:generate_music:3747 - [generate_music] After VAE decode: allocated=4.95GB, max=6.24GB
[ACE-Step] 2026-02-11 19:53:39.157 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:389 - [_load_model_context] Offloading vae to CPU
[ACE-Step] 2026-02-11 19:53:39.363 | INFO     | acestep.core.generation.handler.init_service:_load_model_context:403 - [_load_model_context] Offloaded vae to CPU in 0.2064s
[ACE-Step] 2026-02-11 19:53:39.363 | INFO     | acestep.handler:generate_music:3768 - [generate_music] VAE decode completed. Preparing audio tensors...
[ACE-Step] 2026-02-11 19:53:39.363 | INFO     | acestep.handler:generate_music:3783 - [generate_music] Done! Generated 1 audio tensors.
[ACE-Step] 2026-02-11 19:53:39.367 | INFO     | acestep.inference:generate_music:662 - [Normalization] Audio 0 BEFORE: Peak=1.0000, Target=-1.0dB
[ACE-Step] 2026-02-11 19:53:39.377 | INFO     | acestep.inference:generate_music:667 - [Normalization] Audio 0 AFTER: Peak=0.8913
[ACE-Step] 2026-02-11 19:53:40.527 | DEBUG    | acestep.audio_utils:save_audio:187 - [AudioSaver] Saved audio to /var/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/output/job_1770857610244_hnwbp1j/9c512dfe-fe29-2608-6a24-e4f882e0a6c4.mp3 (mp3, 48000Hz)
Job job_1770857610244_hnwbp1j: Completed via Python in 4.2s with 1 audio files
Originally created by @JoshuaKimsey on GitHub (Feb 12, 2026). Original GitHub issue: https://github.com/fspecii/ace-step-ui/issues/50 _Note: This is a re-post of this issue from https://github.com/cocktailpeanut/ace-step-ui.pinokio/issues/5, since I wanted this report to also go back to the main project and not just the Pinokio wrapper._ In Simple Mode, lyrics simply fail to generate when clicking generate after entering a prompt. Even when the prompt talks about what kind of vocals it should have, lyrics fail to generate. Based on the output in the Pinokio terminal (see below), it seems as though there may be some other missing metadata as well, as songs always seem to generate at 60 seconds in length, and no musical style is ever associated with the song after generation. I encountered this issue on Linux (Bazzite 43), even after the newest February 10th updates adding back in the Gradio backend. There is an error seen in the output in Pinokio, but I'm unsure if it relates to lyric generation. It also seems like it may not be filling in some of the meta data correctly either? Here's the output in the Pinokio terminal when I use Simple mode: ```Bash Job job_1770857610244_hnwbp1j: Queued at position 1 Job job_1770857610244_hnwbp1j: Using Gradio /generation_wrapper { prompt: 'Soulful southern rock ballad with slide guitar, Ha', duration: undefined, batchSize: 1 } Traceback (most recent call last): File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/queueing.py", line 766, in process_events response = await route_utils.call_process_api( ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/route_utils.py", line 355, in call_process_api output = await app.get_blocks().process_api( ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/blocks.py", line 2143, in process_api inputs = await self.preprocess_data( ^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/blocks.py", line 1790, in preprocess_data processed_value = block.preprocess(inputs_cached) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/components/slider.py", line 155, in preprocess Number.raise_if_out_of_bounds(payload, self.minimum, self.maximum) File "/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/env/lib/python3.11/site-packages/gradio/components/number.py", line 134, in raise_if_out_of_bounds raise Error(f"Value {num} is less than minimum value {minimum}.") gradio.exceptions.Error: 'Value False is less than minimum value 0.01.' Job job_1770857610244_hnwbp1j: Gradio generation failed, trying Python spawn fallback { type: 'status', endpoint: '/generation_wrapper', fn_index: 73, time: 2026-02-12T00:53:30.284Z, original_msg: undefined, queue: true, title: 'Error', message: 'Value False is less than minimum value 0.01.', visible: true, duration: 10, stage: 'error', code: undefined, success: false } Job job_1770857610244_hnwbp1j: Using Python spawn (Gradio not available) { prompt: 'Soulful southern rock ballad with slide guitar, Ha', lyricsPreview: '', duration: undefined, batchSize: 1 } [ACE-Step] Skipping import of cpp extensions due to incompatible torch version 2.10.0+cu128 for torchao version 0.15.0 Please see https://github.com/pytorch/ao/issues/2919 for more info [ACE-Step] 2026-02-11 19:53:34.495 | INFO | acestep.handler:initialize_service:712 - [initialize_service] Attempting to load model with attention implementation: sdpa [ACE-Step] `torch_dtype` is deprecated! Use `dtype` instead! [ACE-Step] 2026-02-11 19:53:34.680 | INFO | acestep.handler:initialize_service:738 - [initialize_service] Keeping main model on cuda (persistent) [ACE-Step] 2026-02-11 19:53:36.331 | INFO | acestep.inference:generate_music:403 - [generate_music] LLM usage decision: thinking=True, use_cot_caption=True, use_cot_language=True, use_cot_metas=True, need_lm_for_cot=True, llm_initialized=False, use_lm=False [ACE-Step] 2026-02-11 19:53:36.332 | INFO | acestep.handler:generate_music:3490 - [generate_music] Starting generation... [ACE-Step] 2026-02-11 19:53:36.332 | INFO | acestep.handler:generate_music:3493 - [generate_music] Preparing inputs... [ACE-Step] 2026-02-11 19:53:36.337 | WARNING | acestep.handler:service_generate:2665 - [service_generate] dmd_gan version: infer_steps 12 exceeds maximum 8, clamping to 8 [ACE-Step] 2026-02-11 19:53:36.343 | INFO | acestep.core.generation.handler.init_service:_load_model_context:370 - [_load_model_context] Loading vae to cuda [ACE-Step] 2026-02-11 19:53:36.495 | INFO | acestep.core.generation.handler.init_service:_load_model_context:383 - [_load_model_context] Loaded vae to cuda in 0.1520s [ACE-Step] 2026-02-11 19:53:36.505 | INFO | acestep.core.generation.handler.init_service:_load_model_context:389 - [_load_model_context] Offloading vae to CPU [ACE-Step] 2026-02-11 19:53:36.718 | INFO | acestep.core.generation.handler.init_service:_load_model_context:403 - [_load_model_context] Offloaded vae to CPU in 0.2122s [ACE-Step] 2026-02-11 19:53:36.727 | INFO | acestep.handler:_prepare_batch:2335 - [ACE-Step] ====================================================================== [ACE-Step] 2026-02-11 19:53:36.727 | INFO | acestep.handler:_prepare_batch:2336 - 🔍 [DEBUG] DiT TEXT ENCODER INPUT (Inference) [ACE-Step] 2026-02-11 19:53:36.727 | INFO | acestep.handler:_prepare_batch:2337 - ====================================================================== [ACE-Step] 2026-02-11 19:53:36.727 | INFO | acestep.handler:_prepare_batch:2338 - text_prompt: [ACE-Step] # Instruction [ACE-Step] Fill the audio semantic mask based on the given conditions: [ACE-Step] # Caption [ACE-Step] Soulful southern rock ballad with slide guitar, Hammond organ, and heartfelt vocals telling stories of small-town life and growing up in the American South [ACE-Step] # Metas [ACE-Step] - bpm: N/A [ACE-Step] - timesignature: N/A [ACE-Step] - keyscale: N/A [ACE-Step] - duration: 60 seconds [ACE-Step] <|endoftext|> [ACE-Step] 2026-02-11 19:53:36.727 | INFO | acestep.handler:_prepare_batch:2339 - ====================================================================== [ACE-Step] 2026-02-11 19:53:36.731 | INFO | acestep.core.generation.handler.init_service:_load_model_context:370 - [_load_model_context] Loading vae to cuda [ACE-Step] 2026-02-11 19:53:36.884 | INFO | acestep.core.generation.handler.init_service:_load_model_context:383 - [_load_model_context] Loaded vae to cuda in 0.1530s [ACE-Step] 2026-02-11 19:53:36.888 | INFO | acestep.core.generation.handler.init_service:_load_model_context:389 - [_load_model_context] Offloading vae to CPU [ACE-Step] 2026-02-11 19:53:37.102 | INFO | acestep.core.generation.handler.init_service:_load_model_context:403 - [_load_model_context] Offloaded vae to CPU in 0.2136s [ACE-Step] 2026-02-11 19:53:37.102 | INFO | acestep.handler:preprocess_batch:2558 - [preprocess_batch] Inferring prompt embeddings... [ACE-Step] 2026-02-11 19:53:37.102 | INFO | acestep.core.generation.handler.init_service:_load_model_context:370 - [_load_model_context] Loading text_encoder to cuda [ACE-Step] `loss_type=None` was set in the config but it is unrecognized. Using the default loss: `ForCausalLMLoss`. [ACE-Step] 2026-02-11 19:53:37.303 | INFO | acestep.core.generation.handler.init_service:_load_model_context:383 - [_load_model_context] Loaded text_encoder to cuda in 0.2004s [ACE-Step] 2026-02-11 19:53:37.428 | INFO | acestep.handler:preprocess_batch:2561 - [preprocess_batch] Inferring lyric embeddings... [ACE-Step] 2026-02-11 19:53:37.442 | INFO | acestep.core.generation.handler.init_service:_load_model_context:389 - [_load_model_context] Offloading text_encoder to CPU [ACE-Step] 2026-02-11 19:53:37.723 | INFO | acestep.core.generation.handler.init_service:_load_model_context:403 - [_load_model_context] Offloaded text_encoder to CPU in 0.2808s [ACE-Step] 2026-02-11 19:53:37.723 | INFO | acestep.handler:service_generate:2802 - [service_generate] Generating audio... (DiT backend: PyTorch (cuda)) [ACE-Step] 2026-02-11 19:53:37.828 | INFO | acestep.handler:service_generate:2878 - [service_generate] DiT diffusion via PyTorch (%s)... [ACE-Step] 2026-02-11 19:53:38.461 | INFO | acestep.handler:generate_music:3628 - [generate_music] Model generation completed. Decoding latents... [ACE-Step] 2026-02-11 19:53:38.462 | DEBUG | acestep.handler:generate_music:3647 - [generate_music] pred_latents: torch.Size([1, 1500, 64]), dtype=torch.bfloat16 [ACE-Step] 2026-02-11 19:53:38.462 | DEBUG | acestep.handler:generate_music:3648 - [generate_music] time_costs: {'encoder_time_cost': 0.02520442008972168, 'diffusion_time_cost': 0.6076416969299316, 'diffusion_per_step_time_cost': 0.07595521211624146, 'total_time_cost': 0.6328461170196533, 'offload_time_cost': 1.2119483947753906} [ACE-Step] 2026-02-11 19:53:38.492 | INFO | acestep.handler:generate_music:3665 - [generate_music] Decoding latents with VAE... [ACE-Step] 2026-02-11 19:53:38.492 | INFO | acestep.core.generation.handler.init_service:_load_model_context:370 - [_load_model_context] Loading vae to cuda [ACE-Step] 2026-02-11 19:53:38.635 | INFO | acestep.core.generation.handler.init_service:_load_model_context:383 - [_load_model_context] Loaded vae to cuda in 0.1432s [ACE-Step] 2026-02-11 19:53:38.642 | DEBUG | acestep.handler:generate_music:3692 - [generate_music] Before VAE decode: allocated=4.80GB, max=5.61GB [ACE-Step] 2026-02-11 19:53:38.642 | INFO | acestep.handler:generate_music:3711 - [generate_music] Effective free VRAM before VAE decode: 2.65 GB [ACE-Step] 2026-02-11 19:53:38.642 | INFO | acestep.handler:generate_music:3724 - [generate_music] Using tiled VAE decode to reduce VRAM usage... [ACE-Step] 2026-02-11 19:53:38.642 | DEBUG | acestep.handler:_get_auto_decode_chunk_size:1315 - [_get_auto_decode_chunk_size] Effective free VRAM: 2.65 GB [ACE-Step] 2026-02-11 19:53:38.642 | DEBUG | acestep.handler:_should_offload_wav_to_cpu:1353 - [_should_offload_wav_to_cpu] Effective free VRAM: 2.65 GB [ACE-Step] 2026-02-11 19:53:38.642 | INFO | acestep.handler:tiled_decode:2928 - [tiled_decode] chunk_size=512, offload_wav_to_cpu=True, latents_shape=torch.Size([1, 64, 1500]) [ACE-Step] 2026-02-11 19:53:39.139 | DEBUG | acestep.handler:generate_music:3747 - [generate_music] After VAE decode: allocated=4.95GB, max=6.24GB [ACE-Step] 2026-02-11 19:53:39.157 | INFO | acestep.core.generation.handler.init_service:_load_model_context:389 - [_load_model_context] Offloading vae to CPU [ACE-Step] 2026-02-11 19:53:39.363 | INFO | acestep.core.generation.handler.init_service:_load_model_context:403 - [_load_model_context] Offloaded vae to CPU in 0.2064s [ACE-Step] 2026-02-11 19:53:39.363 | INFO | acestep.handler:generate_music:3768 - [generate_music] VAE decode completed. Preparing audio tensors... [ACE-Step] 2026-02-11 19:53:39.363 | INFO | acestep.handler:generate_music:3783 - [generate_music] Done! Generated 1 audio tensors. [ACE-Step] 2026-02-11 19:53:39.367 | INFO | acestep.inference:generate_music:662 - [Normalization] Audio 0 BEFORE: Peak=1.0000, Target=-1.0dB [ACE-Step] 2026-02-11 19:53:39.377 | INFO | acestep.inference:generate_music:667 - [Normalization] Audio 0 AFTER: Peak=0.8913 [ACE-Step] 2026-02-11 19:53:40.527 | DEBUG | acestep.audio_utils:save_audio:187 - [AudioSaver] Saved audio to /var/home/jkimsey/pinokio/api/ace-step-ui.pinokio.git/app/ACE-Step-1.5/output/job_1770857610244_hnwbp1j/9c512dfe-fe29-2608-6a24-e4f882e0a6c4.mp3 (mp3, 48000Hz) Job job_1770857610244_hnwbp1j: Completed via Python in 4.2s with 1 audio files ```
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
gradio-experimental
No results found.
Labels
Clear labels   
pull-request

Mirrored from GitHub Pull Request

No labels
pull-request
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
starred/ace-step-ui#45
No description provided.