202610301939_h3_FL2AV_0.4mp5s_omnixpu加速_B580_64G_冷启+热启.txt · 2026-10-30 19:39
| 阶段 | 耗时 | 步数 | s/it | staged MB | 其他指标 |
|---|---|---|---|---|---|
| 冷启 | 371.21 s | 4 | 22.11 | 4,965 | CUTE_BHLD_D128 ×3CUTE_H3_VAE_D64 ×3 |
| 热启 | 142.30 s | 4 | 21.95 | 4,965 | — |
[INFO] got prompt [INFO] Model storage policy: fast_disk=False paths=['<本地路径>/minimax_h3_audio_vae_fp32.safetensors'] [INFO] VAE load device: xpu:0, offload device: cpu, dtype: torch.float32 [INFO] Model storage policy: fast_disk=False paths=['<本地路径>/minimax_h3_video_vae_fp16.safetensors'] [INFO] VAE load device: xpu:0, offload device: cpu, dtype: torch.float16 [INFO] gguf qtypes: Q4_K (390), F32 (433), Q6_K (50), Q5_K (27), F16 (2) [INFO] Attenpting to find mmproj file for text encoder... [ERROR] Error: Can't find mmproj file for 'qwen3vl_32b_minimax_h3-Q4_K_M.gguf' (matching:'qwen3vl_32b_minimax_h3')! Qwen-Image-Edit will be broken! [INFO] Model storage policy: fast_disk=False paths=[] [INFO] CLIP/text encoder model load device: xpu:0, offload device: cpu, current: cpu, dtype: torch.float16 [INFO] Requested to load MiniMaxH3TEModel_ [INFO] loaded partially; 7031.00 MB usable, 6679.44 MB loaded, 8703.96 MB offloaded, 351.56 MB buffer reserved, lowvram patches: 0 [INFO] Attempting to release mmap (185) [INFO] [OmniXPU] norm first use: LayerNorm shape=torch.Size([1620, 1152]) [omni_xpu::device] warning: physical Device ID 0xe20b (physical_sku=b580) uses kernel_profile=b580, policy_id=b580-v1, support=experimental; functional support may be accepted independently, performance_claim=false [INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=221.8MiB [INFO] Requested to load MiniMaxH3VideoVAE [INFO] Unloaded partially: 3653.38 MB freed, 3026.05 MB remains loaded, 351.56 MB buffer reserved, lowvram patches: 0 [INFO] Model MiniMaxH3VideoVAE prepared for dynamic VRAM loading. 4965MB Staged. 0 patches attached. Force pre-loaded 128 weights: 352 KB. [INFO] gguf qtypes: F32 (7), F16 (278), Q3_K (250) [INFO] model weight dtype torch.bfloat16, manual cast: None [INFO] model_type FLOW_AV [INFO] Model storage policy: fast_disk=False paths=[] [INFO] [OmniXPU] LoRA memory budget cached: source=1865.5MiB/416 tensors model_added=1865.5MiB model_budget=1865.5MiB clip_added=0.0MiB clip_budget=0.0MiB [INFO] [OmniXPU] LoRA model budget applied: cached=1865.5MiB/416 tensors requested=6620.2->8485.7MiB minimum=3310.1->5175.6MiB xpu_memory=available xpu_allocated=1543.7MiB xpu_reserved=1572.0MiB xpu_free=7282.7MiB xpu_total=11946.2MiB [INFO] Requested to load MiniMaxH3 [INFO] loaded partially; 4778.36 MB usable, 4588.87 MB loaded, 10753.11 MB offloaded, 189.49 MB buffer reserved, lowvram patches: 0 [INFO] [OmniXPU] LoRA model load complete: xpu_memory=available xpu_allocated=4110.6MiB xpu_reserved=4376.0MiB xpu_free=5968.2MiB xpu_total=11946.2MiB [WARNING] [OmniXPU] experimental capability-driven BMG D128 CUTE route enabled; set OMNI_ATTN_BACKEND=torch to disable it if this workflow shows an issue [INFO] [OmniXPU] attention CUTE_BHLD_D128 #1: heads=56 q=427 kv=427 dtype=torch.bfloat16 [INFO] [OmniXPU] attention CUTE_BHLD_D128 #2: heads=56 q=427 kv=427 dtype=torch.bfloat16 0%| | 0/4 [00:00<?, ?it/s, Model Initializing ... ][INFO] [OmniXPU] attention CUTE_BHLD_D128 #3: heads=56 q=16231 kv=16231 dtype=torch.bfloat16 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 4/4 [01:28<00:00, 22.11s/it] [INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=221.7MiB [INFO] Requested to load MiniMaxH3AudioVAE [INFO] Model MiniMaxH3AudioVAE prepared for dynamic VRAM loading. 576MB Staged. 0 patches attached. Force pre-loaded 401 weights: 539 KB. [INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=762.6MiB [INFO] Requested to load MiniMaxH3VideoVAE [INFO] Unloaded partially: 2589.47 MB freed, 1999.40 MB remains loaded, 220.50 MB buffer reserved, lowvram patches: 41 [INFO] Model MiniMaxH3VideoVAE prepared for dynamic VRAM loading. 4965MB Staged. 0 patches attached. Force pre-loaded 128 weights: 352 KB. [INFO] [OmniXPU] attention CUTE_H3_VAE_D64 #1: heads=32 seq=1797 dtype=torch.float16 [INFO] [OmniXPU] attention CUTE_H3_VAE_D64 #2: heads=32 seq=1797 dtype=torch.float16 [INFO] [OmniXPU] attention CUTE_H3_VAE_D64 #3: heads=32 seq=1797 dtype=torch.float16 [INFO] Prompt executed in 371.21 seconds [INFO] got prompt [INFO] [OmniXPU] LoRA model budget applied: cached=1865.5MiB/416 tensors requested=6620.2->8485.7MiB minimum=3310.1->5175.6MiB xpu_memory=available xpu_allocated=1505.7MiB xpu_reserved=1730.0MiB xpu_free=4321.5MiB xpu_total=11946.2MiB [INFO] loaded partially; 4976.40 MB usable, 4779.59 MB loaded, 10562.40 MB offloaded, 189.49 MB buffer reserved, lowvram patches: 0 [INFO] [OmniXPU] LoRA model load complete: xpu_memory=available xpu_allocated=4315.5MiB xpu_reserved=4380.0MiB xpu_free=5362.7MiB xpu_total=11946.2MiB 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 4/4 [01:27<00:00, 21.95s/it] [INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=221.7MiB [INFO] Requested to load MiniMaxH3AudioVAE [INFO] Model MiniMaxH3AudioVAE prepared for dynamic VRAM loading. 576MB Staged. 0 patches attached. Force pre-loaded 401 weights: 539 KB. [INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=762.6MiB [INFO] Requested to load MiniMaxH3VideoVAE [INFO] Unloaded partially: 2780.19 MB freed, 1999.40 MB remains loaded, 220.50 MB buffer reserved, lowvram patches: 85 [INFO] Model MiniMaxH3VideoVAE prepared for dynamic VRAM loading. 4965MB Staged. 0 patches attached. Force pre-loaded 128 weights: 352 KB. [INFO] Prompt executed in 142.30 seconds