[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-Q8_0.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-Q8_0.gguf: 678 tensors, data at offset 56864 [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-turbo-Q8_0.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: CUDA0 (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: 283 ms [Debug] text_hidden: [70, 1024] first4: 3.652379 1.048645 0.229843 -13.063987 [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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-turbo-Q8_0.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, 616.6 MB into backend [Load] CondEncoder: lyric(8L), timbre(4L), text_proj, null_cond [Store] Load CondEnc: 323 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, 19.2 ms [Debug] enc_hidden: [238, 2048] first4: 1.758800 -0.049267 -0.133030 0.058411 [Store] Unload CondEnc (616.6 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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-turbo-Q8_0.gguf: 678 tensors, data at offset 56864 [WeightCtx] Loaded 30 tensors, 106.5 MB into backend [Load] Detokenizer: FSQ(6->2048) + 2L encoder(S=5, 2048->64) [Store] Load FSQ-Detok: 72 ms CUDA Graph id 8 reused ggml_backend_cuda_graph_compute: CUDA graph warmup complete CUDA Graph id 8 reused [Context] Decoded: 434 codes -> 2170 frames (86.8s @ 25Hz) [Build-Context Batch0] Detokenizer: 61.1 ms, 434 codes [Debug] detok_output: [2170, 64] first4: -0.122364 1.438543 0.301741 -0.633332 [Store] Unload FSQ-Detok (106.5 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: -0.122364 1.438543 0.301741 -0.633332 [Init-Noise] Starting: T=2170, S=1085, enc_S=238, steps=8, batch=1 (cover) [Dedup] Previous line repeated 432 times total 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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-turbo-Q8_0.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, 1600.7 MB into backend [Load] DiT: 24 layers, H=2048, Nh=16/8, D=128 [Store] Load DiT: 1014 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: 0.259485 -0.161550 -0.096885 0.051766 [Debug] temb: [2048] first4: 0.000214 -0.132557 -0.035428 0.064847 [Debug] temb_t: [2048] first4: 0.001194 0.026823 -0.052744 0.063762 [Debug] temb_r: [2048] first4: -0.000980 -0.159380 0.017316 0.001084 [Debug] sinusoidal_t: [256] first4: 0.562486 0.789701 0.439822 -0.023583 [Debug] sinusoidal_r: [256] first4: 1.000000 1.000000 1.000000 1.000000 [Debug] temb_lin1_t: [2048] first4: -0.049228 -0.051913 -0.015026 -0.038076 [Debug] temb_lin1_r: [2048] first4: -0.013066 -0.018835 -0.015731 0.008462 [Debug] hidden_after_proj_in: [2048, 1085] first4: -1.038952 -0.960857 0.538357 0.447870 [Debug] proj_in_input: [192, 2170] first4: -0.122364 1.438543 0.301741 -0.633332 [Debug] enc_after_cond_emb: [2048, 238] first4: -0.176740 0.812502 0.336566 -0.559637 [Debug] layer0_sa_input: [2048, 1085] first4: -0.719128 -0.759064 -0.046924 0.261745 [Debug] layer0_q_after_rope: [128, 16] first4: -1.544455 -1.034858 0.189425 0.457322 [Debug] layer0_k_after_rope: [128, 8] first4: -0.176740 0.812502 0.336566 -0.559637 [Debug] layer0_sa_output: [2048, 1085] first4: -1.513538 0.166696 -0.357260 0.506110 [Debug] layer0_attn_out: [2048, 1085] first4: -1.544455 -1.034858 0.189425 0.457322 [Debug] layer0_after_self_attn: [2048, 1085] first4: -1.544455 -1.034858 0.189425 0.457322 [Debug] layer0_after_cross_attn: [2048, 1085] first4: -1.599932 -0.820419 -0.295126 0.493654 [Debug] hidden_after_layer0: [2048, 1085] first4: -9.073871 0.556383 52.114643 -0.861245 [Debug] hidden_after_layer6: [2048, 1085] first4: -21.753124 -0.010392 34.315594 -4.318250 [Debug] hidden_after_layer12: [2048, 1085] first4: -15.882568 -18.143053 72.389709 28.553534 [Debug] hidden_after_layer18: [2048, 1085] first4: -26.068863 15.209676 60.847687 19.705666 [Debug] hidden_after_layer23: [2048, 1085] first4: -4.074146 40.110039 193.608887 143.767899 [Debug] dit_step0_vt: [2170, 64] first4: 0.022585 1.133021 0.320472 2.355647 [Debug] dit_step0_xt: [2170, 64] first4: 0.193309 2.104749 -0.186442 0.740581 [DiT] Step 1/8 t=1.000 ggml_backend_cuda_graph_compute: CUDA graph warmup complete [Debug] dit_step1_vt: [2170, 64] first4: -0.199451 1.338411 -0.105943 1.896582 [Debug] dit_step1_xt: [2170, 64] first4: 0.204188 2.031745 -0.180663 0.637131 [DiT] Step 2/8 t=0.955 [Debug] dit_step2_vt: [2170, 64] first4: -0.026068 1.237643 0.104578 2.386231 [Debug] dit_step2_xt: [2170, 64] first4: 0.205926 1.949235 -0.187635 0.478049 [DiT] Step 3/8 t=0.900 [Debug] dit_step3_vt: [2170, 64] first4: 0.243367 1.084662 0.274099 2.679283 [Debug] dit_step3_xt: [2170, 64] first4: 0.185646 1.858847 -0.210477 0.254776 [DiT] Step 4/8 t=0.833 [Debug] dit_step4_vt: [2170, 64] first4: 0.284440 1.024341 0.106560 2.727204 [Debug] dit_step4_xt: [2170, 64] first4: 0.155170 1.749096 -0.221894 -0.037425 [DiT] Step 5/8 t=0.750 [Debug] dit_step5_vt: [2170, 64] first4: 0.261286 0.944551 -0.261736 2.741330 [Debug] dit_step5_xt: [2170, 64] first4: 0.117844 1.614160 -0.184503 -0.429043 [DiT] Step 6/8 t=0.643 [Debug] dit_step6_vt: [2170, 64] first4: 0.183861 0.617940 -0.818820 2.811596 [Debug] dit_step6_xt: [2170, 64] first4: 0.081071 1.490572 -0.020739 -0.991362 [DiT] Step 7/8 t=0.500 [Debug] dit_step7_vt: [2170, 64] first4: 0.028215 0.224690 -1.458037 3.102695 [Debug] dit_x0: [2170, 64] first4: 0.072607 1.423165 0.416672 -1.922171 [DiT] Step 8/8 t=0.300 [DiT-Generate] Total: 273.8 ms (273.8 ms/sample) [Debug] dit_output: [2170, 64] first4: 0.072607 1.423165 0.416672 -1.922171 [Store] Unload DiT (1600.7 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: CUDA0 (CPU threads: 16) [VAE] Backend: CUDA0, Weight buffer: 161.1 MB [VAE] Loaded: 5 blocks, upsample=1920x, F32 activations [Store] Load VAE-Dec: 670 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: 426.7 ms [Debug] vae_audio: [2, 4166400] first4: 0.000683 0.001266 0.001143 0.001556 [Store] Unload VAE-Dec (161.1 MB) [MP3] Encoding 86.8s @ 128 kbps, 48000 Hz stereo [MP3] 1388928 bytes (12.0:1), 448 ms (193.62x realtime), 32 threads [MP3] Wrote ggml-turbo/request00.mp3 [Ace-Synth] All done 2026-07-12 10:08:53.718 | 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 10:08:55.850 | WARNING | acestep.training.trainer::40 - bitsandbytes not installed. Using standard AdamW. 2026-07-12 10:08:55.855 | 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 10:08:55.925 | 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 10:08:57.565 | 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 10:08:57.565 | INFO | acestep.core.generation.handler.generate_music:generate_music:304 - [generate_music] Starting generation... 2026-07-12 10:08:57.565 | INFO | acestep.core.generation.handler.generate_music:generate_music:307 - [generate_music] Preparing inputs... 2026-07-12 10:08:57.565 | 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 10:08:57.568 | 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 10:08:57.573 | INFO | acestep.core.generation.handler.conditioning_target:_prepare_target_latents_and_wavs:78 - [generate_music] Decoding audio codes for item 0... 2026-07-12 10:08:57.784 | 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 10:08:57.786 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:119 - ====================================================================== 2026-07-12 10:08:57.786 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:120 - πŸ” [DEBUG] DiT TEXT ENCODER INPUT (Inference) 2026-07-12 10:08:57.786 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:121 - ====================================================================== 2026-07-12 10:08:57.786 | 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 10:08:57.786 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:123 - ====================================================================== 2026-07-12 10:08:57.786 | 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 10:08:57.786 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:125 - ====================================================================== 2026-07-12 10:08:57.793 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:110 - [preprocess_batch] Inferring prompt embeddings... 2026-07-12 10:08:57.826 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:113 - [preprocess_batch] Inferring lyric embeddings... 2026-07-12 10:08:57.827 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:161 - [service_generate] Generating audio... (DiT backend: PyTorch (cuda)) 2026-07-12 10:08:57.863 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:263 - [service_generate] DiT diffusion via PyTorch (cuda)... 2026-07-12 10:08:57.871 | 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 10:08:57.955 | 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 10:08:58.180 | INFO | acestep.core.generation.handler.generate_music_decode:_prepare_generate_music_decode_state:42 - [generate_music] Model generation completed. Decoding latents... 2026-07-12 10:08:58.181 | 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 10:08:58.181 | DEBUG | acestep.core.generation.handler.generate_music_decode:_prepare_generate_music_decode_state:65 - [generate_music] time_costs: {'encoder_time_cost': 0.006938934326171875, 'diffusion_time_cost': 0.3101475238800049, 'diffusion_per_step_time_cost': 0.03876844048500061, 'total_time_cost': 0.31708645820617676, 'offload_time_cost': 0.0} 2026-07-12 10:08:58.195 | INFO | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:127 - [generate_music] Decoding latents with VAE... 2026-07-12 10:08:58.197 | 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 10:08:58.198 | INFO | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:154 - [generate_music] Effective free VRAM before VAE decode: 87.73 GB 2026-07-12 10:08:58.198 | 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 10:08:58.198 | DEBUG | acestep.core.generation.handler.memory_utils:_get_auto_decode_chunk_size:108 - [_get_auto_decode_chunk_size] Effective free VRAM: 87.73 GB 2026-07-12 10:08:58.198 | DEBUG | acestep.core.generation.handler.memory_utils:_should_offload_wav_to_cpu:131 - [_should_offload_wav_to_cpu] Effective free VRAM: 87.73 GB 2026-07-12 10:08:58.198 | 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 10:08:58.474 | 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 10:08:58.636 | 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 10:08:58.639 | 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-Q8_0.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-Q8_0.gguf: 678 tensors, data at offset 56800 [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-sft-Q8_0.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: CUDA0 (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: 260 ms [Debug] text_hidden: [70, 1024] first4: 3.652379 1.048645 0.229843 -13.063987 [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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-sft-Q8_0.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, 616.6 MB into backend [Load] CondEncoder: lyric(8L), timbre(4L), text_proj, null_cond [Store] Load CondEnc: 306 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, 15.7 ms [Debug] enc_hidden: [238, 2048] first4: 1.758800 -0.049267 -0.133030 0.058411 [Store] Unload CondEnc (616.6 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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-sft-Q8_0.gguf: 678 tensors, data at offset 56800 [WeightCtx] Loaded 30 tensors, 106.5 MB into backend [Load] Detokenizer: FSQ(6->2048) + 2L encoder(S=5, 2048->64) [Store] Load FSQ-Detok: 54 ms CUDA Graph id 8 reused ggml_backend_cuda_graph_compute: CUDA graph warmup complete CUDA Graph id 8 reused [Context] Decoded: 434 codes -> 2170 frames (86.8s @ 25Hz) [Build-Context Batch0] Detokenizer: 60.9 ms, 434 codes [Debug] detok_output: [2170, 64] first4: -0.122364 1.438543 0.301741 -0.633332 [Store] Unload FSQ-Detok (106.5 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: -0.122364 1.438543 0.301741 -0.633332 [Init-Noise] Starting: T=2170, S=1085, enc_S=238, steps=50, batch=1 (cover) [Dedup] Previous line repeated 432 times total 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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-sft-Q8_0.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, 1600.7 MB into backend [Load] DiT: 24 layers, H=2048, Nh=16/8, D=128 [Store] Load DiT: 872 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: 0.155643 -0.115659 -0.086888 0.081457 [Debug] temb: [2048] first4: -0.002814 -0.176488 0.004230 -0.002015 [Debug] temb_t: [2048] first4: -0.001218 0.003344 -0.013065 -0.002081 [Debug] temb_r: [2048] first4: -0.001596 -0.179832 0.017296 0.000066 [Debug] sinusoidal_t: [256] first4: 0.562486 0.789701 0.439822 -0.023583 [Debug] sinusoidal_r: [256] first4: 1.000000 1.000000 1.000000 1.000000 [Debug] temb_lin1_t: [2048] first4: -0.038611 0.029679 0.028318 -0.024159 [Debug] temb_lin1_r: [2048] first4: 0.001863 0.003353 -0.000552 -0.000197 [Debug] hidden_after_proj_in: [2048, 1085] first4: -1.089495 -0.917714 0.500524 0.441531 [Debug] proj_in_input: [192, 2170] first4: -0.122364 1.438543 0.301741 -0.633332 [Debug] enc_after_cond_emb: [2048, 238] first4: -0.195530 0.868084 0.332281 -0.583266 [Debug] layer0_sa_input: [2048, 1085] first4: -0.913030 -0.710890 -0.037265 0.293969 [Debug] layer0_q_after_rope: [128, 16] first4: -1.800851 -1.366383 -0.115733 0.381449 [Debug] layer0_k_after_rope: [128, 8] first4: -0.195530 0.868084 0.332281 -0.583266 [Debug] layer0_sa_output: [2048, 1085] first4: -1.645180 0.793266 -0.546254 0.536514 [Debug] layer0_attn_out: [2048, 1085] first4: -1.800851 -1.366383 -0.115733 0.381449 [Debug] layer0_after_self_attn: [2048, 1085] first4: -1.800851 -1.366383 -0.115733 0.381449 [Debug] layer0_after_cross_attn: [2048, 1085] first4: -1.960260 -1.030991 -0.403150 0.386583 [Debug] hidden_after_layer0: [2048, 1085] first4: -9.395753 1.192434 57.112274 -1.590155 [Debug] hidden_after_layer6: [2048, 1085] first4: -17.518703 4.160506 57.738289 -3.046195 [Debug] hidden_after_layer12: [2048, 1085] first4: -10.493816 4.078718 -22.709408 -2.909376 [Debug] hidden_after_layer18: [2048, 1085] first4: -4.034719 19.264593 -39.337860 -12.487666 [Debug] hidden_after_layer23: [2048, 1085] first4: 33.825279 63.970184 56.735619 -8.967500 [Debug] dit_step0_vt: [2170, 64] first4: -0.597517 2.529882 -0.171392 1.456065 [Debug] dit_step0_xt: [2170, 64] first4: 0.206286 2.105652 -0.168447 0.818535 [DiT] Step 1/50 t=1.000 ggml_backend_cuda_graph_compute: CUDA graph warmup complete [Debug] dit_step1_vt: [2170, 64] first4: -0.640031 2.546048 -0.002364 1.296107 [Debug] dit_step1_xt: [2170, 64] first4: 0.219087 2.054731 -0.168400 0.792613 [DiT] Step 2/50 t=0.980 [Debug] dit_step2_vt: [2170, 64] first4: -0.633882 2.492483 0.020320 1.300945 [Debug] dit_step2_xt: [2170, 64] first4: 0.231765 2.004882 -0.168806 0.766594 [DiT] Step 3/50 t=0.960 [Debug] dit_step3_vt: [2170, 64] first4: -0.628523 2.448052 0.046230 1.321965 [Debug] dit_step3_xt: [2170, 64] first4: 0.244335 1.955921 -0.169731 0.740155 [DiT] Step 4/50 t=0.940 [Debug] dit_step4_vt: [2170, 64] first4: -0.565812 2.363279 0.022533 1.319726 [Debug] dit_step4_xt: [2170, 64] first4: 0.255651 1.908655 -0.170182 0.713760 [DiT] Step 5/50 t=0.920 [Debug] dit_step5_vt: [2170, 64] first4: -0.511288 2.273364 0.010095 1.294248 [Debug] dit_step5_xt: [2170, 64] first4: 0.265877 1.863188 -0.170383 0.687875 [DiT] Step 6/50 t=0.900 [Debug] dit_step6_vt: [2170, 64] first4: -0.490007 2.232753 -0.013814 1.291401 [Debug] dit_step6_xt: [2170, 64] first4: 0.275677 1.818533 -0.170107 0.662047 [DiT] Step 7/50 t=0.880 [Debug] dit_step7_vt: [2170, 64] first4: -0.450441 2.158698 -0.045945 1.276538 [Debug] dit_step7_xt: [2170, 64] first4: 0.284686 1.775359 -0.169188 0.636516 [DiT] Step 8/50 t=0.860 [Debug] dit_step8_vt: [2170, 64] first4: -0.399656 2.051490 -0.054245 1.250984 [Debug] dit_step8_xt: [2170, 64] first4: 0.292679 1.734329 -0.168103 0.611497 [DiT] Step 9/50 t=0.840 [Debug] dit_step9_vt: [2170, 64] first4: -0.321156 1.968992 -0.058021 1.249884 [Debug] dit_step9_xt: [2170, 64] first4: 0.299102 1.694949 -0.166943 0.586499 [DiT] Step 10/50 t=0.820 [Debug] dit_step10_vt: [2170, 64] first4: -0.268805 1.885387 -0.049984 1.231170 [Debug] dit_step10_xt: [2170, 64] first4: 0.304478 1.657241 -0.165943 0.561876 [DiT] Step 11/50 t=0.800 [Debug] dit_step11_vt: [2170, 64] first4: -0.242812 1.800516 -0.044606 1.231250 [Debug] dit_step11_xt: [2170, 64] first4: 0.309335 1.621231 -0.165051 0.537251 [DiT] Step 12/50 t=0.780 [Debug] dit_step12_vt: [2170, 64] first4: -0.206551 1.749844 -0.054550 1.215934 [Debug] dit_step12_xt: [2170, 64] first4: 0.313466 1.586234 -0.163960 0.512932 [DiT] Step 13/50 t=0.760 [Debug] dit_step13_vt: [2170, 64] first4: -0.192853 1.666518 -0.067789 1.210403 [Debug] dit_step13_xt: [2170, 64] first4: 0.317323 1.552904 -0.162604 0.488724 [DiT] Step 14/50 t=0.740 [Debug] dit_step14_vt: [2170, 64] first4: -0.188553 1.591855 -0.093213 1.183844 [Debug] dit_step14_xt: [2170, 64] first4: 0.321094 1.521067 -0.160740 0.465047 [DiT] Step 15/50 t=0.720 [Debug] dit_step15_vt: [2170, 64] first4: -0.191758 1.514507 -0.130168 1.177348 [Debug] dit_step15_xt: [2170, 64] first4: 0.324929 1.490777 -0.158137 0.441500 [DiT] Step 16/50 t=0.700 [Debug] dit_step16_vt: [2170, 64] first4: -0.183488 1.417093 -0.153953 1.144312 [Debug] dit_step16_xt: [2170, 64] first4: 0.328599 1.462435 -0.155058 0.418614 [DiT] Step 17/50 t=0.680 [Debug] dit_step17_vt: [2170, 64] first4: -0.201386 1.354984 -0.182355 1.146774 [Debug] dit_step17_xt: [2170, 64] first4: 0.332626 1.435335 -0.151411 0.395678 [DiT] Step 18/50 t=0.660 [Debug] dit_step18_vt: [2170, 64] first4: -0.207087 1.268490 -0.207697 1.115252 [Debug] dit_step18_xt: [2170, 64] first4: 0.336768 1.409965 -0.147257 0.373373 [DiT] Step 19/50 t=0.640 [Debug] dit_step19_vt: [2170, 64] first4: -0.234214 1.184564 -0.238867 1.102345 [Debug] dit_step19_xt: [2170, 64] first4: 0.341452 1.386274 -0.142479 0.351326 [DiT] Step 20/50 t=0.620 [Debug] dit_step20_vt: [2170, 64] first4: -0.253691 1.110998 -0.251266 1.079397 [Debug] dit_step20_xt: [2170, 64] first4: 0.346526 1.364054 -0.137454 0.329739 [DiT] Step 21/50 t=0.600 [Debug] dit_step21_vt: [2170, 64] first4: -0.266174 1.026011 -0.276170 1.036771 [Debug] dit_step21_xt: [2170, 64] first4: 0.351850 1.343534 -0.131931 0.309003 [DiT] Step 22/50 t=0.580 [Debug] dit_step22_vt: [2170, 64] first4: -0.293246 0.911232 -0.257626 1.020393 [Debug] dit_step22_xt: [2170, 64] first4: 0.357715 1.325309 -0.126778 0.288595 [DiT] Step 23/50 t=0.560 [Debug] dit_step23_vt: [2170, 64] first4: -0.305231 0.812165 -0.266379 0.996722 [Debug] dit_step23_xt: [2170, 64] first4: 0.363819 1.309066 -0.121450 0.268661 [DiT] Step 24/50 t=0.540 [Debug] dit_step24_vt: [2170, 64] first4: -0.336330 0.712807 -0.258532 0.979090 [Debug] dit_step24_xt: [2170, 64] first4: 0.370546 1.294810 -0.116280 0.249079 [DiT] Step 25/50 t=0.520 [Debug] dit_step25_vt: [2170, 64] first4: -0.350820 0.566485 -0.274051 0.960184 [Debug] dit_step25_xt: [2170, 64] first4: 0.377562 1.283480 -0.110799 0.229875 [DiT] Step 26/50 t=0.500 [Debug] dit_step26_vt: [2170, 64] first4: -0.393387 0.445964 -0.304000 0.913025 [Debug] dit_step26_xt: [2170, 64] first4: 0.385430 1.274561 -0.104719 0.211615 [DiT] Step 27/50 t=0.480 [Debug] dit_step27_vt: [2170, 64] first4: -0.427608 0.347451 -0.283136 0.896502 [Debug] dit_step27_xt: [2170, 64] first4: 0.393982 1.267612 -0.099056 0.193685 [DiT] Step 28/50 t=0.460 [Debug] dit_step28_vt: [2170, 64] first4: -0.421758 0.231981 -0.276398 0.867093 [Debug] dit_step28_xt: [2170, 64] first4: 0.402417 1.262972 -0.093528 0.176343 [DiT] Step 29/50 t=0.440 [Debug] dit_step29_vt: [2170, 64] first4: -0.462086 0.121544 -0.268720 0.845826 [Debug] dit_step29_xt: [2170, 64] first4: 0.411659 1.260541 -0.088154 0.159426 [DiT] Step 30/50 t=0.420 [Debug] dit_step30_vt: [2170, 64] first4: -0.494771 0.037362 -0.219930 0.838203 [Debug] dit_step30_xt: [2170, 64] first4: 0.421554 1.259794 -0.083755 0.142662 [DiT] Step 31/50 t=0.400 [Debug] dit_step31_vt: [2170, 64] first4: -0.494006 -0.104929 -0.237653 0.830174 [Debug] dit_step31_xt: [2170, 64] first4: 0.431435 1.261893 -0.079002 0.126059 [DiT] Step 32/50 t=0.380 [Debug] dit_step32_vt: [2170, 64] first4: -0.515593 -0.237057 -0.239353 0.790560 [Debug] dit_step32_xt: [2170, 64] first4: 0.441746 1.266634 -0.074215 0.110248 [DiT] Step 33/50 t=0.360 [Debug] dit_step33_vt: [2170, 64] first4: -0.544791 -0.338143 -0.191470 0.785250 [Debug] dit_step33_xt: [2170, 64] first4: 0.452642 1.273397 -0.070386 0.094543 [DiT] Step 34/50 t=0.340 [Debug] dit_step34_vt: [2170, 64] first4: -0.537031 -0.421770 -0.187369 0.756133 [Debug] dit_step34_xt: [2170, 64] first4: 0.463383 1.281832 -0.066638 0.079420 [DiT] Step 35/50 t=0.320 [Debug] dit_step35_vt: [2170, 64] first4: -0.530635 -0.541913 -0.169903 0.728508 [Debug] dit_step35_xt: [2170, 64] first4: 0.473996 1.292670 -0.063240 0.064850 [DiT] Step 36/50 t=0.300 [Debug] dit_step36_vt: [2170, 64] first4: -0.563639 -0.643330 -0.155598 0.696895 [Debug] dit_step36_xt: [2170, 64] first4: 0.485268 1.305537 -0.060128 0.050912 [DiT] Step 37/50 t=0.280 [Debug] dit_step37_vt: [2170, 64] first4: -0.555441 -0.734583 -0.119550 0.646566 [Debug] dit_step37_xt: [2170, 64] first4: 0.496377 1.320229 -0.057737 0.037981 [DiT] Step 38/50 t=0.260 [Debug] dit_step38_vt: [2170, 64] first4: -0.556594 -0.805234 -0.100350 0.613970 [Debug] dit_step38_xt: [2170, 64] first4: 0.507509 1.336333 -0.055730 0.025701 [DiT] Step 39/50 t=0.240 [Debug] dit_step39_vt: [2170, 64] first4: -0.550936 -0.888724 -0.057663 0.572371 [Debug] dit_step39_xt: [2170, 64] first4: 0.518528 1.354108 -0.054577 0.014254 [DiT] Step 40/50 t=0.220 [Debug] dit_step40_vt: [2170, 64] first4: -0.546231 -0.972852 0.007619 0.555106 [Debug] dit_step40_xt: [2170, 64] first4: 0.529452 1.373565 -0.054729 0.003152 [DiT] Step 41/50 t=0.200 [Debug] dit_step41_vt: [2170, 64] first4: -0.553115 -1.043566 0.022326 0.495447 [Debug] dit_step41_xt: [2170, 64] first4: 0.540515 1.394436 -0.055176 -0.006757 [DiT] Step 42/50 t=0.180 [Debug] dit_step42_vt: [2170, 64] first4: -0.533621 -1.096646 0.083822 0.438899 [Debug] dit_step42_xt: [2170, 64] first4: 0.551187 1.416369 -0.056852 -0.015535 [DiT] Step 43/50 t=0.160 [Debug] dit_step43_vt: [2170, 64] first4: -0.513800 -1.140474 0.091789 0.385346 [Debug] dit_step43_xt: [2170, 64] first4: 0.561463 1.439178 -0.058688 -0.023242 [DiT] Step 44/50 t=0.140 [Debug] dit_step44_vt: [2170, 64] first4: -0.495923 -1.183225 0.155197 0.338074 [Debug] dit_step44_xt: [2170, 64] first4: 0.571382 1.462843 -0.061792 -0.030004 [DiT] Step 45/50 t=0.120 [Debug] dit_step45_vt: [2170, 64] first4: -0.494819 -1.223503 0.195985 0.305891 [Debug] dit_step45_xt: [2170, 64] first4: 0.581278 1.487313 -0.065712 -0.036121 [DiT] Step 46/50 t=0.100 [Debug] dit_step46_vt: [2170, 64] first4: -0.446902 -1.278054 0.204986 0.242444 [Debug] dit_step46_xt: [2170, 64] first4: 0.590216 1.512874 -0.069811 -0.040970 [DiT] Step 47/50 t=0.080 [Debug] dit_step47_vt: [2170, 64] first4: -0.424753 -1.313939 0.243049 0.205183 [Debug] dit_step47_xt: [2170, 64] first4: 0.598711 1.539153 -0.074672 -0.045074 [DiT] Step 48/50 t=0.060 [Debug] dit_step48_vt: [2170, 64] first4: -0.375129 -1.338454 0.233699 0.156965 [Debug] dit_step48_xt: [2170, 64] first4: 0.606214 1.565922 -0.079346 -0.048213 [DiT] Step 49/50 t=0.040 [Debug] dit_step49_vt: [2170, 64] first4: -0.407742 -1.372256 0.234629 0.141079 [Debug] dit_x0: [2170, 64] first4: 0.614368 1.593367 -0.084039 -0.051035 [DiT] Step 50/50 t=0.020 [DiT-Generate] Total: 1507.0 ms (1507.0 ms/sample) [Debug] dit_output: [2170, 64] first4: 0.614368 1.593367 -0.084039 -0.051035 [Store] Unload DiT (1600.7 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: CUDA0 (CPU threads: 16) [VAE] Backend: CUDA0, Weight buffer: 161.1 MB [VAE] Loaded: 5 blocks, upsample=1920x, F32 activations [Store] Load VAE-Dec: 669 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: 426.9 ms [Debug] vae_audio: [2, 4166400] first4: -0.001575 -0.000918 -0.001169 -0.000848 [Store] Unload VAE-Dec (161.1 MB) [MP3] Encoding 86.8s @ 128 kbps, 48000 Hz stereo [MP3] 1388928 bytes (12.0:1), 395 ms (219.64x realtime), 32 threads [MP3] Wrote ggml-sft/request00.mp3 [Ace-Synth] All done 2026-07-12 10:09:04.313 | 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 10:09:06.141 | INFO | acestep.core.generation.handler.generate_music:generate_music:304 - [generate_music] Starting generation... 2026-07-12 10:09:06.141 | INFO | acestep.core.generation.handler.generate_music:generate_music:307 - [generate_music] Preparing inputs... 2026-07-12 10:09:06.141 | 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 10:09:06.144 | INFO | acestep.core.generation.handler.generate_music:_vram_preflight_check:156 - [generate_music] VRAM pre-flight: 81.69 GB free, ~0.94 GB needed (batch=1, duration=88s, mode=turbo). 2026-07-12 10:09:06.148 | INFO | acestep.core.generation.handler.conditioning_target:_prepare_target_latents_and_wavs:78 - [generate_music] Decoding audio codes for item 0... 2026-07-12 10:09:06.154 | 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 10:09:06.155 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:119 - ====================================================================== 2026-07-12 10:09:06.155 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:120 - πŸ” [DEBUG] DiT TEXT ENCODER INPUT (Inference) 2026-07-12 10:09:06.155 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:121 - ====================================================================== 2026-07-12 10:09:06.155 | 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 10:09:06.155 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:123 - ====================================================================== 2026-07-12 10:09:06.155 | 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 10:09:06.155 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:125 - ====================================================================== 2026-07-12 10:09:06.159 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:110 - [preprocess_batch] Inferring prompt embeddings... 2026-07-12 10:09:06.167 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:113 - [preprocess_batch] Inferring lyric embeddings... 2026-07-12 10:09:06.167 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:161 - [service_generate] Generating audio... (DiT backend: PyTorch (cuda)) 2026-07-12 10:09:06.174 | 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-Q8_0.gguf [GGML] Running acestep-v15-turbo-Q8_0.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 0.999797 lyric_embed 1.000000 enc_hidden 0.999795 detok_output 0.999983 context 0.999989 noise 1.000000 temb_t 0.999997 hidden_after_proj_in 0.999986 enc_after_cond_emb 0.999772 layer0_sa_output 0.999924 hidden_after_layer0 0.999953 hidden_after_layer6 0.999891 hidden_after_layer12 0.999356 hidden_after_layer18 0.996762 hidden_after_layer23 0.993862 dit_step0_vt 0.976249 dit_step0_xt 0.999948 dit_step1_vt 0.978901 dit_step1_xt 0.999833 dit_step2_vt 0.981066 dit_step2_xt 0.999549 dit_step3_vt 0.982308 dit_step3_xt 0.998917 dit_step4_vt 0.980633 dit_step4_xt 0.997490 dit_step5_vt 0.978045 dit_step5_xt 0.994297 dit_step6_vt 0.975257 dit_step6_xt 0.988258 dit_step7_vt 0.969408 dit_x0 0.979327 vae_audio 0.902175 vae_audio (STFT cosine) 0.975911 [Turbo] Error growth GGML vs Python stage cos max_err mean_err mean_A std_A mean_B std_B dit_step0_xt 0.999948 0.137988 0.006588 -0.002313 0.972927 -0.002342 0.972003 dit_step1_xt 0.999833 0.264729 0.011341 -0.005317 0.942672 -0.005313 0.941730 dit_step2_xt 0.999549 0.451595 0.017620 -0.009363 0.909166 -0.009311 0.908527 dit_step3_xt 0.998917 0.631703 0.026073 -0.014723 0.873781 -0.014577 0.873624 dit_step4_xt 0.997490 0.768069 0.037955 -0.021771 0.841924 -0.021660 0.841995 dit_step5_xt 0.994297 1.253134 0.055620 -0.031822 0.825202 -0.032109 0.824593 dit_step6_xt 0.988258 2.055632 0.082325 -0.046063 0.856116 -0.046482 0.855546 [SFT] steps=50, shift=1.0 | acestep-v15-sft-Q8_0.gguf [GGML] Running acestep-v15-sft-Q8_0.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-Q8_0.gguf: 830 tensors, data at offset 69088 [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-turbo-Q8_0.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: CUDA0 (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: 261 ms [Debug] text_hidden: [70, 1024] first4: 3.652379 1.048645 0.229843 -13.063987 [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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-turbo-Q8_0.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, 616.6 MB into backend [Load] CondEncoder: lyric(8L), timbre(4L, CLS), text_proj, null_cond [Store] Load CondEnc: 301 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, 14.9 ms [Debug] enc_hidden: [238, 2048] first4: 0.686921 -0.019141 -0.051705 0.022771 [Store] Unload CondEnc (616.6 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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-turbo-Q8_0.gguf: 830 tensors, data at offset 69088 [WeightCtx] Loaded 30 tensors, 106.5 MB into backend [Load] Detokenizer: FSQ(6->2048) + 2L encoder(S=5, 2048->64) [Store] Load FSQ-Detok: 54 ms CUDA Graph id 8 reused ggml_backend_cuda_graph_compute: CUDA graph warmup complete CUDA Graph id 8 reused [Context] Decoded: 434 codes -> 2170 frames (86.8s @ 25Hz) [Build-Context Batch0] Detokenizer: 60.8 ms, 434 codes [Debug] detok_output: [2170, 64] first4: -0.122364 1.438543 0.301741 -0.633332 [Store] Unload FSQ-Detok (106.5 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: -0.122364 1.438543 0.301741 -0.633332 [Init-Noise] Starting: T=2170, S=1085, enc_S=238, steps=8, batch=1 (cover) [Dedup] Previous line repeated 432 times total 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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-turbo-Q8_0.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, 4230.2 MB into backend [Load] DiT: 32 layers, H=2560, Nh=32/8, D=128 [Store] Load DiT: 2009 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: 0.014266 0.109556 0.062634 -0.073523 [Debug] temb: [2560] first4: 0.026310 -0.023038 -0.016457 0.047174 [Debug] temb_t: [2560] first4: 0.014577 -0.022289 -0.020778 0.048747 [Debug] temb_r: [2560] first4: 0.011734 -0.000750 0.004320 -0.001573 [Debug] sinusoidal_t: [256] first4: 0.562486 0.789701 0.439822 -0.023583 [Debug] sinusoidal_r: [256] first4: 1.000000 1.000000 1.000000 1.000000 [Debug] temb_lin1_t: [2560] first4: 0.063855 -0.019018 0.138265 -0.113400 [Debug] temb_lin1_r: [2560] first4: -0.009850 0.016687 -0.018316 0.012019 [Debug] hidden_after_proj_in: [2560, 1085] first4: 0.341374 -0.579036 0.228710 -0.189157 [Debug] proj_in_input: [192, 2170] first4: -0.122364 1.438543 0.301741 -0.633332 [Debug] enc_after_cond_emb: [2560, 238] first4: 0.034801 0.654381 0.103418 0.877866 [Debug] layer0_sa_input: [2560, 1085] first4: 0.038131 -0.045068 0.091024 -0.133648 [Debug] layer0_q_after_rope: [128, 32] first4: -0.253448 -2.088026 0.317381 0.200957 [Debug] layer0_k_after_rope: [128, 8] first4: 0.055759 2.146986 0.747280 0.541275 [Debug] layer0_sa_output: [2560, 1085] first4: 1.682696 4.311531 -0.538704 5.405484 [Debug] layer0_attn_out: [4096, 1085] first4: 0.475771 -2.245788 -0.166377 -2.789351 [Debug] layer0_after_self_attn: [2560, 1085] first4: 0.475771 -2.245788 -0.166377 -2.789351 [Debug] layer0_after_cross_attn: [2560, 1085] first4: 0.342660 -2.001023 0.006990 -3.745764 [Debug] hidden_after_layer0: [2560, 1085] first4: 1.810615 -2.453848 245.358444 -3.722621 [Debug] hidden_after_layer6: [2560, 1085] first4: -6.045307 -2.184943 324.945160 -12.258566 [Debug] hidden_after_layer12: [2560, 1085] first4: -11.012197 29.280060 268.710419 28.207893 [Debug] hidden_after_layer18: [2560, 1085] first4: 42.290783 13.852589 282.653198 4.488573 [Debug] hidden_after_layer31: [2560, 1085] first4: -78.541496 -222.186264 -158.589630 202.497192 [Debug] dit_step0_vt: [2170, 64] first4: -1.128976 1.416646 0.256877 1.149110 [Debug] dit_step0_xt: [2170, 64] first4: 0.245653 2.091857 -0.183551 0.795424 [DiT] Step 1/8 t=1.000 ggml_backend_cuda_graph_compute: CUDA graph warmup complete [Debug] dit_step1_vt: [2170, 64] first4: -0.868442 2.550477 0.800502 1.187592 [Debug] dit_step1_xt: [2170, 64] first4: 0.293023 1.952740 -0.227215 0.730646 [DiT] Step 2/8 t=0.955 [Debug] dit_step2_vt: [2170, 64] first4: -1.099642 2.142453 0.950984 0.933147 [Debug] dit_step2_xt: [2170, 64] first4: 0.366332 1.809910 -0.290614 0.668436 [DiT] Step 3/8 t=0.900 [Debug] dit_step3_vt: [2170, 64] first4: -1.374540 2.337109 1.118728 0.485022 [Debug] dit_step3_xt: [2170, 64] first4: 0.480877 1.615151 -0.383841 0.628018 [DiT] Step 4/8 t=0.833 [Debug] dit_step4_vt: [2170, 64] first4: -1.287986 2.234877 1.158340 0.364684 [Debug] dit_step4_xt: [2170, 64] first4: 0.618876 1.375700 -0.507949 0.588945 [DiT] Step 5/8 t=0.750 [Debug] dit_step5_vt: [2170, 64] first4: -1.151786 2.202667 0.891626 0.893187 [Debug] dit_step5_xt: [2170, 64] first4: 0.783416 1.061033 -0.635324 0.461346 [DiT] Step 6/8 t=0.643 [Debug] dit_step6_vt: [2170, 64] first4: -0.911400 1.856107 0.660934 0.881348 [Debug] dit_step6_xt: [2170, 64] first4: 0.965696 0.689812 -0.767511 0.285077 [DiT] Step 7/8 t=0.500 [Debug] dit_step7_vt: [2170, 64] first4: -0.875804 1.475198 0.735166 0.727372 [Debug] dit_x0: [2170, 64] first4: 1.228437 0.247253 -0.988061 0.066865 [DiT] Step 8/8 t=0.300 [DiT-Generate] Total: 629.2 ms (629.2 ms/sample) [Debug] dit_output: [2170, 64] first4: 1.228437 0.247253 -0.988061 0.066865 [Store] Unload DiT (4230.2 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: CUDA0 (CPU threads: 16) [VAE] Backend: CUDA0, Weight buffer: 161.1 MB [VAE] Loaded: 5 blocks, upsample=1920x, F32 activations [Store] Load VAE-Dec: 669 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: 428.0 ms [Debug] vae_audio: [2, 4166400] first4: 0.001670 0.004429 0.004696 0.003725 [Store] Unload VAE-Dec (161.1 MB) [MP3] Encoding 86.8s @ 128 kbps, 48000 Hz stereo [MP3] 1388928 bytes (12.0:1), 455 ms (190.68x realtime), 32 threads [MP3] Wrote ggml-xl-turbo/request00.mp3 [Ace-Synth] All done 2026-07-12 10:09:14.377 | 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 0.999797 lyric_embed 1.000000 enc_hidden 0.999795 detok_output 0.999983 context 0.999989 noise 1.000000 temb_t 0.999991 hidden_after_proj_in 0.999986 enc_after_cond_emb 0.999773 layer0_sa_output 0.999911 hidden_after_layer0 0.999958 hidden_after_layer6 0.999825 hidden_after_layer12 0.999446 hidden_after_layer18 0.998711 hidden_after_layer23 0.998971 dit_step0_vt 0.998921 dit_step0_xt 1.000000 dit_step5_vt 0.999478 dit_step5_xt 0.999980 dit_step10_vt 0.999114 dit_step10_xt 0.999935 dit_step15_vt 0.998221 dit_step15_xt 0.999847 dit_step20_vt 0.996374 dit_step20_xt 0.999619 dit_step25_vt 0.993020 dit_step25_xt 0.999150 dit_step30_vt 0.988791 dit_step30_xt 0.998276 dit_step35_vt 0.984350 dit_step35_xt 0.997056 dit_step40_vt 0.981508 dit_step40_xt 0.995739 dit_step45_vt 0.984771 dit_step45_xt 0.994758 dit_step49_vt 0.990024 dit_x0 0.994410 vae_audio 0.968909 vae_audio (STFT cosine) 0.987797 [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 0.020122 0.001458 -0.001654 0.978469 -0.001619 0.978872 dit_step5_xt 0.999980 0.076947 0.004335 -0.008371 0.884869 -0.008480 0.883811 dit_step10_xt 0.999935 0.123988 0.006856 -0.015091 0.801043 -0.015075 0.802324 dit_step15_xt 0.999847 0.170882 0.009640 -0.021775 0.730608 -0.021933 0.731304 dit_step20_xt 0.999619 0.231738 0.014065 -0.028493 0.677835 -0.028882 0.679066 dit_step25_xt 0.999150 0.332772 0.019913 -0.035292 0.647219 -0.035643 0.650780 dit_step30_xt 0.998276 0.564453 0.027687 -0.042140 0.642026 -0.042577 0.647107 dit_step35_xt 0.997056 0.818450 0.036967 -0.048993 0.662927 -0.049471 0.669256 dit_step40_xt 0.995739 1.071709 0.047320 -0.055873 0.707622 -0.056532 0.715783 dit_step45_xt 0.994758 1.297680 0.057113 -0.062859 0.771741 -0.063533 0.780575 [Xl-turbo] steps=8, shift=3.0 | acestep-v15-xl-turbo-Q8_0.gguf [GGML] Running acestep-v15-xl-turbo-Q8_0.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 10:09:19.055 | INFO | acestep.core.generation.handler.generate_music:generate_music:304 - [generate_music] Starting generation... 2026-07-12 10:09:19.055 | INFO | acestep.core.generation.handler.generate_music:generate_music:307 - [generate_music] Preparing inputs... 2026-07-12 10:09:19.055 | 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 10:09:19.057 | 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 10:09:19.062 | INFO | acestep.core.generation.handler.conditioning_target:_prepare_target_latents_and_wavs:78 - [generate_music] Decoding audio codes for item 0... 2026-07-12 10:09:19.067 | 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 10:09:19.069 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:119 - ====================================================================== 2026-07-12 10:09:19.069 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:120 - πŸ” [DEBUG] DiT TEXT ENCODER INPUT (Inference) 2026-07-12 10:09:19.069 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:121 - ====================================================================== 2026-07-12 10:09:19.069 | 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 10:09:19.069 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:123 - ====================================================================== 2026-07-12 10:09:19.069 | 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 10:09:19.069 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:125 - ====================================================================== 2026-07-12 10:09:19.072 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:110 - [preprocess_batch] Inferring prompt embeddings... 2026-07-12 10:09:19.080 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:113 - [preprocess_batch] Inferring lyric embeddings... 2026-07-12 10:09:19.080 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:161 - [service_generate] Generating audio... (DiT backend: PyTorch (cuda)) 2026-07-12 10:09:19.087 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:263 - [service_generate] DiT diffusion via PyTorch (cuda)... 2026-07-12 10:09:19.093 | 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 10:09:19.644 | INFO | acestep.core.generation.handler.generate_music_decode:_prepare_generate_music_decode_state:42 - [generate_music] Model generation completed. Decoding latents... 2026-07-12 10:09:19.645 | 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 10:09:19.645 | DEBUG | acestep.core.generation.handler.generate_music_decode:_prepare_generate_music_decode_state:65 - [generate_music] time_costs: {'encoder_time_cost': 0.006346702575683594, 'diffusion_time_cost': 0.5506305694580078, 'diffusion_per_step_time_cost': 0.06882882118225098, 'total_time_cost': 0.5569772720336914, 'offload_time_cost': 0.0} 2026-07-12 10:09:19.668 | INFO | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:127 - [generate_music] Decoding latents with VAE... 2026-07-12 10:09:19.670 | 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 10:09:19.671 | INFO | acestep.core.generation.handler.generate_music_decode:_decode_generate_music_pred_latents:154 - [generate_music] Effective free VRAM before VAE decode: 76.66 GB 2026-07-12 10:09:19.671 | 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 10:09:19.671 | DEBUG | acestep.core.generation.handler.memory_utils:_get_auto_decode_chunk_size:108 - [_get_auto_decode_chunk_size] Effective free VRAM: 76.66 GB 2026-07-12 10:09:19.671 | DEBUG | acestep.core.generation.handler.memory_utils:_should_offload_wav_to_cpu:131 - [_should_offload_wav_to_cpu] Effective free VRAM: 76.66 GB 2026-07-12 10:09:19.671 | 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 10:09:19.946 | 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 10:09:20.155 | 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 10:09:20.159 | 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-Q8_0.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-Q8_0.gguf: 830 tensors, data at offset 69056 [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-sft-Q8_0.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: CUDA0 (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: 344 ms [Debug] text_hidden: [70, 1024] first4: 3.652379 1.048645 0.229843 -13.063987 [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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-sft-Q8_0.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, 616.6 MB into backend [Load] CondEncoder: lyric(8L), timbre(4L, CLS), text_proj, null_cond [Store] Load CondEnc: 320 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, 15.2 ms [Debug] enc_hidden: [238, 2048] first4: 0.686921 -0.019141 -0.051705 0.022771 [Store] Unload CondEnc (616.6 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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-sft-Q8_0.gguf: 830 tensors, data at offset 69056 [WeightCtx] Loaded 30 tensors, 106.5 MB into backend [Load] Detokenizer: FSQ(6->2048) + 2L encoder(S=5, 2048->64) [Store] Load FSQ-Detok: 57 ms CUDA Graph id 8 reused ggml_backend_cuda_graph_compute: CUDA graph warmup complete CUDA Graph id 8 reused [Context] Decoded: 434 codes -> 2170 frames (86.8s @ 25Hz) [Build-Context Batch0] Detokenizer: 70.8 ms, 434 codes [Debug] detok_output: [2170, 64] first4: -0.122364 1.438543 0.301741 -0.633332 [Store] Unload FSQ-Detok (106.5 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: -0.122364 1.438543 0.301741 -0.633332 [Init-Noise] Starting: T=2170, S=1085, enc_S=238, steps=50, batch=1 (cover) [Dedup] Previous line repeated 432 times total 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: CUDA0 (CPU threads: 16) [GGUF] /mnt/workspace/git/acestep.cpp/models/acestep-v15-xl-sft-Q8_0.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, 4230.2 MB into backend [Load] DiT: 32 layers, H=2560, Nh=32/8, D=128 [Store] Load DiT: 1898 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: 0.012917 0.105451 0.074363 -0.032316 [Debug] temb: [2560] first4: 0.009484 -0.007646 -0.006176 0.004539 [Debug] temb_t: [2560] first4: 0.007823 -0.007880 -0.006709 0.003563 [Debug] temb_r: [2560] first4: 0.001661 0.000234 0.000533 0.000975 [Debug] sinusoidal_t: [256] first4: 0.562486 0.789701 0.439822 -0.023583 [Debug] sinusoidal_r: [256] first4: 1.000000 1.000000 1.000000 1.000000 [Debug] temb_lin1_t: [2560] first4: 0.033000 -0.082023 0.089453 -0.064865 [Debug] temb_lin1_r: [2560] first4: 0.006798 -0.005077 -0.002197 -0.011157 [Debug] hidden_after_proj_in: [2560, 1085] first4: 0.330759 -0.494607 0.099904 -0.212641 [Debug] proj_in_input: [192, 2170] first4: -0.122364 1.438543 0.301741 -0.633332 [Debug] enc_after_cond_emb: [2560, 238] first4: 0.037165 0.644126 0.115663 0.874807 [Debug] layer0_sa_input: [2560, 1085] first4: 0.036994 -0.019313 0.099315 -0.097456 [Debug] layer0_q_after_rope: [128, 32] first4: -0.284471 -2.143278 0.218408 -0.001867 [Debug] layer0_k_after_rope: [128, 8] first4: 0.725674 2.951337 1.173014 2.015921 [Debug] layer0_sa_output: [2560, 1085] first4: 1.896357 4.384690 -0.817818 7.851836 [Debug] layer0_attn_out: [4096, 1085] first4: 0.455381 -2.174521 -0.507434 -4.124167 [Debug] layer0_after_self_attn: [2560, 1085] first4: 0.455381 -2.174521 -0.507434 -4.124167 [Debug] layer0_after_cross_attn: [2560, 1085] first4: 0.288704 -1.901179 -0.425378 -4.206115 [Debug] hidden_after_layer0: [2560, 1085] first4: -1.724511 -1.234198 249.930939 -3.960173 [Debug] hidden_after_layer6: [2560, 1085] first4: -7.840183 3.509132 316.998413 -21.805424 [Debug] hidden_after_layer12: [2560, 1085] first4: -23.799778 39.207905 232.517548 16.782942 [Debug] hidden_after_layer18: [2560, 1085] first4: 9.565565 -13.875763 235.423264 20.098158 [Debug] hidden_after_layer31: [2560, 1085] first4: 64.088287 -107.347534 7.889065 82.122818 [Debug] dit_step0_vt: [2170, 64] first4: -0.323897 2.215685 0.006362 1.701386 [Debug] dit_step0_xt: [2170, 64] first4: 0.200814 2.111936 -0.172002 0.813629 [DiT] Step 1/50 t=1.000 ggml_backend_cuda_graph_compute: CUDA graph warmup complete [Debug] dit_step1_vt: [2170, 64] first4: -0.346586 2.145053 0.052706 1.562979 [Debug] dit_step1_xt: [2170, 64] first4: 0.207746 2.069035 -0.173056 0.782369 [DiT] Step 2/50 t=0.980 [Debug] dit_step2_vt: [2170, 64] first4: -0.321973 2.173384 0.040880 1.579658 [Debug] dit_step2_xt: [2170, 64] first4: 0.214185 2.025568 -0.173874 0.750776 [DiT] Step 3/50 t=0.960 [Debug] dit_step3_vt: [2170, 64] first4: -0.268776 2.136336 0.032855 1.570184 [Debug] dit_step3_xt: [2170, 64] first4: 0.219561 1.982841 -0.174531 0.719372 [DiT] Step 4/50 t=0.940 [Debug] dit_step4_vt: [2170, 64] first4: -0.278986 2.068728 0.002994 1.593700 [Debug] dit_step4_xt: [2170, 64] first4: 0.225140 1.941466 -0.174591 0.687498 [DiT] Step 5/50 t=0.920 [Debug] dit_step5_vt: [2170, 64] first4: -0.265318 2.055711 0.015125 1.592485 [Debug] dit_step5_xt: [2170, 64] first4: 0.230447 1.900352 -0.174893 0.655648 [DiT] Step 6/50 t=0.900 [Debug] dit_step6_vt: [2170, 64] first4: -0.262273 2.019941 0.052514 1.612211 [Debug] dit_step6_xt: [2170, 64] first4: 0.235692 1.859953 -0.175944 0.623404 [DiT] Step 7/50 t=0.880 [Debug] dit_step7_vt: [2170, 64] first4: -0.248257 1.986439 0.100383 1.618337 [Debug] dit_step7_xt: [2170, 64] first4: 0.240657 1.820224 -0.177951 0.591037 [DiT] Step 8/50 t=0.860 [Debug] dit_step8_vt: [2170, 64] first4: -0.252778 1.942454 0.131075 1.614361 [Debug] dit_step8_xt: [2170, 64] first4: 0.245713 1.781375 -0.180573 0.558750 [DiT] Step 9/50 t=0.840 [Debug] dit_step9_vt: [2170, 64] first4: -0.278263 1.874120 0.122342 1.611779 [Debug] dit_step9_xt: [2170, 64] first4: 0.251278 1.743893 -0.183020 0.526515 [DiT] Step 10/50 t=0.820 [Debug] dit_step10_vt: [2170, 64] first4: -0.290077 1.832345 0.116575 1.606775 [Debug] dit_step10_xt: [2170, 64] first4: 0.257080 1.707246 -0.185351 0.494379 [DiT] Step 11/50 t=0.800 [Debug] dit_step11_vt: [2170, 64] first4: -0.294658 1.779309 0.091395 1.585964 [Debug] dit_step11_xt: [2170, 64] first4: 0.262973 1.671660 -0.187179 0.462660 [DiT] Step 12/50 t=0.780 [Debug] dit_step12_vt: [2170, 64] first4: -0.285162 1.740839 0.100956 1.570114 [Debug] dit_step12_xt: [2170, 64] first4: 0.268676 1.636843 -0.189198 0.431258 [DiT] Step 13/50 t=0.760 [Debug] dit_step13_vt: [2170, 64] first4: -0.290540 1.682400 0.076403 1.563554 [Debug] dit_step13_xt: [2170, 64] first4: 0.274487 1.603195 -0.190726 0.399987 [DiT] Step 14/50 t=0.740 [Debug] dit_step14_vt: [2170, 64] first4: -0.296151 1.625634 0.066892 1.561286 [Debug] dit_step14_xt: [2170, 64] first4: 0.280410 1.570682 -0.192064 0.368761 [DiT] Step 15/50 t=0.720 [Debug] dit_step15_vt: [2170, 64] first4: -0.259364 1.582897 0.047121 1.561978 [Debug] dit_step15_xt: [2170, 64] first4: 0.285597 1.539024 -0.193007 0.337521 [DiT] Step 16/50 t=0.700 [Debug] dit_step16_vt: [2170, 64] first4: -0.259939 1.534289 0.022311 1.553477 [Debug] dit_step16_xt: [2170, 64] first4: 0.290796 1.508339 -0.193453 0.306452 [DiT] Step 17/50 t=0.680 [Debug] dit_step17_vt: [2170, 64] first4: -0.228059 1.487940 0.004053 1.536826 [Debug] dit_step17_xt: [2170, 64] first4: 0.295357 1.478580 -0.193534 0.275715 [DiT] Step 18/50 t=0.660 [Debug] dit_step18_vt: [2170, 64] first4: -0.191242 1.446421 -0.000028 1.544416 [Debug] dit_step18_xt: [2170, 64] first4: 0.299182 1.449651 -0.193533 0.244827 [DiT] Step 19/50 t=0.640 [Debug] dit_step19_vt: [2170, 64] first4: -0.177098 1.391099 -0.020943 1.527590 [Debug] dit_step19_xt: [2170, 64] first4: 0.302724 1.421829 -0.193114 0.214275 [DiT] Step 20/50 t=0.620 [Debug] dit_step20_vt: [2170, 64] first4: -0.144281 1.345747 -0.020077 1.526007 [Debug] dit_step20_xt: [2170, 64] first4: 0.305609 1.394915 -0.192713 0.183755 [DiT] Step 21/50 t=0.600 [Debug] dit_step21_vt: [2170, 64] first4: -0.102685 1.287338 -0.049962 1.512565 [Debug] dit_step21_xt: [2170, 64] first4: 0.307663 1.369168 -0.191714 0.153504 [DiT] Step 22/50 t=0.580 [Debug] dit_step22_vt: [2170, 64] first4: -0.088749 1.247683 -0.062848 1.510840 [Debug] dit_step22_xt: [2170, 64] first4: 0.309438 1.344214 -0.190457 0.123287 [DiT] Step 23/50 t=0.560 [Debug] dit_step23_vt: [2170, 64] first4: -0.058805 1.181172 -0.072903 1.480611 [Debug] dit_step23_xt: [2170, 64] first4: 0.310614 1.320590 -0.188999 0.093675 [DiT] Step 24/50 t=0.540 [Debug] dit_step24_vt: [2170, 64] first4: -0.029565 1.125860 -0.069945 1.458336 [Debug] dit_step24_xt: [2170, 64] first4: 0.311206 1.298073 -0.187600 0.064508 [DiT] Step 25/50 t=0.520 [Debug] dit_step25_vt: [2170, 64] first4: -0.034408 1.078380 -0.077417 1.426010 [Debug] dit_step25_xt: [2170, 64] first4: 0.311894 1.276506 -0.186051 0.035988 [DiT] Step 26/50 t=0.500 [Debug] dit_step26_vt: [2170, 64] first4: -0.039958 1.027042 -0.100991 1.423827 [Debug] dit_step26_xt: [2170, 64] first4: 0.312693 1.255965 -0.184032 0.007511 [DiT] Step 27/50 t=0.480 [Debug] dit_step27_vt: [2170, 64] first4: -0.054068 0.967112 -0.106331 1.398099 [Debug] dit_step27_xt: [2170, 64] first4: 0.313774 1.236623 -0.181905 -0.020451 [DiT] Step 28/50 t=0.460 [Debug] dit_step28_vt: [2170, 64] first4: -0.047908 0.885932 -0.103284 1.385783 [Debug] dit_step28_xt: [2170, 64] first4: 0.314732 1.218904 -0.179839 -0.048166 [DiT] Step 29/50 t=0.440 [Debug] dit_step29_vt: [2170, 64] first4: -0.061820 0.819368 -0.132532 1.351105 [Debug] dit_step29_xt: [2170, 64] first4: 0.315969 1.202517 -0.177189 -0.075189 [DiT] Step 30/50 t=0.420 [Debug] dit_step30_vt: [2170, 64] first4: -0.082367 0.729005 -0.175323 1.331093 [Debug] dit_step30_xt: [2170, 64] first4: 0.317616 1.187937 -0.173682 -0.101810 [DiT] Step 31/50 t=0.400 [Debug] dit_step31_vt: [2170, 64] first4: -0.092607 0.664416 -0.171970 1.310458 [Debug] dit_step31_xt: [2170, 64] first4: 0.319468 1.174648 -0.170243 -0.128020 [DiT] Step 32/50 t=0.380 [Debug] dit_step32_vt: [2170, 64] first4: -0.124651 0.590779 -0.172146 1.293164 [Debug] dit_step32_xt: [2170, 64] first4: 0.321961 1.162833 -0.166800 -0.153883 [DiT] Step 33/50 t=0.360 [Debug] dit_step33_vt: [2170, 64] first4: -0.119447 0.509159 -0.199530 1.253217 [Debug] dit_step33_xt: [2170, 64] first4: 0.324350 1.152650 -0.162809 -0.178947 [DiT] Step 34/50 t=0.340 [Debug] dit_step34_vt: [2170, 64] first4: -0.158831 0.395136 -0.208038 1.221351 [Debug] dit_step34_xt: [2170, 64] first4: 0.327527 1.144747 -0.158648 -0.203374 [DiT] Step 35/50 t=0.320 [Debug] dit_step35_vt: [2170, 64] first4: -0.155327 0.281306 -0.206876 1.196917 [Debug] dit_step35_xt: [2170, 64] first4: 0.330633 1.139121 -0.154511 -0.227313 [DiT] Step 36/50 t=0.300 [Debug] dit_step36_vt: [2170, 64] first4: -0.174936 0.198032 -0.180776 1.164100 [Debug] dit_step36_xt: [2170, 64] first4: 0.334132 1.135160 -0.150895 -0.250595 [DiT] Step 37/50 t=0.280 [Debug] dit_step37_vt: [2170, 64] first4: -0.199242 0.093252 -0.168666 1.148115 [Debug] dit_step37_xt: [2170, 64] first4: 0.338117 1.133295 -0.147522 -0.273557 [DiT] Step 38/50 t=0.260 [Debug] dit_step38_vt: [2170, 64] first4: -0.190556 -0.001631 -0.166306 1.082162 [Debug] dit_step38_xt: [2170, 64] first4: 0.341928 1.133328 -0.144196 -0.295200 [DiT] Step 39/50 t=0.240 [Debug] dit_step39_vt: [2170, 64] first4: -0.203725 -0.108529 -0.147637 1.053504 [Debug] dit_step39_xt: [2170, 64] first4: 0.346003 1.135498 -0.141243 -0.316270 [DiT] Step 40/50 t=0.220 [Debug] dit_step40_vt: [2170, 64] first4: -0.225330 -0.207608 -0.145610 0.993122 [Debug] dit_step40_xt: [2170, 64] first4: 0.350509 1.139650 -0.138331 -0.336133 [DiT] Step 41/50 t=0.200 [Debug] dit_step41_vt: [2170, 64] first4: -0.251918 -0.317092 -0.101744 0.964419 [Debug] dit_step41_xt: [2170, 64] first4: 0.355548 1.145992 -0.136296 -0.355421 [DiT] Step 42/50 t=0.180 [Debug] dit_step42_vt: [2170, 64] first4: -0.281289 -0.412267 -0.074401 0.937497 [Debug] dit_step42_xt: [2170, 64] first4: 0.361173 1.154238 -0.134808 -0.374171 [DiT] Step 43/50 t=0.160 [Debug] dit_step43_vt: [2170, 64] first4: -0.329995 -0.484943 -0.038571 0.885055 [Debug] dit_step43_xt: [2170, 64] first4: 0.367773 1.163936 -0.134037 -0.391872 [DiT] Step 44/50 t=0.140 [Debug] dit_step44_vt: [2170, 64] first4: -0.345038 -0.586601 -0.011561 0.836489 [Debug] dit_step44_xt: [2170, 64] first4: 0.374674 1.175668 -0.133805 -0.408602 [DiT] Step 45/50 t=0.120 [Debug] dit_step45_vt: [2170, 64] first4: -0.368256 -0.671218 0.061984 0.766272 [Debug] dit_step45_xt: [2170, 64] first4: 0.382039 1.189093 -0.135045 -0.423927 [DiT] Step 46/50 t=0.100 [Debug] dit_step46_vt: [2170, 64] first4: -0.406667 -0.738156 0.111749 0.719876 [Debug] dit_step46_xt: [2170, 64] first4: 0.390172 1.203856 -0.137280 -0.438325 [DiT] Step 47/50 t=0.080 [Debug] dit_step47_vt: [2170, 64] first4: -0.451282 -0.825371 0.160018 0.679846 [Debug] dit_step47_xt: [2170, 64] first4: 0.399198 1.220363 -0.140480 -0.451922 [DiT] Step 48/50 t=0.060 [Debug] dit_step48_vt: [2170, 64] first4: -0.521380 -0.880302 0.241190 0.659372 [Debug] dit_step48_xt: [2170, 64] first4: 0.409626 1.237969 -0.145304 -0.465109 [DiT] Step 49/50 t=0.040 [Debug] dit_step49_vt: [2170, 64] first4: -0.518100 -0.863318 0.162516 0.678071 [Debug] dit_x0: [2170, 64] first4: 0.419988 1.255236 -0.148555 -0.478671 [DiT] Step 50/50 t=0.020 [DiT-Generate] Total: 3723.2 ms (3723.2 ms/sample) [Debug] dit_output: [2170, 64] first4: 0.419988 1.255236 -0.148555 -0.478671 [Store] Unload DiT (4230.2 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: CUDA0 (CPU threads: 16) [VAE] Backend: CUDA0, Weight buffer: 161.1 MB [VAE] Loaded: 5 blocks, upsample=1920x, F32 activations [Store] Load VAE-Dec: 680 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: 426.8 ms [Debug] vae_audio: [2, 4166400] first4: -0.001262 -0.000604 -0.000712 -0.000224 [Store] Unload VAE-Dec (161.1 MB) [MP3] Encoding 86.8s @ 128 kbps, 48000 Hz stereo [MP3] 1388928 bytes (12.0:1), 433 ms (200.38x realtime), 32 threads [MP3] Wrote ggml-xl-sft/request00.mp3 [Ace-Synth] All done 2026-07-12 10:09:29.301 | 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 0.999797 lyric_embed 1.000000 enc_hidden 0.999528 detok_output 0.999983 context 0.999989 noise 1.000000 temb_t 0.999994 hidden_after_proj_in 0.999977 enc_after_cond_emb 0.999620 layer0_sa_output 0.999503 hidden_after_layer0 0.999391 hidden_after_layer6 0.999817 hidden_after_layer12 0.997528 hidden_after_layer18 0.996241 hidden_after_layer31 0.988312 dit_step0_vt 0.912281 dit_step0_xt 0.999842 dit_step1_vt 0.928889 dit_step1_xt 0.999531 dit_step2_vt 0.930097 dit_step2_xt 0.998766 dit_step3_vt 0.927149 dit_step3_xt 0.996969 dit_step4_vt 0.920821 dit_step4_xt 0.992465 dit_step5_vt 0.912660 dit_step5_xt 0.981715 dit_step6_vt 0.904232 dit_step6_xt 0.960388 dit_step7_vt 0.883222 dit_x0 0.925463 vae_audio 0.731549 vae_audio (STFT cosine) 0.922424 [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 0.159612 0.012474 -0.003014 0.978352 -0.002721 0.976534 dit_step1_xt 0.999531 0.370011 0.021246 -0.005993 0.952094 -0.005567 0.949004 dit_step2_xt 0.998766 0.544029 0.032927 -0.009728 0.923148 -0.009330 0.918931 dit_step3_xt 0.996969 0.806948 0.048938 -0.015404 0.891802 -0.014418 0.887241 dit_step4_xt 0.992465 1.193046 0.073908 -0.023301 0.862516 -0.021453 0.857796 dit_step5_xt 0.981715 1.796404 0.112628 -0.034922 0.846008 -0.032240 0.840033 dit_step6_xt 0.960388 2.556727 0.171440 -0.051931 0.869504 -0.049194 0.863964 [Xl-sft] steps=50, shift=1.0 | acestep-v15-xl-sft-Q8_0.gguf [GGML] Running acestep-v15-xl-sft-Q8_0.gguf... [GGML] Done, 131 dump files [Python] Initializing acestep-v15-xl-sft... Loading checkpoint shards: 0%| | 0/4 [00:00 2026-07-12 10:09:33.850 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:123 - ====================================================================== 2026-07-12 10:09:33.850 | 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 10:09:33.850 | INFO | acestep.core.generation.handler.conditioning_text:_prepare_text_conditioning_inputs:125 - ====================================================================== 2026-07-12 10:09:33.853 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:110 - [preprocess_batch] Inferring prompt embeddings... 2026-07-12 10:09:33.861 | INFO | acestep.core.generation.handler.conditioning_embed:preprocess_batch:113 - [preprocess_batch] Inferring lyric embeddings... 2026-07-12 10:09:33.861 | INFO | acestep.core.generation.handler.service_generate_execute:_execute_service_generate_diffusion:161 - [service_generate] Generating audio... (DiT backend: PyTorch (cuda)) 2026-07-12 10:09:33.867 | 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