[Request] Parsed ggml-turbo/request0.json [Request] seed=42 lm_batch=1 synth_batch=1 [Request] caption: An upbeat and anthemic pop-rock track driven by bright, slig... [Request] lyrics: 490 bytes [Request] bpm=83 dur=88 key=G major ts=4 lang=fr [Request] lm: temp=0.85 cfg=2.0 top_p=0.90 top_k=0 [Request] dit: steps=8 guidance=0.0 shift=3.0 [Request] task_type: text2music [Request] solver: euler (stork_substeps=10) [Request] lm_mode: generate [Request] output_format: mp3 [Request] synth_model: acestep-v15-turbo-Q6_K.gguf [Request] vae: vae-BF16.gguf [Request] audio_codes: (present) [Ace-Synth] Batch: 1 request(s) [Registry] Qwen3-Embedding-0.6B-BF16.gguf -> Text-Enc [Registry] Qwen3-Embedding-0.6B-Q8_0.gguf -> Text-Enc [Registry] acestep-5Hz-lm-0.6B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-0.6B-Q8_0.gguf -> LM [Registry] acestep-5Hz-lm-1.7B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-1.7B-Q8_0.gguf -> LM [Registry] acestep-5Hz-lm-4B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q5_K_M.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q6_K.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q8_0.gguf -> LM [Registry] acestep-v15-base-BF16.gguf -> DiT [Registry] acestep-v15-base-Q4_K_M.gguf -> DiT [Registry] acestep-v15-base-Q5_K_M.gguf -> DiT [Registry] acestep-v15-base-Q6_K.gguf -> DiT [Registry] acestep-v15-base-Q8_0.gguf -> DiT [Registry] acestep-v15-sft-BF16.gguf -> DiT [Registry] acestep-v15-sft-Q4_K_M.gguf -> DiT [Registry] acestep-v15-sft-Q5_K_M.gguf -> DiT [Registry] acestep-v15-sft-Q6_K.gguf -> DiT [Registry] acestep-v15-sft-Q8_0.gguf -> DiT [Registry] acestep-v15-sftturbo50-BF16.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q4_K_M.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q5_K_M.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q6_K.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-BF16.gguf -> DiT [Registry] acestep-v15-turbo-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-continuous-BF16.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-shift1-BF16.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-shift3-BF16.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-base-BF16.gguf -> DiT [Registry] acestep-v15-xl-base-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-base-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-base-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-base-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-sft-BF16.gguf -> DiT [Registry] acestep-v15-xl-sft-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-sft-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-sft-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-sft-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-BF16.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-turbo-BF16.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q8_0.gguf -> DiT [Registry] scragvae-BF16.gguf -> VAE [Registry] vae-BF16.gguf -> VAE [Store] Created (policy=STRICT) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-turbo-Q6_K.gguf: 678 tensors, data at offset 56864 [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-turbo-Q6_K.gguf: 678 tensors, data at offset 56864 [Synth-Load] Ready: turbo=yes, fa=no, batch_cfg=yes [Resolve-Params] max audio codes across batch: 434 (86.8s @ 5Hz) [Resolve-T] T=2170, S=1085 [Resolve-T] seed=42, steps=8, guidance=1.0, shift=3.0, duration=88.0s [BPE] Loaded from GGUF: 151643 vocab, 151387 merges ggml_cuda_init: found 1 CUDA devices (Total VRAM: 97247 MiB): Device 0: NVIDIA RTX PRO 6000 Blackwell Workstation Edition, compute capability 12.0, VMM: yes, VRAM: 97247 MiB load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so ggml_vulkan: Found 1 Vulkan devices: ggml_vulkan: 0 = NVIDIA RTX PRO 6000 Blackwell Workstation Edition (NVIDIA) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 49152 | int dot: 1 | matrix cores: NV_coopmat2 load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] TextEncoder backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/Qwen3-Embedding-0.6B-BF16.gguf: 310 tensors, data at offset 5337568 [Load] TextEncoder: 28L, H=1024, Nh=16/8 [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [WeightCtx] Loaded 310 tensors, 1136.5 MB into backend [Store] Load TextEnc: 257 ms [Debug] text_hidden: [70, 1024] first4: 3.704526 2.436253 0.222853 -13.131872 [Debug] lyric_embed: [167, 1024] first4: 0.029175 0.032227 -0.022339 -0.028809 [Store] Unload TextEnc (1136.5 MB) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] CondEncoder backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-turbo-Q6_K.gguf: 678 tensors, data at offset 56864 [Load] LyricEncoder: 8L [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [Load] TimbreEncoder: 4L [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [WeightCtx] Loaded 140 tensors, 476.3 MB into backend [Load] CondEncoder: lyric(8L), timbre(4L), text_proj, null_cond [Store] Load CondEnc: 302 ms [CondEnc] Lyric sliding mask: 167x167, window=128 [CondEnc] Timbre sliding mask: 1x1, window=128 [Encode] Packed: lyric=167 + timbre=1 + text=70 = 238 tokens [Encode-Text Batch0] 70+167 tokens -> enc_S=238, 98.9 ms [Debug] enc_hidden: [238, 2048] first4: 1.762379 -0.051814 -0.131966 0.057631 [Store] Unload CondEnc (476.3 MB) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] Detokenizer backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-turbo-Q6_K.gguf: 678 tensors, data at offset 56864 [WeightCtx] Loaded 30 tensors, 82.2 MB into backend [Load] Detokenizer: FSQ(6->2048) + 2L encoder(S=5, 2048->64) [Store] Load FSQ-Detok: 60 ms [Context] Decoded: 434 codes -> 2170 frames (86.8s @ 25Hz) [Build-Context Batch0] Detokenizer: 351.6 ms, 434 codes [Debug] detok_output: [2170, 64] first4: -0.151355 1.462444 0.326907 -0.627213 [Store] Unload FSQ-Detok (82.2 MB) [Init-Noise Batch0] Philox noise seed=42, [2170, 64] solver=euler [Debug] noise: [2170, 64] first4: 0.194336 2.156250 -0.171875 0.847656 [Debug] context: [2170, 128] first4: -0.151355 1.462444 0.326907 -0.627213 [Init-Noise] Starting: T=2170, S=1085, enc_S=238, steps=8, batch=1 (cover) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] DiT backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-turbo-Q6_K.gguf: 678 tensors, data at offset 56864 [DiT] Self-attn: Q+K+V fused [DiT] Cross-attn: Q+K+V fused [DiT] MLP: gate+up fused [Load] null_condition_emb found (CFG available) [WeightCtx] Loaded 478 tensors, 1237.2 MB into backend [Load] DiT: 24 layers, H=2048, Nh=16/8, D=128 [Store] Load DiT: 927 ms [DiT] Batch N=1, T=2170, S=1085, enc_S=238 [DiT] Graph: 2129 nodes [DiT] Solver: ODE Euler (1 NFE/step, order 1) [Debug] tproj: [12288] first4: 0.261809 -0.161156 -0.099489 0.050901 [Debug] temb: [2048] first4: 0.000441 -0.132284 -0.035603 0.064823 [Debug] temb_t: [2048] first4: 0.001519 0.026983 -0.052936 0.063921 [Debug] temb_r: [2048] first4: -0.001078 -0.159268 0.017333 0.000903 [Debug] sinusoidal_t: [256] first4: 0.562379 0.789627 0.439928 -0.023645 [Debug] sinusoidal_r: [256] first4: 1.000000 1.000000 1.000000 1.000000 [Debug] temb_lin1_t: [2048] first4: -0.049242 -0.050737 -0.017494 -0.036973 [Debug] temb_lin1_r: [2048] first4: -0.014408 -0.020609 -0.015729 0.003875 [Debug] hidden_after_proj_in: [2048, 1085] first4: -1.041706 -0.935163 0.543316 0.447904 [Debug] proj_in_input: [192, 2170] first4: -0.151355 1.462444 0.326907 -0.627213 [Debug] enc_after_cond_emb: [2048, 238] first4: -0.166439 0.831665 0.328897 -0.589582 [Debug] layer0_sa_input: [2048, 1085] first4: -0.719262 -0.743265 -0.048909 0.260726 [Debug] layer0_q_after_rope: [128, 16] first4: -1.546695 -1.029571 0.220500 0.458906 [Debug] layer0_k_after_rope: [128, 8] first4: -0.166439 0.831665 0.328897 -0.589582 [Debug] layer0_sa_output: [2048, 1085] first4: -1.511177 0.212656 -0.330983 0.523252 [Debug] layer0_attn_out: [2048, 1085] first4: -1.546695 -1.029571 0.220500 0.458906 [Debug] layer0_after_self_attn: [2048, 1085] first4: -1.546695 -1.029571 0.220500 0.458906 [Debug] layer0_after_cross_attn: [2048, 1085] first4: -1.609587 -0.798885 -0.285115 0.500374 [Debug] hidden_after_layer0: [2048, 1085] first4: -9.238791 0.594274 52.893948 -0.948797 [Debug] hidden_after_layer6: [2048, 1085] first4: -21.246397 -0.218955 36.391987 -4.624691 [Debug] hidden_after_layer12: [2048, 1085] first4: -15.351803 -16.536158 76.883438 31.149061 [Debug] hidden_after_layer18: [2048, 1085] first4: -28.788742 13.380124 63.779480 20.173618 [Debug] hidden_after_layer23: [2048, 1085] first4: -20.930138 54.459087 193.014542 140.836411 [Debug] dit_step0_vt: [2170, 64] first4: 0.096821 1.129415 0.285259 2.376315 [Debug] dit_step0_xt: [2170, 64] first4: 0.189935 2.104913 -0.184841 0.739642 [DiT] Step 1/8 t=1.000 [Debug] dit_step1_vt: [2170, 64] first4: -0.109089 1.309798 -0.244168 1.987816 [Debug] dit_step1_xt: [2170, 64] first4: 0.195885 2.033469 -0.171523 0.631216 [DiT] Step 2/8 t=0.955 [Debug] dit_step2_vt: [2170, 64] first4: 0.094821 1.245678 0.101010 2.413914 [Debug] dit_step2_xt: [2170, 64] first4: 0.189564 1.950424 -0.178257 0.470288 [DiT] Step 3/8 t=0.900 [Debug] dit_step3_vt: [2170, 64] first4: 0.313540 1.102525 0.261468 2.697613 [Debug] dit_step3_xt: [2170, 64] first4: 0.163436 1.858547 -0.200046 0.245487 [DiT] Step 4/8 t=0.833 [Debug] dit_step4_vt: [2170, 64] first4: 0.285723 0.994373 0.111937 2.733400 [Debug] dit_step4_xt: [2170, 64] first4: 0.132822 1.752007 -0.212039 -0.047377 [DiT] Step 5/8 t=0.750 [Debug] dit_step5_vt: [2170, 64] first4: 0.362832 0.887874 -0.305495 2.771905 [Debug] dit_step5_xt: [2170, 64] first4: 0.080989 1.625168 -0.168397 -0.443364 [DiT] Step 6/8 t=0.643 [Debug] dit_step6_vt: [2170, 64] first4: 0.296376 0.640049 -0.759461 2.749637 [Debug] dit_step6_xt: [2170, 64] first4: 0.021714 1.497158 -0.016505 -0.993291 [DiT] Step 7/8 t=0.500 [Debug] dit_step7_vt: [2170, 64] first4: 0.215550 0.167503 -1.350311 2.964794 [Debug] dit_x0: [2170, 64] first4: -0.042951 1.446907 0.388589 -1.882729 [DiT] Step 8/8 t=0.300 [DiT-Generate] Total: 21510.8 ms (21510.8 ms/sample) [Debug] dit_output: [2170, 64] first4: -0.042951 1.446907 0.388589 -1.882729 [Store] Unload DiT (1237.2 MB) [GGUF] /mnt/workspace/git/acestep.cpp/models/vae-BF16.gguf: 365 tensors, data at offset 30048 load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] VAE backend: CPU (CPU threads: 16) [VAE] Backend: CPU, Weight buffer: 161.1 MB [VAE] Loaded: 5 blocks, upsample=1920x, F32 activations [Store] Load VAE-Dec: 711 ms [VAE] Tiled decode: 3 tiles (chunk=1024, overlap=64, stride=896) [VAE] Graph: 479 nodes, T_latent=960 [VAE] Upsample factor: 1920.00 (expected ~1920) [VAE] Graph: 479 nodes, T_latent=1024 [VAE] Graph: 479 nodes, T_latent=442 [VAE] Tiled decode done: 3 tiles -> T_audio=4166400 (86.80s @ 48kHz) [VAE-Decode Batch0] Decode: 31898.2 ms [Debug] vae_audio: [2, 4166400] first4: 0.000737 0.001471 0.001250 0.001778 [Store] Unload VAE-Dec (161.1 MB) [MP3] Encoding 86.8s @ 128 kbps, 48000 Hz stereo [MP3] 1388928 bytes (12.0:1), 499 ms (173.84x realtime), 32 threads [MP3] Wrote ggml-turbo/request00.mp3 [Ace-Synth] All done 2026-07-12 09:40:46.805 | DEBUG | acestep.core.generation.handler.init_service_memory_basic:_apply_malloc_mmap_threshold:50 - [memory] Set M_MMAP_THRESHOLD=131072 for immediate OS reclaim of large frees Skipping import of cpp extensions due to incompatible torch version. Please upgrade to torch >= 2.11.0 (found 2.10.0+cu128). 2026-07-12 09:40:48.499 | WARNING | acestep.training.trainer::40 - bitsandbytes not installed. Using standard AdamW. 2026-07-12 09:40:48.501 | INFO | acestep.gpu_config:get_gpu_memory_gb:578 - CUDA GPU detected: NVIDIA RTX PRO 6000 Blackwell Workstation Edition (95.0 GB) 2026-07-12 09:40:48.580 | INFO | acestep.core.generation.handler.init_service_loader:_load_main_model_from_checkpoint:174 - [initialize_service] Attempting to load model with attention implementation: sdpa Unable to import `torchao` Tensor objects. This may affect loading checkpoints serialized with `torchao` 2026-07-12 09:40:50.083 | INFO | acestep.core.generation.handler.generate_music:generate_music:289 - [generate_music] Turbo model detected: overriding guidance_scale 0.0 -> 1.0 (turbo does not use CFG). 2026-07-12 09:40:50.083 | INFO | acestep.core.generation.handler.generate_music:generate_music:304 - [generate_music] Starting generation... 2026-07-12 09:40:50.083 | INFO | acestep.core.generation.handler.generate_music:generate_music:307 - [generate_music] Preparing inputs... 2026-07-12 09:40:50.084 | INFO | acestep.core.generation.handler.generate_music_request:_prepare_reference_and_source_audio:179 - [generate_music] cover task: no src_audio but audio codes provided, proceeding with codes 2026-07-12 09:40:50.087 | INFO | acestep.core.generation.handler.generate_music:_vram_preflight_check:156 - [generate_music] VRAM pre-flight: 87.83 GB free, ~0.94 GB needed (batch=1, duration=88s, mode=turbo). 2026-07-12 09:40:50.092 | INFO | acestep.core.generation.handler.conditioning_target:_prepare_target_latents_and_wavs:78 - [generate_music] Decoding audio codes for item 0... 2026-07-12 09:40:50.290 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_precomputed_lm_hints:31 - [generate_music] Decoding audio codes for LM hints for item 0... 2026-07-12 09:40:50.292 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:119 - ====================================================================== 2026-07-12 09:40:50.293 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:120 - πŸ” [DEBUG] DiT TEXT ENCODER INPUT (Inference) 2026-07-12 09:40:50.293 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:121 - ====================================================================== 2026-07-12 09:40:50.293 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:122 - text_prompt: # Instruction Generate audio semantic tokens based on the given conditions: # Caption An upbeat and anthemic pop-rock track driven by bright, slightly overdriven # Metas - bpm: 83 - timesignature: 4 - keyscale: G major - duration: 88 seconds <|endoftext|> 2026-07-12 09:40:50.293 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:123 - ====================================================================== 2026-07-12 09:40:50.293 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:124 - lyrics_text: # Languages fr # Lyric # Lyric [Intro - Guitar Riff] [Verse 1] Dans le monde des tutos virtuels G ta toise en nouvelle passion Avec Ggendoline et PumbΓ© Γ  midi La communautΓ©, c'est l'unitΓ© Quel joie, une clΓ© [Chorus] Dans le monde des tutos virtuels GΓ’ndoline et PumbΓ© Γ  midi Une famille Γ  connecter, c'est vrai D'un enfant qui voit toi fusionner [Guitar Solo] [Verse 2] Dans le monde des tutos virtuels GΓ’ndoline, PumbΓ© Γ  midi Une famille Γ  connecter, c'est vrai D'un enfant qui voit toi fusionner<|endoftext|> 2026-07-12 09:40:50.293 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:125 - ====================================================================== 2026-07-12 09:40:50.299 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:110 - [preprocess_batch] Inferring prompt embeddings... 2026-07-12 09:40:50.314 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:113 - [preprocess_batch] Inferring lyric embeddings... 2026-07-12 09:40:50.314 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:161 - [service_generate] Generating audio... (DiT backend: PyTorch (cuda)) 2026-07-12 09:40:50.341 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:263 - [service_generate] DiT diffusion via PyTorch (cuda)... 2026-07-12 09:40:50.349 | INFO | acestep.models.common.dcw_correction:__init__:98 - [DCW] Active β€” mode=double, scaler=0.0500, high_scaler=0.0200, wavelet='haar' 2026-07-12 09:40:50.434 | WARNING | acestep.models.common.dcw_loader:_try_import:39 - DCW is enabled but 'pytorch_wavelets' is not installed. Install with `pip install pytorch_wavelets PyWavelets` to use Differential Correction in Wavelet domain. Falling back to no-op for this generation. 2026-07-12 09:40:50.658 | INFO | acestep.core.generation.handler.generate_music_decode:_prepare_generate_music_decode_state:42 - [generate_music] Model generation completed. Decoding latents... 2026-07-12 09:40:50.667 | DEBUG | acestep.core.generation.handler.generate_music_decode:_prepare_generate_music_decode_state:64 - [generate_music] pred_latents: torch.Size([1, 2170, 64]), dtype=torch.bfloat16 2026-07-12 09:40:50.667 | DEBUG | acestep.core.generation.handler.generate_music_decode:_prepare_generate_music_decode_state:65 - [generate_music] time_costs: {'encoder_time_cost': 0.006986379623413086, 'diffusion_time_cost': 0.309267520904541, 'diffusion_per_step_time_cost': 0.03865844011306763, 'total_time_cost': 0.3162539005279541, 'offload_time_cost': 0.0} 2026-07-12 09:40:50.672 | INFO | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:127 - [generate_music] Decoding latents with VAE... 2026-07-12 09:40:50.675 | DEBUG | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:136 - [generate_music] Before VAE decode: allocated=5.98GB, max=7.29GB 2026-07-12 09:40:50.675 | INFO | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:154 - [generate_music] Effective free VRAM before VAE decode: 87.72 GB 2026-07-12 09:40:50.675 | INFO | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:172 - [generate_music] Using tiled VAE decode to reduce VRAM usage... 2026-07-12 09:40:50.675 | DEBUG | acestep.core.generation.handler.memory_utils:_get_auto_decode_chunk_size:108 - [_get_auto_decode_chunk_size] Effective free VRAM: 87.72 GB 2026-07-12 09:40:50.675 | DEBUG | acestep.core.generation.handler.memory_utils:_should_offload_wav_to_cpu:131 - [_should_offload_wav_to_cpu] Effective free VRAM: 87.72 GB 2026-07-12 09:40:50.675 | INFO | acestep.core.generation.handler.vae_decode:tiled_decode:56 - [tiled_decode] chunk_size=512, offload_wav_to_cpu=False, latents_shape=torch.Size([1, 64, 2170]) 2026-07-12 09:40:50.950 | DEBUG | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:193 - [generate_music] After VAE decode: allocated=6.15GB, max=7.44GB 2026-07-12 09:40:51.112 | INFO | acestep.core.generation.handler.generate_music_payload:_build_generate_music_success_payload:38 - [generate_music] VAE decode completed. Preparing audio tensors... 2026-07-12 09:40:51.115 | INFO | acestep.core.generation.handler.generate_music_payload:_build_generate_music_success_payload:50 - [generate_music] Done! Generated 1 audio tensors. [Request] Parsed ggml-sft/request0.json [Request] seed=42 lm_batch=1 synth_batch=1 [Request] caption: An upbeat and anthemic pop-rock track driven by bright, slig... [Request] lyrics: 490 bytes [Request] bpm=83 dur=88 key=G major ts=4 lang=fr [Request] lm: temp=0.85 cfg=2.0 top_p=0.90 top_k=0 [Request] dit: steps=50 guidance=1.0 shift=1.0 [Request] task_type: text2music [Request] solver: euler (stork_substeps=10) [Request] lm_mode: generate [Request] output_format: mp3 [Request] synth_model: acestep-v15-sft-Q6_K.gguf [Request] vae: vae-BF16.gguf [Request] audio_codes: (present) [Ace-Synth] Batch: 1 request(s) [Registry] Qwen3-Embedding-0.6B-BF16.gguf -> Text-Enc [Registry] Qwen3-Embedding-0.6B-Q8_0.gguf -> Text-Enc [Registry] acestep-5Hz-lm-0.6B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-0.6B-Q8_0.gguf -> LM [Registry] acestep-5Hz-lm-1.7B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-1.7B-Q8_0.gguf -> LM [Registry] acestep-5Hz-lm-4B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q5_K_M.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q6_K.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q8_0.gguf -> LM [Registry] acestep-v15-base-BF16.gguf -> DiT [Registry] acestep-v15-base-Q4_K_M.gguf -> DiT [Registry] acestep-v15-base-Q5_K_M.gguf -> DiT [Registry] acestep-v15-base-Q6_K.gguf -> DiT [Registry] acestep-v15-base-Q8_0.gguf -> DiT [Registry] acestep-v15-sft-BF16.gguf -> DiT [Registry] acestep-v15-sft-Q4_K_M.gguf -> DiT [Registry] acestep-v15-sft-Q5_K_M.gguf -> DiT [Registry] acestep-v15-sft-Q6_K.gguf -> DiT [Registry] acestep-v15-sft-Q8_0.gguf -> DiT [Registry] acestep-v15-sftturbo50-BF16.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q4_K_M.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q5_K_M.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q6_K.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-BF16.gguf -> DiT [Registry] acestep-v15-turbo-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-continuous-BF16.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-shift1-BF16.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-shift3-BF16.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-base-BF16.gguf -> DiT [Registry] acestep-v15-xl-base-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-base-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-base-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-base-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-sft-BF16.gguf -> DiT [Registry] acestep-v15-xl-sft-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-sft-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-sft-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-sft-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-BF16.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-turbo-BF16.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q8_0.gguf -> DiT [Registry] scragvae-BF16.gguf -> VAE [Registry] vae-BF16.gguf -> VAE [Store] Created (policy=STRICT) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-sft-Q6_K.gguf: 678 tensors, data at offset 56800 [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-sft-Q6_K.gguf: 678 tensors, data at offset 56800 [Synth-Load] Ready: turbo=no, fa=no, batch_cfg=yes [Resolve-Params] max audio codes across batch: 434 (86.8s @ 5Hz) [Resolve-T] T=2170, S=1085 [Resolve-T] seed=42, steps=50, guidance=1.0, shift=1.0, duration=88.0s [BPE] Loaded from GGUF: 151643 vocab, 151387 merges ggml_cuda_init: found 1 CUDA devices (Total VRAM: 97247 MiB): Device 0: NVIDIA RTX PRO 6000 Blackwell Workstation Edition, compute capability 12.0, VMM: yes, VRAM: 97247 MiB load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so ggml_vulkan: Found 1 Vulkan devices: ggml_vulkan: 0 = NVIDIA RTX PRO 6000 Blackwell Workstation Edition (NVIDIA) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 49152 | int dot: 1 | matrix cores: NV_coopmat2 load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] TextEncoder backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/Qwen3-Embedding-0.6B-BF16.gguf: 310 tensors, data at offset 5337568 [Load] TextEncoder: 28L, H=1024, Nh=16/8 [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [WeightCtx] Loaded 310 tensors, 1136.5 MB into backend [Store] Load TextEnc: 221 ms [Debug] text_hidden: [70, 1024] first4: 3.704526 2.436253 0.222853 -13.131872 [Debug] lyric_embed: [167, 1024] first4: 0.029175 0.032227 -0.022339 -0.028809 [Store] Unload TextEnc (1136.5 MB) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] CondEncoder backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-sft-Q6_K.gguf: 678 tensors, data at offset 56800 [Load] LyricEncoder: 8L [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [Load] TimbreEncoder: 4L [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [WeightCtx] Loaded 140 tensors, 476.3 MB into backend [Load] CondEncoder: lyric(8L), timbre(4L), text_proj, null_cond [Store] Load CondEnc: 305 ms [CondEnc] Lyric sliding mask: 167x167, window=128 [CondEnc] Timbre sliding mask: 1x1, window=128 [Encode] Packed: lyric=167 + timbre=1 + text=70 = 238 tokens [Encode-Text Batch0] 70+167 tokens -> enc_S=238, 97.8 ms [Debug] enc_hidden: [238, 2048] first4: 1.762379 -0.051814 -0.131966 0.057631 [Store] Unload CondEnc (476.3 MB) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] Detokenizer backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-sft-Q6_K.gguf: 678 tensors, data at offset 56800 [WeightCtx] Loaded 30 tensors, 82.2 MB into backend [Load] Detokenizer: FSQ(6->2048) + 2L encoder(S=5, 2048->64) [Store] Load FSQ-Detok: 57 ms [Context] Decoded: 434 codes -> 2170 frames (86.8s @ 25Hz) [Build-Context Batch0] Detokenizer: 353.4 ms, 434 codes [Debug] detok_output: [2170, 64] first4: -0.151355 1.462444 0.326907 -0.627213 [Store] Unload FSQ-Detok (82.2 MB) [Init-Noise Batch0] Philox noise seed=42, [2170, 64] solver=euler [Debug] noise: [2170, 64] first4: 0.194336 2.156250 -0.171875 0.847656 [Debug] context: [2170, 128] first4: -0.151355 1.462444 0.326907 -0.627213 [Init-Noise] Starting: T=2170, S=1085, enc_S=238, steps=50, batch=1 (cover) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] DiT backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-sft-Q6_K.gguf: 678 tensors, data at offset 56800 [DiT] Self-attn: Q+K+V fused [DiT] Cross-attn: Q+K+V fused [DiT] MLP: gate+up fused [Load] null_condition_emb found (CFG available) [WeightCtx] Loaded 478 tensors, 1237.2 MB into backend [Load] DiT: 24 layers, H=2048, Nh=16/8, D=128 [Store] Load DiT: 896 ms [DiT] Batch N=1, T=2170, S=1085, enc_S=238 [DiT] Graph: 2129 nodes [DiT] Solver: ODE Euler (1 NFE/step, order 1) [Debug] tproj: [12288] first4: 0.154743 -0.116957 -0.090544 0.080574 [Debug] temb: [2048] first4: -0.002145 -0.176665 0.004243 -0.002255 [Debug] temb_t: [2048] first4: -0.000777 0.003197 -0.013362 -0.002273 [Debug] temb_r: [2048] first4: -0.001369 -0.179863 0.017604 0.000017 [Debug] sinusoidal_t: [256] first4: 0.562379 0.789627 0.439928 -0.023645 [Debug] sinusoidal_r: [256] first4: 1.000000 1.000000 1.000000 1.000000 [Debug] temb_lin1_t: [2048] first4: -0.041318 0.030674 0.028131 -0.025225 [Debug] temb_lin1_r: [2048] first4: 0.004272 0.006720 0.000208 -0.005103 [Debug] hidden_after_proj_in: [2048, 1085] first4: -1.092314 -0.891880 0.504853 0.441786 [Debug] proj_in_input: [192, 2170] first4: -0.151355 1.462444 0.326907 -0.627213 [Debug] enc_after_cond_emb: [2048, 238] first4: -0.193801 0.898684 0.328701 -0.595298 [Debug] layer0_sa_input: [2048, 1085] first4: -0.916144 -0.696019 -0.040361 0.292980 [Debug] layer0_q_after_rope: [128, 16] first4: -1.800593 -1.339359 -0.144539 0.381733 [Debug] layer0_k_after_rope: [128, 8] first4: -0.193801 0.898684 0.328701 -0.595298 [Debug] layer0_sa_output: [2048, 1085] first4: -1.640291 0.793311 -0.575851 0.538407 [Debug] layer0_attn_out: [2048, 1085] first4: -1.800593 -1.339359 -0.144539 0.381733 [Debug] layer0_after_self_attn: [2048, 1085] first4: -1.800593 -1.339359 -0.144539 0.381733 [Debug] layer0_after_cross_attn: [2048, 1085] first4: -1.972986 -1.027412 -0.424850 0.386348 [Debug] hidden_after_layer0: [2048, 1085] first4: -9.390988 1.432513 58.367939 -1.779012 [Debug] hidden_after_layer6: [2048, 1085] first4: -17.649006 2.919505 58.992424 -1.699171 [Debug] hidden_after_layer12: [2048, 1085] first4: -8.396528 6.754469 -18.786980 -4.312500 [Debug] hidden_after_layer18: [2048, 1085] first4: -3.990694 18.373602 -40.167282 -13.808522 [Debug] hidden_after_layer23: [2048, 1085] first4: 33.229965 65.295158 57.146408 -15.140586 [Debug] dit_step0_vt: [2170, 64] first4: -0.613155 2.547581 -0.196654 1.428258 [Debug] dit_step0_xt: [2170, 64] first4: 0.206599 2.105299 -0.167942 0.819091 [DiT] Step 1/50 t=1.000 [Debug] dit_step1_vt: [2170, 64] first4: -0.677913 2.516037 -0.081329 1.264512 [Debug] dit_step1_xt: [2170, 64] first4: 0.220157 2.054978 -0.166315 0.793801 [DiT] Step 2/50 t=0.980 [Debug] dit_step2_vt: [2170, 64] first4: -0.686019 2.497601 0.016064 1.292563 [Debug] dit_step2_xt: [2170, 64] first4: 0.233878 2.005026 -0.166637 0.767950 [DiT] Step 3/50 t=0.960 [Debug] dit_step3_vt: [2170, 64] first4: -0.686576 2.438186 0.013939 1.312539 [Debug] dit_step3_xt: [2170, 64] first4: 0.247609 1.956262 -0.166915 0.741699 [DiT] Step 4/50 t=0.940 [Debug] dit_step4_vt: [2170, 64] first4: -0.595693 2.333536 0.020800 1.354952 [Debug] dit_step4_xt: [2170, 64] first4: 0.259523 1.909591 -0.167331 0.714600 [DiT] Step 5/50 t=0.920 [Debug] dit_step5_vt: [2170, 64] first4: -0.538042 2.306288 -0.058144 1.307326 [Debug] dit_step5_xt: [2170, 64] first4: 0.270284 1.863465 -0.166168 0.688453 [DiT] Step 6/50 t=0.900 [Debug] dit_step6_vt: [2170, 64] first4: -0.577163 2.249364 -0.075185 1.291071 [Debug] dit_step6_xt: [2170, 64] first4: 0.281827 1.818478 -0.164665 0.662632 [DiT] Step 7/50 t=0.880 [Debug] dit_step7_vt: [2170, 64] first4: -0.491654 2.161266 -0.084423 1.281742 [Debug] dit_step7_xt: [2170, 64] first4: 0.291660 1.775253 -0.162976 0.636997 [DiT] Step 8/50 t=0.860 [Debug] dit_step8_vt: [2170, 64] first4: -0.448943 2.076568 -0.103649 1.254388 [Debug] dit_step8_xt: [2170, 64] first4: 0.300639 1.733721 -0.160903 0.611909 [DiT] Step 9/50 t=0.840 [Debug] dit_step9_vt: [2170, 64] first4: -0.388061 1.968806 -0.085346 1.243821 [Debug] dit_step9_xt: [2170, 64] first4: 0.308400 1.694345 -0.159196 0.587033 [DiT] Step 10/50 t=0.820 [Debug] dit_step10_vt: [2170, 64] first4: -0.336021 1.881550 -0.101911 1.252710 [Debug] dit_step10_xt: [2170, 64] first4: 0.315121 1.656714 -0.157158 0.561979 [DiT] Step 11/50 t=0.800 [Debug] dit_step11_vt: [2170, 64] first4: -0.315319 1.841578 -0.050343 1.209821 [Debug] dit_step11_xt: [2170, 64] first4: 0.321427 1.619883 -0.156151 0.537782 [DiT] Step 12/50 t=0.780 [Debug] dit_step12_vt: [2170, 64] first4: -0.277157 1.740461 -0.095502 1.215509 [Debug] dit_step12_xt: [2170, 64] first4: 0.326970 1.585074 -0.154241 0.513472 [DiT] Step 13/50 t=0.760 [Debug] dit_step13_vt: [2170, 64] first4: -0.245912 1.634442 -0.111691 1.196837 [Debug] dit_step13_xt: [2170, 64] first4: 0.331888 1.552385 -0.152008 0.489535 [DiT] Step 14/50 t=0.740 [Debug] dit_step14_vt: [2170, 64] first4: -0.260961 1.577526 -0.137106 1.172146 [Debug] dit_step14_xt: [2170, 64] first4: 0.337108 1.520834 -0.149265 0.466092 [DiT] Step 15/50 t=0.720 [Debug] dit_step15_vt: [2170, 64] first4: -0.240335 1.489366 -0.157901 1.149845 [Debug] dit_step15_xt: [2170, 64] first4: 0.341914 1.491047 -0.146107 0.443095 [DiT] Step 16/50 t=0.700 [Debug] dit_step16_vt: [2170, 64] first4: -0.212750 1.414127 -0.180037 1.136789 [Debug] dit_step16_xt: [2170, 64] first4: 0.346169 1.462764 -0.142507 0.420360 [DiT] Step 17/50 t=0.680 [Debug] dit_step17_vt: [2170, 64] first4: -0.243286 1.357383 -0.200700 1.102206 [Debug] dit_step17_xt: [2170, 64] first4: 0.351035 1.435617 -0.138493 0.398316 [DiT] Step 18/50 t=0.660 [Debug] dit_step18_vt: [2170, 64] first4: -0.243793 1.256381 -0.244863 1.082416 [Debug] dit_step18_xt: [2170, 64] first4: 0.355911 1.410489 -0.133595 0.376667 [DiT] Step 19/50 t=0.640 [Debug] dit_step19_vt: [2170, 64] first4: -0.278751 1.153520 -0.285861 1.082474 [Debug] dit_step19_xt: [2170, 64] first4: 0.361486 1.387419 -0.127878 0.355018 [DiT] Step 20/50 t=0.620 [Debug] dit_step20_vt: [2170, 64] first4: -0.288792 1.097326 -0.270182 1.043576 [Debug] dit_step20_xt: [2170, 64] first4: 0.367262 1.365472 -0.122475 0.334146 [DiT] Step 21/50 t=0.600 [Debug] dit_step21_vt: [2170, 64] first4: -0.298348 0.985489 -0.254788 0.985212 [Debug] dit_step21_xt: [2170, 64] first4: 0.373229 1.345762 -0.117379 0.314442 [DiT] Step 22/50 t=0.580 [Debug] dit_step22_vt: [2170, 64] first4: -0.345301 0.889261 -0.301036 0.974880 [Debug] dit_step22_xt: [2170, 64] first4: 0.380135 1.327977 -0.111358 0.294944 [DiT] Step 23/50 t=0.560 [Debug] dit_step23_vt: [2170, 64] first4: -0.310827 0.786733 -0.277415 0.936082 [Debug] dit_step23_xt: [2170, 64] first4: 0.386351 1.312242 -0.105810 0.276223 [DiT] Step 24/50 t=0.540 [Debug] dit_step24_vt: [2170, 64] first4: -0.360205 0.727257 -0.267813 0.911230 [Debug] dit_step24_xt: [2170, 64] first4: 0.393556 1.297697 -0.100453 0.257998 [DiT] Step 25/50 t=0.520 [Debug] dit_step25_vt: [2170, 64] first4: -0.352901 0.557441 -0.277952 0.896063 [Debug] dit_step25_xt: [2170, 64] first4: 0.400614 1.286548 -0.094894 0.240077 [DiT] Step 26/50 t=0.500 [Debug] dit_step26_vt: [2170, 64] first4: -0.402939 0.448277 -0.323576 0.879598 [Debug] dit_step26_xt: [2170, 64] first4: 0.408672 1.277583 -0.088423 0.222485 [DiT] Step 27/50 t=0.480 [Debug] dit_step27_vt: [2170, 64] first4: -0.439950 0.363509 -0.276971 0.881568 [Debug] dit_step27_xt: [2170, 64] first4: 0.417471 1.270313 -0.082884 0.204854 [DiT] Step 28/50 t=0.460 [Debug] dit_step28_vt: [2170, 64] first4: -0.450573 0.257335 -0.222470 0.869812 [Debug] dit_step28_xt: [2170, 64] first4: 0.426483 1.265166 -0.078434 0.187457 [DiT] Step 29/50 t=0.440 [Debug] dit_step29_vt: [2170, 64] first4: -0.468659 0.130336 -0.248406 0.875034 [Debug] dit_step29_xt: [2170, 64] first4: 0.435856 1.262559 -0.073466 0.169957 [DiT] Step 30/50 t=0.420 [Debug] dit_step30_vt: [2170, 64] first4: -0.474624 0.061799 -0.211201 0.803282 [Debug] dit_step30_xt: [2170, 64] first4: 0.445348 1.261323 -0.069242 0.153891 [DiT] Step 31/50 t=0.400 [Debug] dit_step31_vt: [2170, 64] first4: -0.511891 -0.091393 -0.236198 0.817607 [Debug] dit_step31_xt: [2170, 64] first4: 0.455586 1.263151 -0.064518 0.137539 [DiT] Step 32/50 t=0.380 [Debug] dit_step32_vt: [2170, 64] first4: -0.528051 -0.175315 -0.229105 0.816554 [Debug] dit_step32_xt: [2170, 64] first4: 0.466147 1.266657 -0.059936 0.121208 [DiT] Step 33/50 t=0.360 [Debug] dit_step33_vt: [2170, 64] first4: -0.538468 -0.265782 -0.210837 0.789438 [Debug] dit_step33_xt: [2170, 64] first4: 0.476917 1.271973 -0.055719 0.105419 [DiT] Step 34/50 t=0.340 [Debug] dit_step34_vt: [2170, 64] first4: -0.621642 -0.397833 -0.196285 0.774127 [Debug] dit_step34_xt: [2170, 64] first4: 0.489349 1.279930 -0.051793 0.089937 [DiT] Step 35/50 t=0.320 [Debug] dit_step35_vt: [2170, 64] first4: -0.571940 -0.527526 -0.205302 0.736066 [Debug] dit_step35_xt: [2170, 64] first4: 0.500788 1.290480 -0.047687 0.075215 [DiT] Step 36/50 t=0.300 [Debug] dit_step36_vt: [2170, 64] first4: -0.592785 -0.596799 -0.153226 0.755388 [Debug] dit_step36_xt: [2170, 64] first4: 0.512644 1.302416 -0.044623 0.060107 [DiT] Step 37/50 t=0.280 [Debug] dit_step37_vt: [2170, 64] first4: -0.617620 -0.665467 -0.146912 0.652840 [Debug] dit_step37_xt: [2170, 64] first4: 0.524996 1.315725 -0.041685 0.047051 [DiT] Step 38/50 t=0.260 [Debug] dit_step38_vt: [2170, 64] first4: -0.594819 -0.720013 -0.154671 0.641198 [Debug] dit_step38_xt: [2170, 64] first4: 0.536893 1.330126 -0.038591 0.034227 [DiT] Step 39/50 t=0.240 [Debug] dit_step39_vt: [2170, 64] first4: -0.535285 -0.850494 -0.105562 0.627716 [Debug] dit_step39_xt: [2170, 64] first4: 0.547598 1.347136 -0.036480 0.021672 [DiT] Step 40/50 t=0.220 [Debug] dit_step40_vt: [2170, 64] first4: -0.573615 -0.927160 -0.097847 0.542948 [Debug] dit_step40_xt: [2170, 64] first4: 0.559071 1.365679 -0.034523 0.010813 [DiT] Step 41/50 t=0.200 [Debug] dit_step41_vt: [2170, 64] first4: -0.567948 -0.995886 0.006643 0.517907 [Debug] dit_step41_xt: [2170, 64] first4: 0.570430 1.385596 -0.034656 0.000455 [DiT] Step 42/50 t=0.180 [Debug] dit_step42_vt: [2170, 64] first4: -0.523054 -1.044437 0.033272 0.484856 [Debug] dit_step42_xt: [2170, 64] first4: 0.580891 1.406485 -0.035321 -0.009242 [DiT] Step 43/50 t=0.160 [Debug] dit_step43_vt: [2170, 64] first4: -0.525441 -1.076839 0.049187 0.470658 [Debug] dit_step43_xt: [2170, 64] first4: 0.591400 1.428022 -0.036305 -0.018655 [DiT] Step 44/50 t=0.140 [Debug] dit_step44_vt: [2170, 64] first4: -0.522604 -1.123449 0.068296 0.394463 [Debug] dit_step44_xt: [2170, 64] first4: 0.601852 1.450491 -0.037671 -0.026544 [DiT] Step 45/50 t=0.120 [Debug] dit_step45_vt: [2170, 64] first4: -0.475719 -1.181196 0.175192 0.354841 [Debug] dit_step45_xt: [2170, 64] first4: 0.611366 1.474115 -0.041175 -0.033641 [DiT] Step 46/50 t=0.100 [Debug] dit_step46_vt: [2170, 64] first4: -0.495847 -1.243702 0.220221 0.279768 [Debug] dit_step46_xt: [2170, 64] first4: 0.621283 1.498989 -0.045579 -0.039236 [DiT] Step 47/50 t=0.080 [Debug] dit_step47_vt: [2170, 64] first4: -0.452964 -1.285210 0.188186 0.253003 [Debug] dit_step47_xt: [2170, 64] first4: 0.630342 1.524693 -0.049343 -0.044296 [DiT] Step 48/50 t=0.060 [Debug] dit_step48_vt: [2170, 64] first4: -0.399792 -1.320184 0.233392 0.169236 [Debug] dit_step48_xt: [2170, 64] first4: 0.638338 1.551097 -0.054011 -0.047681 [DiT] Step 49/50 t=0.040 [Debug] dit_step49_vt: [2170, 64] first4: -0.422784 -1.348173 0.195202 0.184336 [Debug] dit_x0: [2170, 64] first4: 0.646794 1.578060 -0.057915 -0.051368 [DiT] Step 50/50 t=0.020 [DiT-Generate] Total: 137099.5 ms (137099.5 ms/sample) [Debug] dit_output: [2170, 64] first4: 0.646794 1.578060 -0.057915 -0.051368 [Store] Unload DiT (1237.2 MB) [GGUF] /mnt/workspace/git/acestep.cpp/models/vae-BF16.gguf: 365 tensors, data at offset 30048 load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] VAE backend: CPU (CPU threads: 16) [VAE] Backend: CPU, Weight buffer: 161.1 MB [VAE] Loaded: 5 blocks, upsample=1920x, F32 activations [Store] Load VAE-Dec: 712 ms [VAE] Tiled decode: 3 tiles (chunk=1024, overlap=64, stride=896) [VAE] Graph: 479 nodes, T_latent=960 [VAE] Upsample factor: 1920.00 (expected ~1920) [VAE] Graph: 479 nodes, T_latent=1024 [VAE] Graph: 479 nodes, T_latent=442 [VAE] Tiled decode done: 3 tiles -> T_audio=4166400 (86.80s @ 48kHz) [VAE-Decode Batch0] Decode: 32186.6 ms [Debug] vae_audio: [2, 4166400] first4: -0.001599 -0.001103 -0.001384 -0.000975 [Store] Unload VAE-Dec (161.1 MB) [MP3] Encoding 86.8s @ 128 kbps, 48000 Hz stereo [MP3] 1388928 bytes (12.0:1), 431 ms (201.28x realtime), 32 threads [MP3] Wrote ggml-sft/request00.mp3 [Ace-Synth] All done 2026-07-12 09:43:44.548 | INFO | acestep.core.generation.handler.init_service_loader:_load_main_model_from_checkpoint:174 - [initialize_service] Attempting to load model with attention implementation: sdpa 2026-07-12 09:43:46.352 | INFO | acestep.core.generation.handler.generate_music:generate_music:304 - [generate_music] Starting generation... 2026-07-12 09:43:46.352 | INFO | acestep.core.generation.handler.generate_music:generate_music:307 - [generate_music] Preparing inputs... 2026-07-12 09:43:46.353 | INFO | acestep.core.generation.handler.generate_music_request:_prepare_reference_and_source_audio:179 - [generate_music] cover task: no src_audio but audio codes provided, proceeding with codes 2026-07-12 09:43:46.355 | INFO | acestep.core.generation.handler.generate_music:_vram_preflight_check:156 - [generate_music] VRAM pre-flight: 81.68 GB free, ~0.94 GB needed (batch=1, duration=88s, mode=turbo). 2026-07-12 09:43:46.360 | INFO | acestep.core.generation.handler.conditioning_target:_prepare_target_latents_and_wavs:78 - [generate_music] Decoding audio codes for item 0... 2026-07-12 09:43:46.365 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_precomputed_lm_hints:31 - [generate_music] Decoding audio codes for LM hints for item 0... 2026-07-12 09:43:46.366 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:119 - ====================================================================== 2026-07-12 09:43:46.366 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:120 - πŸ” [DEBUG] DiT TEXT ENCODER INPUT (Inference) 2026-07-12 09:43:46.366 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:121 - ====================================================================== 2026-07-12 09:43:46.366 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:122 - text_prompt: # Instruction Generate audio semantic tokens based on the given conditions: # Caption An upbeat and anthemic pop-rock track driven by bright, slightly overdriven # Metas - bpm: 83 - timesignature: 4 - keyscale: G major - duration: 88 seconds <|endoftext|> 2026-07-12 09:43:46.366 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:123 - ====================================================================== 2026-07-12 09:43:46.366 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:124 - lyrics_text: # Languages fr # Lyric # Lyric [Intro - Guitar Riff] [Verse 1] Dans le monde des tutos virtuels G ta toise en nouvelle passion Avec Ggendoline et PumbΓ© Γ  midi La communautΓ©, c'est l'unitΓ© Quel joie, une clΓ© [Chorus] Dans le monde des tutos virtuels GΓ’ndoline et PumbΓ© Γ  midi Une famille Γ  connecter, c'est vrai D'un enfant qui voit toi fusionner [Guitar Solo] [Verse 2] Dans le monde des tutos virtuels GΓ’ndoline, PumbΓ© Γ  midi Une famille Γ  connecter, c'est vrai D'un enfant qui voit toi fusionner<|endoftext|> 2026-07-12 09:43:46.366 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:125 - ====================================================================== 2026-07-12 09:43:46.370 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:110 - [preprocess_batch] Inferring prompt embeddings... 2026-07-12 09:43:46.378 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:113 - [preprocess_batch] Inferring lyric embeddings... 2026-07-12 09:43:46.378 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:161 - [service_generate] Generating audio... (DiT backend: PyTorch (cuda)) 2026-07-12 09:43:46.384 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:263 - [service_generate] DiT diffusion via PyTorch (cuda)... [Request] Loaded request0.json [Turbo] steps=8, shift=3.0 | acestep-v15-turbo-Q6_K.gguf [GGML] Running acestep-v15-turbo-Q6_K.gguf... [GGML] Done, 47 dump files [Python] Initializing acestep-v15-turbo... [Python] Generating (acestep-v15-turbo, 8 steps)... Using precomputed LM hints Using precomputed LM hints [Python] Wrote python-turbo/output.wav: 4166400 samples (86.80s @ 48kHz stereo) [Python] Done, 40 dump files [Turbo] Cosine similarities GGML vs Python stage GGML vs Python text_hidden 0.999810 lyric_embed 1.000000 enc_hidden 0.999627 detok_output 0.999927 context 0.999954 noise 1.000000 temb_t 0.999986 hidden_after_proj_in 0.999975 enc_after_cond_emb 0.999615 layer0_sa_output 0.999718 hidden_after_layer0 0.999825 hidden_after_layer6 0.999788 hidden_after_layer12 0.998912 hidden_after_layer18 0.995926 hidden_after_layer23 0.992285 dit_step0_vt 0.971194 dit_step0_xt 0.999936 dit_step1_vt 0.974424 dit_step1_xt 0.999799 dit_step2_vt 0.977923 dit_step2_xt 0.999471 dit_step3_vt 0.977600 dit_step3_xt 0.998717 dit_step4_vt 0.973239 dit_step4_xt 0.996917 dit_step5_vt 0.968741 dit_step5_xt 0.992644 dit_step6_vt 0.964555 dit_step6_xt 0.984241 dit_step7_vt 0.956646 dit_x0 0.971565 vae_audio 0.889755 vae_audio (STFT cosine) 0.967968 [Turbo] Error growth GGML vs Python stage cos max_err mean_err mean_A std_A mean_B std_B dit_step0_xt 0.999936 0.148925 0.007243 -0.002255 0.972904 -0.002342 0.972003 dit_step1_xt 0.999799 0.293564 0.012580 -0.005196 0.942625 -0.005313 0.941730 dit_step2_xt 0.999471 0.475913 0.019361 -0.009152 0.909045 -0.009311 0.908527 dit_step3_xt 0.998717 0.731060 0.028859 -0.014498 0.873512 -0.014577 0.873624 dit_step4_xt 0.996917 1.059091 0.042926 -0.021663 0.841485 -0.021660 0.841995 dit_step5_xt 0.992644 1.536603 0.064512 -0.031963 0.824626 -0.032109 0.824593 dit_step6_xt 0.984241 2.192945 0.097266 -0.046555 0.855647 -0.046482 0.855546 [SFT] steps=50, shift=1.0 | acestep-v15-sft-Q6_K.gguf [GGML] Running acestep-v15-sft-Q6_K.gguf... [GGML] Done, 131 dump files [Python] Initializing acestep-v15-sft... [Python] Generating (acestep-v15-sft, 50 steps)... Using precomputed LM hints Using precomputed LM hints 0%| | 0/50 [00:00 Text-Enc [Registry] Qwen3-Embedding-0.6B-Q8_0.gguf -> Text-Enc [Registry] acestep-5Hz-lm-0.6B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-0.6B-Q8_0.gguf -> LM [Registry] acestep-5Hz-lm-1.7B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-1.7B-Q8_0.gguf -> LM [Registry] acestep-5Hz-lm-4B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q5_K_M.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q6_K.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q8_0.gguf -> LM [Registry] acestep-v15-base-BF16.gguf -> DiT [Registry] acestep-v15-base-Q4_K_M.gguf -> DiT [Registry] acestep-v15-base-Q5_K_M.gguf -> DiT [Registry] acestep-v15-base-Q6_K.gguf -> DiT [Registry] acestep-v15-base-Q8_0.gguf -> DiT [Registry] acestep-v15-sft-BF16.gguf -> DiT [Registry] acestep-v15-sft-Q4_K_M.gguf -> DiT [Registry] acestep-v15-sft-Q5_K_M.gguf -> DiT [Registry] acestep-v15-sft-Q6_K.gguf -> DiT [Registry] acestep-v15-sft-Q8_0.gguf -> DiT [Registry] acestep-v15-sftturbo50-BF16.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q4_K_M.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q5_K_M.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q6_K.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-BF16.gguf -> DiT [Registry] acestep-v15-turbo-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-continuous-BF16.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-shift1-BF16.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-shift3-BF16.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-base-BF16.gguf -> DiT [Registry] acestep-v15-xl-base-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-base-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-base-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-base-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-sft-BF16.gguf -> DiT [Registry] acestep-v15-xl-sft-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-sft-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-sft-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-sft-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-BF16.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-turbo-BF16.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q8_0.gguf -> DiT [Registry] scragvae-BF16.gguf -> VAE [Registry] vae-BF16.gguf -> VAE [Store] Created (policy=STRICT) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-turbo-Q6_K.gguf: 830 tensors, data at offset 69088 [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-turbo-Q6_K.gguf: 830 tensors, data at offset 69088 [Synth-Load] Ready: turbo=yes, fa=no, batch_cfg=yes [Resolve-Params] max audio codes across batch: 434 (86.8s @ 5Hz) [Resolve-T] T=2170, S=1085 [Resolve-T] seed=42, steps=8, guidance=1.0, shift=3.0, duration=88.0s [BPE] Loaded from GGUF: 151643 vocab, 151387 merges ggml_cuda_init: found 1 CUDA devices (Total VRAM: 97247 MiB): Device 0: NVIDIA RTX PRO 6000 Blackwell Workstation Edition, compute capability 12.0, VMM: yes, VRAM: 97247 MiB load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so ggml_vulkan: Found 1 Vulkan devices: ggml_vulkan: 0 = NVIDIA RTX PRO 6000 Blackwell Workstation Edition (NVIDIA) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 49152 | int dot: 1 | matrix cores: NV_coopmat2 load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] TextEncoder backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/Qwen3-Embedding-0.6B-BF16.gguf: 310 tensors, data at offset 5337568 [Load] TextEncoder: 28L, H=1024, Nh=16/8 [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [WeightCtx] Loaded 310 tensors, 1136.5 MB into backend [Store] Load TextEnc: 222 ms [Debug] text_hidden: [70, 1024] first4: 3.704526 2.436253 0.222853 -13.131872 [Debug] lyric_embed: [167, 1024] first4: 0.029175 0.032227 -0.022339 -0.028809 [Store] Unload TextEnc (1136.5 MB) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] CondEncoder backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-turbo-Q6_K.gguf: 830 tensors, data at offset 69088 [Load] LyricEncoder: 8L [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [Load] TimbreEncoder: 4L [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [WeightCtx] Loaded 141 tensors, 476.3 MB into backend [Load] CondEncoder: lyric(8L), timbre(4L, CLS), text_proj, null_cond [Store] Load CondEnc: 306 ms [CondEnc] Lyric sliding mask: 167x167, window=128 [CondEnc] Timbre sliding mask: 2x2, window=128 (CLS) [Encode] Packed: lyric=167 + timbre=1 + text=70 = 238 tokens [Encode-Text Batch0] 70+167 tokens -> enc_S=238, 95.4 ms [Debug] enc_hidden: [238, 2048] first4: 0.687624 -0.020412 -0.050765 0.022580 [Store] Unload CondEnc (476.3 MB) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] Detokenizer backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-turbo-Q6_K.gguf: 830 tensors, data at offset 69088 [WeightCtx] Loaded 30 tensors, 82.2 MB into backend [Load] Detokenizer: FSQ(6->2048) + 2L encoder(S=5, 2048->64) [Store] Load FSQ-Detok: 49 ms [Context] Decoded: 434 codes -> 2170 frames (86.8s @ 25Hz) [Build-Context Batch0] Detokenizer: 361.2 ms, 434 codes [Debug] detok_output: [2170, 64] first4: -0.151355 1.462444 0.326907 -0.627213 [Store] Unload FSQ-Detok (82.2 MB) [Init-Noise Batch0] Philox noise seed=42, [2170, 64] solver=euler [Debug] noise: [2170, 64] first4: 0.194336 2.156250 -0.171875 0.847656 [Debug] context: [2170, 128] first4: -0.151355 1.462444 0.326907 -0.627213 [Init-Noise] Starting: T=2170, S=1085, enc_S=238, steps=8, batch=1 (cover) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] DiT backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-turbo-Q6_K.gguf: 830 tensors, data at offset 69088 [DiT] Self-attn: Q+K+V fused [DiT] Cross-attn: Q+K+V fused [DiT] MLP: gate+up fused [Load] null_condition_emb found (CFG available) [WeightCtx] Loaded 630 tensors, 3267.8 MB into backend [Load] DiT: 32 layers, H=2560, Nh=32/8, D=128 [Store] Load DiT: 2390 ms [DiT] Batch N=1, T=2170, S=1085, enc_S=238 [DiT] Graph: 2825 nodes [DiT] Solver: ODE Euler (1 NFE/step, order 1) [Debug] tproj: [15360] first4: 0.018624 0.109874 0.060427 -0.073897 [Debug] temb: [2560] first4: 0.026456 -0.022590 -0.016414 0.047046 [Debug] temb_t: [2560] first4: 0.014595 -0.022043 -0.020712 0.048631 [Debug] temb_r: [2560] first4: 0.011861 -0.000548 0.004298 -0.001585 [Debug] sinusoidal_t: [256] first4: 0.562379 0.789627 0.439928 -0.023645 [Debug] sinusoidal_r: [256] first4: 1.000000 1.000000 1.000000 1.000000 [Debug] temb_lin1_t: [2560] first4: 0.062874 -0.017498 0.137456 -0.111371 [Debug] temb_lin1_r: [2560] first4: -0.011992 0.015308 -0.017417 0.011656 [Debug] hidden_after_proj_in: [2560, 1085] first4: 0.343898 -0.584744 0.251996 -0.196074 [Debug] proj_in_input: [192, 2170] first4: -0.151355 1.462444 0.326907 -0.627213 [Debug] enc_after_cond_emb: [2560, 238] first4: 0.038852 0.672015 0.109741 0.865113 [Debug] layer0_sa_input: [2560, 1085] first4: 0.042658 -0.046178 0.089744 -0.136220 [Debug] layer0_q_after_rope: [128, 32] first4: -0.252196 -2.086671 0.334540 0.257373 [Debug] layer0_k_after_rope: [128, 8] first4: 0.045117 2.257508 0.595099 0.560088 [Debug] layer0_sa_output: [2560, 1085] first4: 1.651073 4.306114 -0.571291 5.315104 [Debug] layer0_attn_out: [4096, 1085] first4: 0.477567 -2.248265 -0.164023 -2.760866 [Debug] layer0_after_self_attn: [2560, 1085] first4: 0.477567 -2.248265 -0.164023 -2.760866 [Debug] layer0_after_cross_attn: [2560, 1085] first4: 0.348397 -2.015417 0.012216 -3.753975 [Debug] hidden_after_layer0: [2560, 1085] first4: 1.754470 -2.430678 236.301865 -3.720348 [Debug] hidden_after_layer6: [2560, 1085] first4: -6.237708 -3.360382 314.148407 -12.385801 [Debug] hidden_after_layer12: [2560, 1085] first4: -9.822474 30.221588 255.693390 31.853754 [Debug] hidden_after_layer18: [2560, 1085] first4: 42.540726 9.600193 272.448303 6.007025 [Debug] hidden_after_layer31: [2560, 1085] first4: -45.973820 -257.666687 -114.022362 190.562195 [Debug] dit_step0_vt: [2170, 64] first4: -0.989399 1.727090 0.880368 1.221861 [Debug] dit_step0_xt: [2170, 64] first4: 0.239309 2.077746 -0.211892 0.792117 [DiT] Step 1/8 t=1.000 [Debug] dit_step1_vt: [2170, 64] first4: -0.868939 2.387661 0.733166 1.260771 [Debug] dit_step1_xt: [2170, 64] first4: 0.286705 1.947510 -0.251883 0.723348 [DiT] Step 2/8 t=0.955 [Debug] dit_step2_vt: [2170, 64] first4: -0.859114 1.441048 1.028573 1.007871 [Debug] dit_step2_xt: [2170, 64] first4: 0.343980 1.851440 -0.320454 0.656156 [DiT] Step 3/8 t=0.900 [Debug] dit_step3_vt: [2170, 64] first4: -0.330849 1.391282 0.968648 0.994241 [Debug] dit_step3_xt: [2170, 64] first4: 0.371550 1.735500 -0.401175 0.573303 [DiT] Step 4/8 t=0.833 [Debug] dit_step4_vt: [2170, 64] first4: -1.684619 2.501694 0.988279 0.393451 [Debug] dit_step4_xt: [2170, 64] first4: 0.552045 1.467461 -0.507062 0.531147 [DiT] Step 5/8 t=0.750 [Debug] dit_step5_vt: [2170, 64] first4: 0.530749 2.165804 1.245435 0.790729 [Debug] dit_step5_xt: [2170, 64] first4: 0.476224 1.158061 -0.684981 0.418186 [DiT] Step 6/8 t=0.643 [Debug] dit_step6_vt: [2170, 64] first4: 0.799476 1.901758 0.526168 0.763218 [Debug] dit_step6_xt: [2170, 64] first4: 0.316329 0.777709 -0.790215 0.265543 [DiT] Step 7/8 t=0.500 [Debug] dit_step7_vt: [2170, 64] first4: -0.909951 2.714055 1.416051 0.718708 [Debug] dit_x0: [2170, 64] first4: 0.589314 -0.036508 -1.215030 0.049930 [DiT] Step 8/8 t=0.300 [DiT-Generate] Total: 58561.8 ms (58561.8 ms/sample) [Debug] dit_output: [2170, 64] first4: 0.589314 -0.036508 -1.215030 0.049930 [Store] Unload DiT (3267.8 MB) [GGUF] /mnt/workspace/git/acestep.cpp/models/vae-BF16.gguf: 365 tensors, data at offset 30048 load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] VAE backend: CPU (CPU threads: 16) [VAE] Backend: CPU, Weight buffer: 161.1 MB [VAE] Loaded: 5 blocks, upsample=1920x, F32 activations [Store] Load VAE-Dec: 719 ms [VAE] Tiled decode: 3 tiles (chunk=1024, overlap=64, stride=896) [VAE] Graph: 479 nodes, T_latent=960 [VAE] Upsample factor: 1920.00 (expected ~1920) [VAE] Graph: 479 nodes, T_latent=1024 [VAE] Graph: 479 nodes, T_latent=442 [VAE] Tiled decode done: 3 tiles -> T_audio=4166400 (86.80s @ 48kHz) [VAE-Decode Batch0] Decode: 33499.0 ms [Debug] vae_audio: [2, 4166400] first4: 0.004886 0.007499 0.007158 0.006476 [Store] Unload VAE-Dec (161.1 MB) [MP3] Encoding 86.8s @ 128 kbps, 48000 Hz stereo [MP3] 1388928 bytes (12.0:1), 497 ms (174.79x realtime), 32 threads [MP3] Wrote ggml-xl-turbo/request00.mp3 [Ace-Synth] All done 2026-07-12 09:45:26.465 | INFO | acestep.core.generation.handler.init_service_loader:_load_main_model_from_checkpoint:174 - [initialize_service] Attempting to load model with attention implementation: sdpa [Python] Wrote python-sft/output.wav: 4166400 samples (86.80s @ 48kHz stereo) [Python] Done, 166 dump files [SFT] Cosine similarities GGML vs Python stage GGML vs Python text_hidden 0.999810 lyric_embed 1.000000 enc_hidden 0.999627 detok_output 0.999927 context 0.999954 noise 1.000000 temb_t 0.999952 hidden_after_proj_in 0.999974 enc_after_cond_emb 0.999618 layer0_sa_output 0.999729 hidden_after_layer0 0.999852 hidden_after_layer6 0.999753 hidden_after_layer12 0.999301 hidden_after_layer18 0.998453 hidden_after_layer23 0.998713 dit_step0_vt 0.998620 dit_step0_xt 1.000000 dit_step5_vt 0.999134 dit_step5_xt 0.999974 dit_step10_vt 0.998523 dit_step10_xt 0.999911 dit_step15_vt 0.996987 dit_step15_xt 0.999766 dit_step20_vt 0.994493 dit_step20_xt 0.999403 dit_step25_vt 0.990333 dit_step25_xt 0.998690 dit_step30_vt 0.985274 dit_step30_xt 0.997453 dit_step35_vt 0.979859 dit_step35_xt 0.995810 dit_step40_vt 0.976408 dit_step40_xt 0.994100 dit_step45_vt 0.980266 dit_step45_xt 0.992863 dit_step49_vt 0.986685 dit_x0 0.992434 vae_audio 0.956546 vae_audio (STFT cosine) 0.980987 [SFT] Error growth GGML vs Python stage cos max_err mean_err mean_A std_A mean_B std_B dit_step0_xt 1.000000 0.019246 0.001510 -0.001664 0.978610 -0.001619 0.978872 dit_step5_xt 0.999974 0.082588 0.005102 -0.008383 0.885580 -0.008480 0.883811 dit_step10_xt 0.999911 0.102561 0.008142 -0.015045 0.801936 -0.015075 0.802324 dit_step15_xt 0.999766 0.138471 0.012115 -0.021626 0.731622 -0.021933 0.731304 dit_step20_xt 0.999403 0.219700 0.017741 -0.028216 0.679038 -0.028882 0.679066 dit_step25_xt 0.998690 0.369784 0.024622 -0.034858 0.648634 -0.035643 0.650780 dit_step30_xt 0.997453 0.568217 0.033513 -0.041517 0.643547 -0.042577 0.647107 dit_step35_xt 0.995810 0.829981 0.043971 -0.048165 0.664549 -0.049471 0.669256 dit_step40_xt 0.994100 1.092357 0.055584 -0.054832 0.709226 -0.056532 0.715783 dit_step45_xt 0.992863 1.325922 0.066603 -0.061580 0.773328 -0.063533 0.780575 [Xl-turbo] steps=8, shift=3.0 | acestep-v15-xl-turbo-Q6_K.gguf [GGML] Running acestep-v15-xl-turbo-Q6_K.gguf... [GGML] Done, 47 dump files [Python] Initializing acestep-v15-xl-turbo... Loading checkpoint shards: 0%| | 0/4 [00:00 1.0 (turbo does not use CFG). 2026-07-12 09:45:31.159 | INFO | acestep.core.generation.handler.generate_music:generate_music:304 - [generate_music] Starting generation... 2026-07-12 09:45:31.159 | INFO | acestep.core.generation.handler.generate_music:generate_music:307 - [generate_music] Preparing inputs... 2026-07-12 09:45:31.160 | INFO | acestep.core.generation.handler.generate_music_request:_prepare_reference_and_source_audio:179 - [generate_music] cover task: no src_audio but audio codes provided, proceeding with codes 2026-07-12 09:45:31.166 | INFO | acestep.core.generation.handler.generate_music:_vram_preflight_check:156 - [generate_music] VRAM pre-flight: 76.75 GB free, ~1.23 GB needed (batch=1, duration=88s, mode=xl_turbo). 2026-07-12 09:45:31.170 | INFO | acestep.core.generation.handler.conditioning_target:_prepare_target_latents_and_wavs:78 - [generate_music] Decoding audio codes for item 0... 2026-07-12 09:45:31.175 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_precomputed_lm_hints:31 - [generate_music] Decoding audio codes for LM hints for item 0... 2026-07-12 09:45:31.177 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:119 - ====================================================================== 2026-07-12 09:45:31.177 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:120 - πŸ” [DEBUG] DiT TEXT ENCODER INPUT (Inference) 2026-07-12 09:45:31.177 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:121 - ====================================================================== 2026-07-12 09:45:31.177 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:122 - text_prompt: # Instruction Generate audio semantic tokens based on the given conditions: # Caption An upbeat and anthemic pop-rock track driven by bright, slightly overdriven # Metas - bpm: 83 - timesignature: 4 - keyscale: G major - duration: 88 seconds <|endoftext|> 2026-07-12 09:45:31.177 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:123 - ====================================================================== 2026-07-12 09:45:31.177 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:124 - lyrics_text: # Languages fr # Lyric # Lyric [Intro - Guitar Riff] [Verse 1] Dans le monde des tutos virtuels G ta toise en nouvelle passion Avec Ggendoline et PumbΓ© Γ  midi La communautΓ©, c'est l'unitΓ© Quel joie, une clΓ© [Chorus] Dans le monde des tutos virtuels GΓ’ndoline et PumbΓ© Γ  midi Une famille Γ  connecter, c'est vrai D'un enfant qui voit toi fusionner [Guitar Solo] [Verse 2] Dans le monde des tutos virtuels GΓ’ndoline, PumbΓ© Γ  midi Une famille Γ  connecter, c'est vrai D'un enfant qui voit toi fusionner<|endoftext|> 2026-07-12 09:45:31.177 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:125 - ====================================================================== 2026-07-12 09:45:31.180 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:110 - [preprocess_batch] Inferring prompt embeddings... 2026-07-12 09:45:31.188 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:113 - [preprocess_batch] Inferring lyric embeddings... 2026-07-12 09:45:31.188 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:161 - [service_generate] Generating audio... (DiT backend: PyTorch (cuda)) 2026-07-12 09:45:31.195 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:263 - [service_generate] DiT diffusion via PyTorch (cuda)... 2026-07-12 09:45:31.202 | INFO | acestep.models.common.dcw_correction:__init__:98 - [DCW] Active β€” mode=double, scaler=0.0500, high_scaler=0.0200, wavelet='haar' 2026-07-12 09:45:31.748 | INFO | acestep.core.generation.handler.generate_music_decode:_prepare_generate_music_decode_state:42 - [generate_music] Model generation completed. Decoding latents... 2026-07-12 09:45:31.748 | DEBUG | acestep.core.generation.handler.generate_music_decode:_prepare_generate_music_decode_state:64 - [generate_music] pred_latents: torch.Size([1, 2170, 64]), dtype=torch.bfloat16 2026-07-12 09:45:31.749 | DEBUG | acestep.core.generation.handler.generate_music_decode:_prepare_generate_music_decode_state:65 - [generate_music] time_costs: {'encoder_time_cost': 0.006348609924316406, 'diffusion_time_cost': 0.5453572273254395, 'diffusion_per_step_time_cost': 0.06816965341567993, 'total_time_cost': 0.5517058372497559, 'offload_time_cost': 0.0} 2026-07-12 09:45:31.771 | INFO | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:127 - [generate_music] Decoding latents with VAE... 2026-07-12 09:45:31.774 | DEBUG | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:136 - [generate_music] Before VAE decode: allocated=17.04GB, max=18.34GB 2026-07-12 09:45:31.774 | INFO | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:154 - [generate_music] Effective free VRAM before VAE decode: 76.65 GB 2026-07-12 09:45:31.774 | INFO | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:172 - [generate_music] Using tiled VAE decode to reduce VRAM usage... 2026-07-12 09:45:31.774 | DEBUG | acestep.core.generation.handler.memory_utils:_get_auto_decode_chunk_size:108 - [_get_auto_decode_chunk_size] Effective free VRAM: 76.65 GB 2026-07-12 09:45:31.774 | DEBUG | acestep.core.generation.handler.memory_utils:_should_offload_wav_to_cpu:131 - [_should_offload_wav_to_cpu] Effective free VRAM: 76.65 GB 2026-07-12 09:45:31.774 | INFO | acestep.core.generation.handler.vae_decode:tiled_decode:56 - [tiled_decode] chunk_size=512, offload_wav_to_cpu=False, latents_shape=torch.Size([1, 64, 2170]) 2026-07-12 09:45:32.049 | DEBUG | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:193 - [generate_music] After VAE decode: allocated=17.21GB, max=18.50GB 2026-07-12 09:45:32.256 | INFO | acestep.core.generation.handler.generate_music_payload:_build_generate_music_success_payload:38 - [generate_music] VAE decode completed. Preparing audio tensors... 2026-07-12 09:45:32.260 | INFO | acestep.core.generation.handler.generate_music_payload:_build_generate_music_success_payload:50 - [generate_music] Done! Generated 1 audio tensors. [Request] Parsed ggml-xl-sft/request0.json [Request] seed=42 lm_batch=1 synth_batch=1 [Request] caption: An upbeat and anthemic pop-rock track driven by bright, slig... [Request] lyrics: 490 bytes [Request] bpm=83 dur=88 key=G major ts=4 lang=fr [Request] lm: temp=0.85 cfg=2.0 top_p=0.90 top_k=0 [Request] dit: steps=50 guidance=1.0 shift=1.0 [Request] task_type: text2music [Request] solver: euler (stork_substeps=10) [Request] lm_mode: generate [Request] output_format: mp3 [Request] synth_model: acestep-v15-xl-sft-Q6_K.gguf [Request] vae: vae-BF16.gguf [Request] audio_codes: (present) [Ace-Synth] Batch: 1 request(s) [Registry] Qwen3-Embedding-0.6B-BF16.gguf -> Text-Enc [Registry] Qwen3-Embedding-0.6B-Q8_0.gguf -> Text-Enc [Registry] acestep-5Hz-lm-0.6B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-0.6B-Q8_0.gguf -> LM [Registry] acestep-5Hz-lm-1.7B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-1.7B-Q8_0.gguf -> LM [Registry] acestep-5Hz-lm-4B-BF16.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q5_K_M.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q6_K.gguf -> LM [Registry] acestep-5Hz-lm-4B-Q8_0.gguf -> LM [Registry] acestep-v15-base-BF16.gguf -> DiT [Registry] acestep-v15-base-Q4_K_M.gguf -> DiT [Registry] acestep-v15-base-Q5_K_M.gguf -> DiT [Registry] acestep-v15-base-Q6_K.gguf -> DiT [Registry] acestep-v15-base-Q8_0.gguf -> DiT [Registry] acestep-v15-sft-BF16.gguf -> DiT [Registry] acestep-v15-sft-Q4_K_M.gguf -> DiT [Registry] acestep-v15-sft-Q5_K_M.gguf -> DiT [Registry] acestep-v15-sft-Q6_K.gguf -> DiT [Registry] acestep-v15-sft-Q8_0.gguf -> DiT [Registry] acestep-v15-sftturbo50-BF16.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q4_K_M.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q5_K_M.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q6_K.gguf -> DiT [Registry] acestep-v15-sftturbo50-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-BF16.gguf -> DiT [Registry] acestep-v15-turbo-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-continuous-BF16.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-continuous-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-shift1-BF16.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-shift1-Q8_0.gguf -> DiT [Registry] acestep-v15-turbo-shift3-BF16.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q4_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q5_K_M.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q6_K.gguf -> DiT [Registry] acestep-v15-turbo-shift3-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-base-BF16.gguf -> DiT [Registry] acestep-v15-xl-base-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-base-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-base-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-base-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-sft-BF16.gguf -> DiT [Registry] acestep-v15-xl-sft-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-sft-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-sft-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-sft-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-BF16.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-sftturbo50-Q8_0.gguf -> DiT [Registry] acestep-v15-xl-turbo-BF16.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q4_K_M.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q5_K_M.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q6_K.gguf -> DiT [Registry] acestep-v15-xl-turbo-Q8_0.gguf -> DiT [Registry] scragvae-BF16.gguf -> VAE [Registry] vae-BF16.gguf -> VAE [Store] Created (policy=STRICT) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-sft-Q6_K.gguf: 830 tensors, data at offset 69056 [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-sft-Q6_K.gguf: 830 tensors, data at offset 69056 [Synth-Load] Ready: turbo=no, fa=no, batch_cfg=yes [Resolve-Params] max audio codes across batch: 434 (86.8s @ 5Hz) [Resolve-T] T=2170, S=1085 [Resolve-T] seed=42, steps=50, guidance=1.0, shift=1.0, duration=88.0s [BPE] Loaded from GGUF: 151643 vocab, 151387 merges ggml_cuda_init: found 1 CUDA devices (Total VRAM: 97247 MiB): Device 0: NVIDIA RTX PRO 6000 Blackwell Workstation Edition, compute capability 12.0, VMM: yes, VRAM: 97247 MiB load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so ggml_vulkan: Found 1 Vulkan devices: ggml_vulkan: 0 = NVIDIA RTX PRO 6000 Blackwell Workstation Edition (NVIDIA) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 49152 | int dot: 1 | matrix cores: NV_coopmat2 load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] TextEncoder backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/Qwen3-Embedding-0.6B-BF16.gguf: 310 tensors, data at offset 5337568 [Load] TextEncoder: 28L, H=1024, Nh=16/8 [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [WeightCtx] Loaded 310 tensors, 1136.5 MB into backend [Store] Load TextEnc: 240 ms [Debug] text_hidden: [70, 1024] first4: 3.704526 2.436253 0.222853 -13.131872 [Debug] lyric_embed: [167, 1024] first4: 0.029175 0.032227 -0.022339 -0.028809 [Store] Unload TextEnc (1136.5 MB) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] CondEncoder backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-sft-Q6_K.gguf: 830 tensors, data at offset 69056 [Load] LyricEncoder: 8L [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [Load] TimbreEncoder: 4L [Qwen3] Attn: Q+K+V fused [Qwen3] MLP: gate+up fused [WeightCtx] Loaded 141 tensors, 476.3 MB into backend [Load] CondEncoder: lyric(8L), timbre(4L, CLS), text_proj, null_cond [Store] Load CondEnc: 302 ms [CondEnc] Lyric sliding mask: 167x167, window=128 [CondEnc] Timbre sliding mask: 2x2, window=128 (CLS) [Encode] Packed: lyric=167 + timbre=1 + text=70 = 238 tokens [Encode-Text Batch0] 70+167 tokens -> enc_S=238, 95.1 ms [Debug] enc_hidden: [238, 2048] first4: 0.687624 -0.020412 -0.050765 0.022580 [Store] Unload CondEnc (476.3 MB) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] Detokenizer backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-sft-Q6_K.gguf: 830 tensors, data at offset 69056 [WeightCtx] Loaded 30 tensors, 82.2 MB into backend [Load] Detokenizer: FSQ(6->2048) + 2L encoder(S=5, 2048->64) [Store] Load FSQ-Detok: 58 ms [Context] Decoded: 434 codes -> 2170 frames (86.8s @ 25Hz) [Build-Context Batch0] Detokenizer: 369.4 ms, 434 codes [Debug] detok_output: [2170, 64] first4: -0.151355 1.462444 0.326907 -0.627213 [Store] Unload FSQ-Detok (82.2 MB) [Init-Noise Batch0] Philox noise seed=42, [2170, 64] solver=euler [Debug] noise: [2170, 64] first4: 0.194336 2.156250 -0.171875 0.847656 [Debug] context: [2170, 128] first4: -0.151355 1.462444 0.326907 -0.627213 [Init-Noise] Starting: T=2170, S=1085, enc_S=238, steps=50, batch=1 (cover) load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] DiT backend: CPU (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-sft-Q6_K.gguf: 830 tensors, data at offset 69056 [DiT] Self-attn: Q+K+V fused [DiT] Cross-attn: Q+K+V fused [DiT] MLP: gate+up fused [Load] null_condition_emb found (CFG available) [WeightCtx] Loaded 630 tensors, 3267.8 MB into backend [Load] DiT: 32 layers, H=2560, Nh=32/8, D=128 [Store] Load DiT: 2050 ms [DiT] Batch N=1, T=2170, S=1085, enc_S=238 [DiT] Graph: 2825 nodes [DiT] Solver: ODE Euler (1 NFE/step, order 1) [Debug] tproj: [15360] first4: 0.009506 0.101242 0.073779 -0.034931 [Debug] temb: [2560] first4: 0.009704 -0.007647 -0.006160 0.004419 [Debug] temb_t: [2560] first4: 0.008233 -0.007846 -0.006666 0.003417 [Debug] temb_r: [2560] first4: 0.001471 0.000199 0.000506 0.001001 [Debug] sinusoidal_t: [256] first4: 0.562379 0.789627 0.439928 -0.023645 [Debug] sinusoidal_r: [256] first4: 1.000000 1.000000 1.000000 1.000000 [Debug] temb_lin1_t: [2560] first4: 0.034102 -0.082811 0.088383 -0.064701 [Debug] temb_lin1_r: [2560] first4: 0.009306 -0.004837 -0.002965 -0.011076 [Debug] hidden_after_proj_in: [2560, 1085] first4: 0.333331 -0.500216 0.123159 -0.219648 [Debug] proj_in_input: [192, 2170] first4: -0.151355 1.462444 0.326907 -0.627213 [Debug] enc_after_cond_emb: [2560, 238] first4: 0.046758 0.645454 0.121823 0.862796 [Debug] layer0_sa_input: [2560, 1085] first4: 0.033760 -0.024849 0.099830 -0.102151 [Debug] layer0_q_after_rope: [128, 32] first4: -0.262700 -2.150678 0.318646 0.004803 [Debug] layer0_k_after_rope: [128, 8] first4: 0.759861 2.969560 1.271230 1.949076 [Debug] layer0_sa_output: [2560, 1085] first4: 1.965966 4.395340 -0.761963 7.744899 [Debug] layer0_attn_out: [4096, 1085] first4: 0.463727 -2.179160 -0.440477 -4.073095 [Debug] layer0_after_self_attn: [2560, 1085] first4: 0.463727 -2.179160 -0.440477 -4.073095 [Debug] layer0_after_cross_attn: [2560, 1085] first4: 0.299487 -1.915637 -0.360635 -4.138309 [Debug] hidden_after_layer0: [2560, 1085] first4: -1.565226 -1.272830 243.935760 -4.055056 [Debug] hidden_after_layer6: [2560, 1085] first4: -7.243199 3.910677 305.820038 -21.876902 [Debug] hidden_after_layer12: [2560, 1085] first4: -22.202169 38.657459 221.398972 15.539539 [Debug] hidden_after_layer18: [2560, 1085] first4: 10.920046 -13.138327 226.405396 19.980648 [Debug] hidden_after_layer31: [2560, 1085] first4: 65.760437 -106.984703 13.166077 90.631363 [Debug] dit_step0_vt: [2170, 64] first4: -0.312118 2.165119 0.054935 1.683498 [Debug] dit_step0_xt: [2170, 64] first4: 0.200578 2.112948 -0.172974 0.813986 [DiT] Step 1/50 t=1.000 [Debug] dit_step1_vt: [2170, 64] first4: -0.341468 2.139199 0.102816 1.531485 [Debug] dit_step1_xt: [2170, 64] first4: 0.207408 2.070164 -0.175030 0.783357 [DiT] Step 2/50 t=0.980 [Debug] dit_step2_vt: [2170, 64] first4: -0.314436 2.139104 0.077881 1.559264 [Debug] dit_step2_xt: [2170, 64] first4: 0.213696 2.027382 -0.176588 0.752171 [DiT] Step 3/50 t=0.960 [Debug] dit_step3_vt: [2170, 64] first4: -0.257257 2.114859 0.061832 1.547791 [Debug] dit_step3_xt: [2170, 64] first4: 0.218842 1.985085 -0.177824 0.721215 [DiT] Step 4/50 t=0.940 [Debug] dit_step4_vt: [2170, 64] first4: -0.255625 2.043399 0.083981 1.570178 [Debug] dit_step4_xt: [2170, 64] first4: 0.223954 1.944216 -0.179504 0.689812 [DiT] Step 5/50 t=0.920 [Debug] dit_step5_vt: [2170, 64] first4: -0.243652 2.033262 0.065760 1.572274 [Debug] dit_step5_xt: [2170, 64] first4: 0.228827 1.903551 -0.180819 0.658366 [DiT] Step 6/50 t=0.900 [Debug] dit_step6_vt: [2170, 64] first4: -0.254241 2.009290 0.113264 1.579486 [Debug] dit_step6_xt: [2170, 64] first4: 0.233912 1.863366 -0.183084 0.626777 [DiT] Step 7/50 t=0.880 [Debug] dit_step7_vt: [2170, 64] first4: -0.236812 1.974128 0.136143 1.604053 [Debug] dit_step7_xt: [2170, 64] first4: 0.238648 1.823883 -0.185807 0.594696 [DiT] Step 8/50 t=0.860 [Debug] dit_step8_vt: [2170, 64] first4: -0.247458 1.930153 0.172126 1.584265 [Debug] dit_step8_xt: [2170, 64] first4: 0.243597 1.785280 -0.189250 0.563010 [DiT] Step 9/50 t=0.840 [Debug] dit_step9_vt: [2170, 64] first4: -0.284515 1.854430 0.170759 1.572295 [Debug] dit_step9_xt: [2170, 64] first4: 0.249288 1.748191 -0.192665 0.531564 [DiT] Step 10/50 t=0.820 [Debug] dit_step10_vt: [2170, 64] first4: -0.275452 1.823810 0.122923 1.556191 [Debug] dit_step10_xt: [2170, 64] first4: 0.254797 1.711715 -0.195123 0.500441 [DiT] Step 11/50 t=0.800 [Debug] dit_step11_vt: [2170, 64] first4: -0.284720 1.753454 0.094287 1.542906 [Debug] dit_step11_xt: [2170, 64] first4: 0.260491 1.676646 -0.197009 0.469582 [DiT] Step 12/50 t=0.780 [Debug] dit_step12_vt: [2170, 64] first4: -0.301886 1.711960 0.079967 1.508062 [Debug] dit_step12_xt: [2170, 64] first4: 0.266529 1.642407 -0.198608 0.439421 [DiT] Step 13/50 t=0.760 [Debug] dit_step13_vt: [2170, 64] first4: -0.280330 1.658566 0.077758 1.506024 [Debug] dit_step13_xt: [2170, 64] first4: 0.272135 1.609236 -0.200164 0.409301 [DiT] Step 14/50 t=0.740 [Debug] dit_step14_vt: [2170, 64] first4: -0.267667 1.600592 0.063099 1.498749 [Debug] dit_step14_xt: [2170, 64] first4: 0.277489 1.577224 -0.201426 0.379326 [DiT] Step 15/50 t=0.720 [Debug] dit_step15_vt: [2170, 64] first4: -0.255903 1.553735 0.041921 1.493726 [Debug] dit_step15_xt: [2170, 64] first4: 0.282607 1.546149 -0.202264 0.349451 [DiT] Step 16/50 t=0.700 [Debug] dit_step16_vt: [2170, 64] first4: -0.233831 1.502487 0.040443 1.510615 [Debug] dit_step16_xt: [2170, 64] first4: 0.287283 1.516099 -0.203073 0.319239 [DiT] Step 17/50 t=0.680 [Debug] dit_step17_vt: [2170, 64] first4: -0.219107 1.444100 0.012710 1.501600 [Debug] dit_step17_xt: [2170, 64] first4: 0.291665 1.487217 -0.203327 0.289207 [DiT] Step 18/50 t=0.660 [Debug] dit_step18_vt: [2170, 64] first4: -0.159824 1.387677 0.007008 1.501050 [Debug] dit_step18_xt: [2170, 64] first4: 0.294862 1.459464 -0.203467 0.259186 [DiT] Step 19/50 t=0.640 [Debug] dit_step19_vt: [2170, 64] first4: -0.143866 1.346933 -0.000380 1.494974 [Debug] dit_step19_xt: [2170, 64] first4: 0.297739 1.432525 -0.203460 0.229287 [DiT] Step 20/50 t=0.620 [Debug] dit_step20_vt: [2170, 64] first4: -0.133790 1.320113 -0.032064 1.483299 [Debug] dit_step20_xt: [2170, 64] first4: 0.300415 1.406123 -0.202818 0.199621 [DiT] Step 21/50 t=0.600 [Debug] dit_step21_vt: [2170, 64] first4: -0.095077 1.266860 -0.038616 1.492514 [Debug] dit_step21_xt: [2170, 64] first4: 0.302317 1.380786 -0.202046 0.169770 [DiT] Step 22/50 t=0.580 [Debug] dit_step22_vt: [2170, 64] first4: -0.088900 1.218120 -0.040673 1.461899 [Debug] dit_step22_xt: [2170, 64] first4: 0.304095 1.356423 -0.201233 0.140532 [DiT] Step 23/50 t=0.560 [Debug] dit_step23_vt: [2170, 64] first4: -0.079205 1.171221 -0.056169 1.426304 [Debug] dit_step23_xt: [2170, 64] first4: 0.305679 1.332999 -0.200109 0.112006 [DiT] Step 24/50 t=0.540 [Debug] dit_step24_vt: [2170, 64] first4: -0.037089 1.112305 -0.050957 1.427209 [Debug] dit_step24_xt: [2170, 64] first4: 0.306421 1.310753 -0.199090 0.083462 [DiT] Step 25/50 t=0.520 [Debug] dit_step25_vt: [2170, 64] first4: 0.002067 1.041147 -0.075665 1.374744 [Debug] dit_step25_xt: [2170, 64] first4: 0.306379 1.289930 -0.197577 0.055967 [DiT] Step 26/50 t=0.500 [Debug] dit_step26_vt: [2170, 64] first4: -0.036625 0.974199 -0.082610 1.408507 [Debug] dit_step26_xt: [2170, 64] first4: 0.307112 1.270446 -0.195925 0.027797 [DiT] Step 27/50 t=0.480 [Debug] dit_step27_vt: [2170, 64] first4: -0.065509 0.932928 -0.083212 1.343793 [Debug] dit_step27_xt: [2170, 64] first4: 0.308422 1.251787 -0.194260 0.000921 [DiT] Step 28/50 t=0.460 [Debug] dit_step28_vt: [2170, 64] first4: -0.086871 0.879863 -0.092483 1.314061 [Debug] dit_step28_xt: [2170, 64] first4: 0.310159 1.234190 -0.192411 -0.025360 [DiT] Step 29/50 t=0.440 [Debug] dit_step29_vt: [2170, 64] first4: -0.064771 0.832469 -0.095781 1.292529 [Debug] dit_step29_xt: [2170, 64] first4: 0.311455 1.217541 -0.190495 -0.051211 [DiT] Step 30/50 t=0.420 [Debug] dit_step30_vt: [2170, 64] first4: -0.060545 0.727658 -0.106520 1.283506 [Debug] dit_step30_xt: [2170, 64] first4: 0.312666 1.202988 -0.188365 -0.076881 [DiT] Step 31/50 t=0.400 [Debug] dit_step31_vt: [2170, 64] first4: -0.127297 0.672579 -0.111527 1.211460 [Debug] dit_step31_xt: [2170, 64] first4: 0.315212 1.189536 -0.186134 -0.101110 [DiT] Step 32/50 t=0.380 [Debug] dit_step32_vt: [2170, 64] first4: -0.124123 0.580164 -0.097863 1.236331 [Debug] dit_step32_xt: [2170, 64] first4: 0.317694 1.177933 -0.184177 -0.125837 [DiT] Step 33/50 t=0.360 [Debug] dit_step33_vt: [2170, 64] first4: -0.132405 0.494401 -0.097390 1.225479 [Debug] dit_step33_xt: [2170, 64] first4: 0.320342 1.168045 -0.182229 -0.150346 [DiT] Step 34/50 t=0.340 [Debug] dit_step34_vt: [2170, 64] first4: -0.099086 0.403934 -0.101936 1.146895 [Debug] dit_step34_xt: [2170, 64] first4: 0.322324 1.159966 -0.180190 -0.173284 [DiT] Step 35/50 t=0.320 [Debug] dit_step35_vt: [2170, 64] first4: -0.110905 0.282513 -0.074015 1.096174 [Debug] dit_step35_xt: [2170, 64] first4: 0.324542 1.154316 -0.178710 -0.195208 [DiT] Step 36/50 t=0.300 [Debug] dit_step36_vt: [2170, 64] first4: -0.138414 0.218764 -0.068279 1.017159 [Debug] dit_step36_xt: [2170, 64] first4: 0.327310 1.149940 -0.177344 -0.215551 [DiT] Step 37/50 t=0.280 [Debug] dit_step37_vt: [2170, 64] first4: -0.129680 0.079816 -0.113666 1.027041 [Debug] dit_step37_xt: [2170, 64] first4: 0.329904 1.148344 -0.175071 -0.236091 [DiT] Step 38/50 t=0.260 [Debug] dit_step38_vt: [2170, 64] first4: -0.130753 -0.041390 -0.112635 0.998467 [Debug] dit_step38_xt: [2170, 64] first4: 0.332519 1.149172 -0.172818 -0.256061 [DiT] Step 39/50 t=0.240 [Debug] dit_step39_vt: [2170, 64] first4: -0.212455 -0.157115 -0.100421 0.968666 [Debug] dit_step39_xt: [2170, 64] first4: 0.336768 1.152314 -0.170810 -0.275434 [DiT] Step 40/50 t=0.220 [Debug] dit_step40_vt: [2170, 64] first4: -0.220059 -0.279065 -0.010904 0.872365 [Debug] dit_step40_xt: [2170, 64] first4: 0.341169 1.157895 -0.170592 -0.292881 [DiT] Step 41/50 t=0.200 [Debug] dit_step41_vt: [2170, 64] first4: -0.238877 -0.357423 -0.020465 0.841890 [Debug] dit_step41_xt: [2170, 64] first4: 0.345947 1.165044 -0.170183 -0.309719 [DiT] Step 42/50 t=0.180 [Debug] dit_step42_vt: [2170, 64] first4: -0.264151 -0.454071 0.034757 0.822858 [Debug] dit_step42_xt: [2170, 64] first4: 0.351230 1.174125 -0.170878 -0.326176 [DiT] Step 43/50 t=0.160 [Debug] dit_step43_vt: [2170, 64] first4: -0.334725 -0.540851 0.010944 0.780511 [Debug] dit_step43_xt: [2170, 64] first4: 0.357924 1.184942 -0.171097 -0.341787 [DiT] Step 44/50 t=0.140 [Debug] dit_step44_vt: [2170, 64] first4: -0.362583 -0.636920 0.076876 0.708881 [Debug] dit_step44_xt: [2170, 64] first4: 0.365176 1.197681 -0.172634 -0.355964 [DiT] Step 45/50 t=0.120 [Debug] dit_step45_vt: [2170, 64] first4: -0.351643 -0.676087 0.116158 0.678409 [Debug] dit_step45_xt: [2170, 64] first4: 0.372209 1.211203 -0.174957 -0.369532 [DiT] Step 46/50 t=0.100 [Debug] dit_step46_vt: [2170, 64] first4: -0.397799 -0.769951 0.123705 0.640916 [Debug] dit_step46_xt: [2170, 64] first4: 0.380165 1.226602 -0.177431 -0.382351 [DiT] Step 47/50 t=0.080 [Debug] dit_step47_vt: [2170, 64] first4: -0.471523 -0.841422 0.191995 0.629210 [Debug] dit_step47_xt: [2170, 64] first4: 0.389595 1.243430 -0.181271 -0.394935 [DiT] Step 48/50 t=0.060 [Debug] dit_step48_vt: [2170, 64] first4: -0.500148 -0.940853 0.261135 0.583724 [Debug] dit_step48_xt: [2170, 64] first4: 0.399598 1.262247 -0.186494 -0.406609 [DiT] Step 49/50 t=0.040 [Debug] dit_step49_vt: [2170, 64] first4: -0.529853 -0.882358 0.194948 0.611176 [Debug] dit_x0: [2170, 64] first4: 0.410195 1.279894 -0.190393 -0.418833 [DiT] Step 50/50 t=0.020 [DiT-Generate] Total: 366804.0 ms (366804.0 ms/sample) [Debug] dit_output: [2170, 64] first4: 0.410195 1.279894 -0.190393 -0.418833 [Store] Unload DiT (3267.8 MB) [GGUF] /mnt/workspace/git/acestep.cpp/models/vae-BF16.gguf: 365 tensors, data at offset 30048 load_backend: loaded CUDA backend from /mnt/workspace/git/acestep.cpp/build/libggml-cuda.so load_backend: loaded Vulkan backend from /mnt/workspace/git/acestep.cpp/build/libggml-vulkan.so load_backend: loaded CPU backend from /mnt/workspace/git/acestep.cpp/build/libggml-cpu-zen4.so [Load] VAE backend: CPU (CPU threads: 16) [VAE] Backend: CPU, Weight buffer: 161.1 MB [VAE] Loaded: 5 blocks, upsample=1920x, F32 activations [Store] Load VAE-Dec: 717 ms [VAE] Tiled decode: 3 tiles (chunk=1024, overlap=64, stride=896) [VAE] Graph: 479 nodes, T_latent=960 [VAE] Upsample factor: 1920.00 (expected ~1920) [VAE] Graph: 479 nodes, T_latent=1024 [VAE] Graph: 479 nodes, T_latent=442 [VAE] Tiled decode done: 3 tiles -> T_audio=4166400 (86.80s @ 48kHz) [VAE-Decode Batch0] Decode: 33421.6 ms [Debug] vae_audio: [2, 4166400] first4: -0.001430 -0.000715 -0.000854 -0.000356 [Store] Unload VAE-Dec (161.1 MB) [MP3] Encoding 86.8s @ 128 kbps, 48000 Hz stereo [MP3] 1388928 bytes (12.0:1), 471 ms (184.24x realtime), 32 threads [MP3] Wrote ggml-xl-sft/request00.mp3 [Ace-Synth] All done 2026-07-12 09:52:17.940 | INFO | acestep.core.generation.handler.init_service_loader:_load_main_model_from_checkpoint:174 - [initialize_service] Attempting to load model with attention implementation: sdpa [Python] Generating (acestep-v15-xl-turbo, 8 steps)... Using precomputed LM hints Using precomputed LM hints [Python] Wrote python-xl-turbo/output.wav: 4166400 samples (86.80s @ 48kHz stereo) [Python] Done, 40 dump files [Xl-turbo] Cosine similarities GGML vs Python stage GGML vs Python text_hidden 0.999810 lyric_embed 1.000000 enc_hidden 0.999359 detok_output 0.999927 context 0.999954 noise 1.000000 temb_t 0.999968 hidden_after_proj_in 0.999961 enc_after_cond_emb 0.999393 layer0_sa_output 0.999156 hidden_after_layer0 0.999316 hidden_after_layer6 0.999727 hidden_after_layer12 0.996090 hidden_after_layer18 0.996132 hidden_after_layer31 0.988087 dit_step0_vt 0.911993 dit_step0_xt 0.999842 dit_step1_vt 0.926791 dit_step1_xt 0.999523 dit_step2_vt 0.928856 dit_step2_xt 0.998744 dit_step3_vt 0.919484 dit_step3_xt 0.996800 dit_step4_vt 0.916476 dit_step4_xt 0.992119 dit_step5_vt 0.909024 dit_step5_xt 0.980905 dit_step6_vt 0.902058 dit_step6_xt 0.958807 dit_step7_vt 0.881859 dit_x0 0.923452 vae_audio 0.717186 vae_audio (STFT cosine) 0.920635 [Xl-turbo] Error growth GGML vs Python stage cos max_err mean_err mean_A std_A mean_B std_B dit_step0_xt 0.999842 0.235044 0.012536 -0.002976 0.978420 -0.002721 0.976534 dit_step1_xt 0.999523 0.306101 0.021399 -0.005971 0.952195 -0.005567 0.949004 dit_step2_xt 0.998744 0.542181 0.033244 -0.009756 0.923704 -0.009330 0.918931 dit_step3_xt 0.996800 0.802876 0.050476 -0.015259 0.893128 -0.014418 0.887241 dit_step4_xt 0.992119 1.270764 0.075816 -0.022871 0.863841 -0.021453 0.857796 dit_step5_xt 0.980905 2.267337 0.115177 -0.034042 0.847625 -0.032240 0.840033 dit_step6_xt 0.958807 3.799029 0.174737 -0.051150 0.873138 -0.049194 0.863964 [Xl-sft] steps=50, shift=1.0 | acestep-v15-xl-sft-Q6_K.gguf [GGML] Running acestep-v15-xl-sft-Q6_K.gguf... [GGML] Done, 131 dump files [Python] Initializing acestep-v15-xl-sft... Loading checkpoint shards: 0%| | 0/4 [00:00 2026-07-12 09:52:21.851 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:123 - ====================================================================== 2026-07-12 09:52:21.851 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:124 - lyrics_text: # Languages fr # Lyric # Lyric [Intro - Guitar Riff] [Verse 1] Dans le monde des tutos virtuels G ta toise en nouvelle passion Avec Ggendoline et PumbΓ© Γ  midi La communautΓ©, c'est l'unitΓ© Quel joie, une clΓ© [Chorus] Dans le monde des tutos virtuels GΓ’ndoline et PumbΓ© Γ  midi Une famille Γ  connecter, c'est vrai D'un enfant qui voit toi fusionner [Guitar Solo] [Verse 2] Dans le monde des tutos virtuels GΓ’ndoline, PumbΓ© Γ  midi Une famille Γ  connecter, c'est vrai D'un enfant qui voit toi fusionner<|endoftext|> 2026-07-12 09:52:21.851 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:125 - ====================================================================== 2026-07-12 09:52:21.855 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:110 - [preprocess_batch] Inferring prompt embeddings... 2026-07-12 09:52:21.862 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:113 - [preprocess_batch] Inferring lyric embeddings... 2026-07-12 09:52:21.863 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:161 - [service_generate] Generating audio... (DiT backend: PyTorch (cuda)) 2026-07-12 09:52:21.869 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:263 - [service_generate] DiT diffusion via PyTorch (cuda)... [Python] Generating (acestep-v15-xl-sft, 50 steps)... Using precomputed LM hints Using precomputed LM hints 0%| | 0/50 [00:00