MiniMaxMusic3TextEncode: Expected a cuda device, but got: cpu

[LocalCount]

Custom Node Testing

• I have tried disabling custom nodes and the issue persists (see how to disable custom nodes if you need help)

Expected Behavior

.

Actual Behavior

.

Steps to Reproduce

Default Minimax Music3 workflow on Windows 11 on GTX 1660S

Debug Logs

Node threw an error during execution.


# ComfyUI Error Report
## Error Details
- **Node ID:** 37:13
- **Node Type:** MiniMaxMusic3TextEncode
- **Exception Type:** ValueError
- **Exception Message:** ValueError: Expected a cuda device, but got: cpu

## Stack Trace

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\execution.py", line 545, in execute
    output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
                                                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\execution.py", line 344, in get_output_data
    return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\execution.py", line 318, in _async_map_node_over_list
    await process_inputs(input_dict, i)

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\execution.py", line 306, in process_inputs
    result = f(**inputs)

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy_api\internal\__init__.py", line 149, in wrapped_func
    return method(locked_class, **inputs)

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy_api\latest\_io.py", line 1990, in EXECUTE_NORMALIZED
    to_return = cls.execute(*args, **kwargs)

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy_extras\nodes_minimax_music.py", line 37, in execute
    conditioning = clip.encode_from_tokens_scheduled(tokens)

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy\sd.py", line 340, in encode_from_tokens_scheduled
    pooled_dict = self.encode_from_tokens(tokens, return_pooled=return_pooled, return_dict=True)

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy\sd.py", line 409, in encode_from_tokens
    o = self.cond_stage_model.encode_token_weights(tokens)

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy\text_encoders\minimax_music.py", line 98, in encode_token_weights
    hidden = self.generate(input_ids, seed, max_audio_f****es, self.execution_device, cfg_scale, top_k)

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy\ldm\minimax_music\ar.py", line 252, in generate
    past = self.model.init_kv_cache(2, prompt_tokens + decode_limit + 1, device, execution_dtype)

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy\text_encoders\llama.py", line 794, in init_kv_cache
    fixed_kv = self.fixed_kv and comfy_kitchen.flash_attention_decode_is_available(device)
                                 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Lib\site-packages\comfy_kitchen\flash_attention.py", line 22, in is_available
    return torch.cuda.get_device_capability(device) >= _MINIMUM_CAPABILITY
           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Lib\site-packages\torch\cuda\__init__.py", line 615, in get_device_capability
    prop = get_device_properties(device)

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Lib\site-packages\torch\cuda\__init__.py", line 633, in get_device_properties
    device = _get_device_index(device, optional=True)

  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Lib\site-packages\torch\cuda\_utils.py", line 525, in _get_device_index
    raise ValueError(f"Expected a cuda device, but got: {device}")


## System Information
- **ComfyUI Version:** 0.33.0
- **Arguments:** ComfyUI\main.py --feature-flag show_signin_button=true --enable-manager --extra-model-paths-config C:\Users\****y\AppData\Roaming\Comfy Desktop\shared_model_paths.yaml --input-directory C:\Users\****y\ComfyUI-Shared\input --output-directory C:\Users\****y\ComfyUI-Shared\output
- **OS:** win32
- **Python Version:** 3.13.12 (main, Feb 12 2026, 00:38:53) [MSC v.1944 64 bit (AMD64)]
- **Embedded Python:** false
- **PyTorch Version:** 2.10.0+cu130
## Devices

- **Name:** cuda:0 NVIDIA GeForce GTX 1660 SUPER : cudaMallocAsync
  - **Type:** cuda
  - **V**** Total:** 6441992192
  - **V**** Free:** 5387583488
  - **Torch V**** Total:** 0
  - **Torch V**** Free:** 0

## Logs

2026-08-14T08:33:36.113993 - �[32m[INFO]�[0m setup plugin alembic.autogenerate.schemas
2026-08-14T08:33:36.114100 - �[32m[INFO]�[0m setup plugin alembic.autogenerate.tables
2026-08-14T08:33:36.114149 - �[32m[INFO]�[0m setup plugin alembic.autogenerate.types
2026-08-14T08:33:36.114196 - �[32m[INFO]�[0m setup plugin alembic.autogenerate.constraints
2026-08-14T08:33:36.114239 - �[32m[INFO]�[0m setup plugin alembic.autogenerate.defaults
2026-08-14T08:33:36.114283 - �[32m[INFO]�[0m setup plugin alembic.autogenerate.comments
2026-08-14T08:33:36.709094 - �[32m[INFO]�[0m Adding extra search path checkpoints C:\Users\****y\ComfyUI-Shared\models\checkpoints
2026-08-14T08:33:36.709150 - �[32m[INFO]�[0m Adding extra search path classifiers C:\Users\****y\ComfyUI-Shared\models\classifiers
2026-08-14T08:33:36.709180 - �[32m[INFO]�[0m Adding extra search path clip_vision C:\Users\****y\ComfyUI-Shared\models\clip_vision
2026-08-14T08:33:36.709205 - �[32m[INFO]�[0m Adding extra search path configs C:\Users\****y\ComfyUI-Shared\models\configs
2026-08-14T08:33:36.709227 - �[32m[INFO]�[0m Adding extra search path controlnet C:\Users\****y\ComfyUI-Shared\models\controlnet
2026-08-14T08:33:36.709253 - �[32m[INFO]�[0m Adding extra search path controlnet C:\Users\****y\ComfyUI-Shared\models\t2i_adapter
2026-08-14T08:33:36.709274 - �[32m[INFO]�[0m Adding extra search path diffusers C:\Users\****y\ComfyUI-Shared\models\diffusers
2026-08-14T08:33:36.709295 - �[32m[INFO]�[0m Adding extra search path diffusion_models C:\Users\****y\ComfyUI-Shared\models\diffusion_models
2026-08-14T08:33:36.709316 - �[32m[INFO]�[0m Adding extra search path embeddings C:\Users\****y\ComfyUI-Shared\models\embeddings
2026-08-14T08:33:36.709336 - �[32m[INFO]�[0m Adding extra search path gligen C:\Users\****y\ComfyUI-Shared\models\gligen
2026-08-14T08:33:36.709356 - �[32m[INFO]�[0m Adding extra search path hypernetworks C:\Users\****y\ComfyUI-Shared\models\hypernetworks
2026-08-14T08:33:36.709377 - �[32m[INFO]�[0m Adding extra search path latent_upscale_models C:\Users\****y\ComfyUI-Shared\models\latent_upscale_models
2026-08-14T08:33:36.709396 - �[32m[INFO]�[0m Adding extra search path loras C:\Users\****y\ComfyUI-Shared\models\loras
2026-08-14T08:33:36.709418 - �[32m[INFO]�[0m Adding extra search path model_patches C:\Users\****y\ComfyUI-Shared\models\model_patches
2026-08-14T08:33:36.709438 - �[32m[INFO]�[0m Adding extra search path audio_encoders C:\Users\****y\ComfyUI-Shared\models\audio_encoders
2026-08-14T08:33:36.709457 - �[32m[INFO]�[0m Adding extra search path photomaker C:\Users\****y\ComfyUI-Shared\models\photomaker
2026-08-14T08:33:36.709476 - �[32m[INFO]�[0m Adding extra search path style_models C:\Users\****y\ComfyUI-Shared\models\style_models
2026-08-14T08:33:36.709498 - �[32m[INFO]�[0m Adding extra search path text_encoders C:\Users\****y\ComfyUI-Shared\models\text_encoders
2026-08-14T08:33:36.709516 - �[32m[INFO]�[0m Adding extra search path upscale_models C:\Users\****y\ComfyUI-Shared\models\upscale_models
2026-08-14T08:33:36.709535 - �[32m[INFO]�[0m Adding extra search path background_removal C:\Users\****y\ComfyUI-Shared\models\background_removal
2026-08-14T08:33:36.709555 - �[32m[INFO]�[0m Adding extra search path f****e_interpolation C:\Users\****y\ComfyUI-Shared\models\f****e_interpolation
2026-08-14T08:33:36.709574 - �[32m[INFO]�[0m Adding extra search path geometry_estimation C:\Users\****y\ComfyUI-Shared\models\geometry_estimation
2026-08-14T08:33:36.709593 - �[32m[INFO]�[0m Adding extra search path optical_flow C:\Users\****y\ComfyUI-Shared\models\optical_flow
2026-08-14T08:33:36.709612 - �[32m[INFO]�[0m Adding extra search path detection C:\Users\****y\ComfyUI-Shared\models\detection
2026-08-14T08:33:36.709630 - �[32m[INFO]�[0m Adding extra search path vae C:\Users\****y\ComfyUI-Shared\models\vae
2026-08-14T08:33:36.709650 - �[32m[INFO]�[0m Adding extra search path vae_approx C:\Users\****y\ComfyUI-Shared\models\vae_approx
2026-08-14T08:33:36.709669 - �[32m[INFO]�[0m Adding extra search path CogVideo C:\Users\****y\ComfyUI-Shared\models\CogVideo
2026-08-14T08:33:36.709689 - �[32m[INFO]�[0m Adding extra search path Joy_caption C:\Users\****y\ComfyUI-Shared\models\Joy_caption
2026-08-14T08:33:36.709708 - �[32m[INFO]�[0m Adding extra search path LLM C:\Users\****y\ComfyUI-Shared\models\LLM
2026-08-14T08:33:36.709730 - �[32m[INFO]�[0m Adding extra search path SEEDVR2 C:\Users\****y\ComfyUI-Shared\models\SEEDVR2
2026-08-14T08:33:36.709749 - �[32m[INFO]�[0m Adding extra search path animatediff_models C:\Users\****y\ComfyUI-Shared\models\animatediff_models
2026-08-14T08:33:36.709774 - �[32m[INFO]�[0m Adding extra search path animatediff_motion_lora C:\Users\****y\ComfyUI-Shared\models\animatediff_motion_lora
2026-08-14T08:33:36.709796 - �[32m[INFO]�[0m Adding extra search path animatediff_video_formats C:\Users\****y\ComfyUI-Shared\models\animatediff_video_formats
2026-08-14T08:33:36.709815 - �[32m[INFO]�[0m Adding extra search path audio C:\Users\****y\ComfyUI-Shared\models\audio
2026-08-14T08:33:36.709834 - �[32m[INFO]�[0m Adding extra search path blip C:\Users\****y\ComfyUI-Shared\models\blip
2026-08-14T08:33:36.709852 - �[32m[INFO]�[0m Adding extra search path insightface C:\Users\****y\ComfyUI-Shared\models\insightface
2026-08-14T08:33:36.709871 - �[32m[INFO]�[0m Adding extra search path layerstyle C:\Users\****y\ComfyUI-Shared\models\layerstyle
2026-08-14T08:33:36.709896 - �[32m[INFO]�[0m Adding extra search path liveportrait C:\Users\****y\ComfyUI-Shared\models\liveportrait
2026-08-14T08:33:36.709915 - �[32m[INFO]�[0m Adding extra search path sams C:\Users\****y\ComfyUI-Shared\models\sams
2026-08-14T08:33:36.709934 - �[32m[INFO]�[0m Adding extra search path ultralytics C:\Users\****y\ComfyUI-Shared\models\ultralytics
2026-08-14T08:33:36.709952 - �[32m[INFO]�[0m Adding extra search path xlabs C:\Users\****y\ComfyUI-Shared\models\xlabs
2026-08-14T08:33:36.709970 - �[32m[INFO]�[0m Adding extra search path yolo C:\Users\****y\ComfyUI-Shared\models\yolo
2026-08-14T08:33:36.709990 - �[32m[INFO]�[0m Adding extra search path clip C:\Users\****y\ComfyUI-Shared\models\clip
2026-08-14T08:33:36.710008 - �[32m[INFO]�[0m Adding extra search path unet C:\Users\****y\ComfyUI-Shared\models\unet
2026-08-14T08:33:36.710036 - �[32m[INFO]�[0m Setting output directory to: C:\Users\****y\ComfyUI-Shared\output
2026-08-14T08:33:36.710068 - �[32m[INFO]�[0m Setting input directory to: C:\Users\****y\ComfyUI-Shared\input
2026-08-14T08:33:37.409177 - [START] Security scan2026-08-14T08:33:37.409190 - 
2026-08-14T08:33:39.358084 - [DONE] Security scan2026-08-14T08:33:39.358098 - 
2026-08-14T08:33:39.359582 - ** ComfyUI startup time:2026-08-14T08:33:39.359625 -  2026-08-14T08:33:39.359651 - 2026-08-14 08:33:39.3592026-08-14T08:33:39.359674 - 
2026-08-14T08:33:39.359700 - ** Platform:2026-08-14T08:33:39.359722 -  2026-08-14T08:33:39.359741 - Windows2026-08-14T08:33:39.359761 - 
2026-08-14T08:33:39.359782 - ** Python version:2026-08-14T08:33:39.359800 -  2026-08-14T08:33:39.359819 - 3.13.12 (main, Feb 12 2026, 00:38:53) [MSC v.1944 64 bit (AMD64)]2026-08-14T08:33:39.359838 - 
2026-08-14T08:33:39.359858 - ** Python executable:2026-08-14T08:33:39.359877 -  2026-08-14T08:33:39.359897 - C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Scripts\python.exe2026-08-14T08:33:39.359916 - 
2026-08-14T08:33:39.359936 - ** ComfyUI Path:2026-08-14T08:33:39.359954 -  2026-08-14T08:33:39.359973 - C:\Users\****y\ComfyUI-Installs\Local\ComfyUI2026-08-14T08:33:39.359992 - 
2026-08-14T08:33:39.360013 - ** ComfyUI Base Folder Path:2026-08-14T08:33:39.360032 -  2026-08-14T08:33:39.360050 - C:\Users\****y\ComfyUI-Installs\Local\ComfyUI2026-08-14T08:33:39.360069 - 
2026-08-14T08:33:39.360089 - ** User directory:2026-08-14T08:33:39.360107 -  2026-08-14T08:33:39.360126 - C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\user2026-08-14T08:33:39.360144 - 
2026-08-14T08:33:39.360164 - ** ComfyUI-Manager config path:2026-08-14T08:33:39.360182 -  2026-08-14T08:33:39.360360 - C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\user\__manager\config.ini2026-08-14T08:33:39.360440 - 
2026-08-14T08:33:39.360482 - ** Log path:2026-08-14T08:33:39.360502 -  2026-08-14T08:33:39.360520 - C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\user\comfyui.log2026-08-14T08:33:39.360537 - 
2026-08-14T08:33:40.389528 - �[32m[INFO]�[0m [PRE] ComfyUI-Manager
2026-08-14T08:33:45.987035 - �[32m[INFO]�[0m Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope1_', 'apply_rope_', 'apply_rope_split_half', 'apply_rope_split_half1', 'apply_rope_split_half1_', 'apply_rope_split_half_', 'convrot_w4a4_linear', 'dequantize_convrot_w4a4_weight', 'dequantize_int8_convrot_weight', 'dequantize_int8_convrot_weight_dtype', 'dequantize_int8_embedding', 'dequantize_int8_simple', 'dequantize_int8_simple_dtype', 'dequantize_mxfp8', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'dequantize_w4a8_int8_weight', 'gemv_awq_w4a16', 'int8_linear', 'na3d', 'prepare_int4_weight_for_int8_linear', 'quantize_and_rotate_rowwise', 'quantize_convrot_w4a4_weight', 'quantize_int8_convrot_weight', 'quantize_int8_rowwise', 'quantize_int8_tensorwise', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'quantize_svdquant_w4a4', 'quantize_w4a8_int8_weight', 'rms_adaln', 'rms_rope', 'rms_rope1', 'rms_rope1_', 'rms_rope_', 'rms_rope_split_half', 'rms_rope_split_half1', 'rms_rope_split_half1_', 'rms_rope_split_half_', 'rotate_int8_convrot_weight', 'scaled_mm_mxfp8', 'scaled_mm_nvfp4', 'scaled_mm_svdquant_w4a4', 'stochastic_rounding_fp8', 'w4a8_int8_linear']}
2026-08-14T08:33:45.987155 - �[32m[INFO]�[0m Found comfy_kitchen backend hip: {'available': False, 'disabled': False, 'unavailable_reason': 'PyTorch ROCm/HIP runtime not available', 'capabilities': []}
2026-08-14T08:33:45.987211 - �[32m[INFO]�[0m Found comfy_kitchen backend cuda: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope1_', 'apply_rope_', 'apply_rope_split_half', 'apply_rope_split_half1', 'apply_rope_split_half1_', 'apply_rope_split_half_', 'convrot_w4a4_linear', 'dequantize_convrot_w4a4_weight', 'dequantize_int8_convrot_weight', 'dequantize_int8_convrot_weight_dtype', 'dequantize_int8_simple', 'dequantize_int8_simple_dtype', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'dequantize_w4a8_int8_weight', 'gemv_awq_w4a16', 'int8_linear', 'na3d', 'prepare_int4_weight_for_int8_linear', 'quantize_and_rotate_rowwise', 'quantize_convrot_w4a4_weight', 'quantize_int8_convrot_weight', 'quantize_int8_rowwise', 'quantize_int8_tensorwise', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'quantize_svdquant_w4a4', 'quantize_w4a8_int8_weight', 'rms_adaln', 'rms_rope', 'rms_rope1', 'rms_rope1_', 'rms_rope_', 'rms_rope_split_half', 'rms_rope_split_half1', 'rms_rope_split_half1_', 'rms_rope_split_half_', 'rotate_int8_convrot_weight', 'scaled_mm_nvfp4', 'scaled_mm_svdquant_w4a4', 'stochastic_rounding_fp8', 'w4a8_int8_linear']}
2026-08-14T08:33:45.987262 - �[32m[INFO]�[0m Found comfy_kitchen backend triton: {'available': False, 'disabled': True, 'unavailable_reason': "ImportError: No module named 'triton'", 'capabilities': []}
2026-08-14T08:33:46.018120 - �[32m[INFO]�[0m Checkpoint files will always be loaded safely.
2026-08-14T08:33:46.160438 - �[32m[INFO]�[0m Total V**** 6144 MB, total **** 32669 MB
2026-08-14T08:33:46.160583 - �[32m[INFO]�[0m pytorch version: 2.10.0+cu130
2026-08-14T08:33:46.161046 - �[32m[INFO]�[0m Set v**** state to: NORMAL_V****
2026-08-14T08:33:46.161428 - �[32m[INFO]�[0m Device: cuda:0 NVIDIA GeForce GTX 1660 SUPER : cudaMallocAsync
2026-08-14T08:33:46.181151 - �[32m[INFO]�[0m Using async weight offloading with 2 streams
2026-08-14T08:33:46.181507 - �[32m[INFO]�[0m Enabled pinned memory 13067.0
2026-08-14T08:33:46.282976 - �[32m[INFO]�[0m working around nvidia conv3d memory bug.
2026-08-14T08:33:46.302515 - �[32m[INFO]�[0m Using pytorch attention
2026-08-14T08:34:05.169563 - �[32m[INFO]�[0m aimdo: src-win/cuda-detour.c:38:INFO:aimdo_setup_hooks: installing 6 hooks
2026-08-14T08:34:05.203687 - �[32m[INFO]�[0m aimdo: src/control.c:262:INFO:comfy-aimdo NVML pressure enabled
2026-08-14T08:34:05.235450 - �[32m[INFO]�[0m aimdo: src-win/shmem-detect.c:80:INFO:comfy-aimdo WDDM adapter match: NVIDIA GeForce GTX 1660 SUPER runtime_luid=00000000:0000ebfd dxgi_luid=00000000:0000ebfd
2026-08-14T08:34:05.235561 - �[32m[INFO]�[0m aimdo: src/control.c:277:INFO:comfy-aimdo inited for GPU: NVIDIA GeForce GTX 1660 SUPER (V****: 6143 MB)
2026-08-14T08:34:05.235644 - �[32m[INFO]�[0m DynamicV**** support detected and enabled
2026-08-14T08:34:05.235695 - �[32m[INFO]�[0m Python version: 3.13.12 (main, Feb 12 2026, 00:38:53) [MSC v.1944 64 bit (AMD64)]
2026-08-14T08:34:05.235734 - �[32m[INFO]�[0m ComfyUI version: 0.33.0
2026-08-14T08:34:05.236426 - �[32m[INFO]�[0m comfy-aimdo version: 0.4.13
2026-08-14T08:34:05.238286 - �[32m[INFO]�[0m comfy-kitchen version: 0.2.31
2026-08-14T08:34:05.242936 - �[32m[INFO]�[0m comfyui-frontend-package version: 1.49.6
2026-08-14T08:34:05.243027 - �[32m[INFO]�[0m comfyui-workflow-templates version: 0.11.41
2026-08-14T08:34:05.243079 - �[32m[INFO]�[0m comfyui-embedded-docs version: 0.5.9
2026-08-14T08:34:05.243133 - �[32m[INFO]�[0m comfy-kitchen version: 0.2.31
2026-08-14T08:34:05.243179 - �[32m[INFO]�[0m comfy-aimdo version: 0.4.13
2026-08-14T08:34:05.245423 - �[32m[INFO]�[0m [Prompt Server] web root: C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Lib\site-packages\comfyui_frontend_package\static
2026-08-14T08:34:05.246264 - �[32m[INFO]�[0m Asset seeder disabled
2026-08-14T08:34:05.246376 - �[32m[INFO]�[0m [START] ComfyUI-Manager
2026-08-14T08:34:05.390593 - [ComfyUI-Manager] Using GitPython backend2026-08-14T08:34:05.390634 - 
2026-08-14T08:34:05.434991 - �[32m[INFO]�[0m [ComfyUI-Manager] network_mode: public
2026-08-14T08:34:05.445198 - �[1m�[33m[WARNING]�[0m [ComfyUI-Manager] The matrix sharing feature has been disabled because the `matrix-nio` dependency is not installed.
	To use this feature, please run the following command:
	C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Scripts\python.exe -m pip install matrix-nio

2026-08-14T08:34:07.263892 - �[32m[INFO]�[0m No OpenGL_accelerate module loaded: No module named 'OpenGL_accelerate'
2026-08-14T08:34:08.154620 - �[32m[INFO]�[0m 
Import times for custom nodes:
2026-08-14T08:34:08.154720 - �[32m[INFO]�[0m    0.0 seconds: C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\custom_nodes\websocket_image_save.py
2026-08-14T08:34:08.154768 - �[32m[INFO]�[0m    0.0 seconds: C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\custom_nodes\comfyui-node-organizer
2026-08-14T08:34:08.154806 - �[32m[INFO]�[0m    0.0 seconds: C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\custom_nodes\cable-management
2026-08-14T08:34:08.154842 - �[32m[INFO]�[0m 
2026-08-14T08:34:08.165597 - �[32m[INFO]�[0m Context impl SQLiteImpl.
2026-08-14T08:34:08.165667 - �[32m[INFO]�[0m Will assume non-transactional DDL.
2026-08-14T08:34:08.250071 - �[32m[INFO]�[0m Using **** pressure cache.
2026-08-14T08:34:08.250635 - �[32m[INFO]�[0m Starting server

2026-08-14T08:34:08.251040 - �[32m[INFO]�[0m To see the GUI go to: http://127.0.0.1:8188
2026-08-14T08:34:10.370889 - �[32m[INFO]�[0m [ComfyUI-Manager] The ComfyRegistry cache update is still in progress, so an outdated cache is being used.
2026-08-14T08:34:10.464158 - FETCH DATA from: C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Lib\site-packages\comfyui_manager\custom-node-list.json2026-08-14T08:34:10.464189 - 2026-08-14T08:34:10.495406 -  [DONE]2026-08-14T08:34:10.495452 - 
2026-08-14T08:38:29.984269 - �[1m�[33m[WARNING]�[0m Cannot connect to comfyregistry.
2026-08-14T08:38:29.987378 - FETCH DATA from: C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\user\__manager\cache\1514988643_custom-node-list.json2026-08-14T08:38:29.987412 - 2026-08-14T08:38:30.018992 -  [DONE]2026-08-14T08:38:30.019038 - 
2026-08-14T08:38:30.058267 - �[32m[INFO]�[0m [ComfyUI-Manager] All startup tasks have been completed.
2026-08-14T08:39:36.954598 - �[32m[INFO]�[0m got prompt
2026-08-14T08:39:37.158731 - �[32m[INFO]�[0m VAE load device: cuda:0, offload device: cpu, dtype: torch.float32
2026-08-14T08:39:37.220730 - �[32m[INFO]�[0m Found quantization metadata version 1
2026-08-14T08:39:38.460824 - �[32m[INFO]�[0m Requested to load MiniMaxMusic3TEModel
2026-08-14T08:39:38.488103 - �[32m[INFO]�[0m loaded completely;  8759.61 MB loaded, full load: True
2026-08-14T08:39:38.497855 - �[32m[INFO]�[0m CLIP/text encoder model load device: cpu, offload device: cpu, current: cpu, dtype: torch.float16
2026-08-14T08:39:38.516567 - �[1m�[31m[ERROR]�[0m !!! Exception during processing !!! Expected a cuda device, but got: cpu
2026-08-14T08:39:38.582435 - �[1m�[31m[ERROR]�[0m Traceback (most recent call last):
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\execution.py", line 545, in execute
    output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
                                                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\execution.py", line 344, in get_output_data
    return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\execution.py", line 318, in _async_map_node_over_list
    await process_inputs(input_dict, i)
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\execution.py", line 306, in process_inputs
    result = f(**inputs)
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy_api\internal\__init__.py", line 149, in wrapped_func
    return method(locked_class, **inputs)
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy_api\latest\_io.py", line 1990, in EXECUTE_NORMALIZED
    to_return = cls.execute(*args, **kwargs)
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy_extras\nodes_minimax_music.py", line 37, in execute
    conditioning = clip.encode_from_tokens_scheduled(tokens)
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy\sd.py", line 340, in encode_from_tokens_scheduled
    pooled_dict = self.encode_from_tokens(tokens, return_pooled=return_pooled, return_dict=True)
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy\sd.py", line 409, in encode_from_tokens
    o = self.cond_stage_model.encode_token_weights(tokens)
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy\text_encoders\minimax_music.py", line 98, in encode_token_weights
    hidden = self.generate(input_ids, seed, max_audio_f****es, self.execution_device, cfg_scale, top_k)
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy\ldm\minimax_music\ar.py", line 252, in generate
    past = self.model.init_kv_cache(2, prompt_tokens + decode_limit + 1, device, execution_dtype)
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\comfy\text_encoders\llama.py", line 794, in init_kv_cache
    fixed_kv = self.fixed_kv and comfy_kitchen.flash_attention_decode_is_available(device)
                                 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Lib\site-packages\comfy_kitchen\flash_attention.py", line 22, in is_available
    return torch.cuda.get_device_capability(device) >= _MINIMUM_CAPABILITY
           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Lib\site-packages\torch\cuda\__init__.py", line 615, in get_device_capability
    prop = get_device_properties(device)
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Lib\site-packages\torch\cuda\__init__.py", line 633, in get_device_properties
    device = _get_device_index(device, optional=True)
  File "C:\Users\****y\ComfyUI-Installs\Local\ComfyUI\.venv\Lib\site-packages\torch\cuda\_utils.py", line 525, in _get_device_index
    raise ValueError(f"Expected a cuda device, but got: {device}")
ValueError: Expected a cuda device, but got: cpu

2026-08-14T08:39:38.585236 - �[32m[INFO]�[0m �[32mPrompt executed in 1.63 seconds�[0m


## Attached Workflow
Please make sure that workflow does not contain any sensitive information such as API keys or passwords.

Workflow too large. Please manually upload the workflow from local file system.


## Additional Context
(Please add any additional context or steps to reproduce the error here)

Other

No response

[GlamoramaAttack]

I have still --cuda-device 0 in my start commands. Even though it's displayed as cuda:0, offload device: cpu in your log, try it with the start command.

[LocalCount]

I have still --cuda-device 0 in my start commands. Even though it's displayed as cuda:0, offload device: cpu in your log, try it with the start command.

Tried it just now and getting the same error

[GlamoramaAttack]

Then my next guess:

Set v**** state to: NORMAL_V****
2026-08-14T08:33:46.161428 - �[32m[INFO]�[0m Device: cuda:0 NVIDIA GeForce GTX 1660 SUPER : cudaMallocAsync

If I'm correct GTX 1660 has 6 GB VRAM. 6 GB isn't enough to keep the MiniMax Music text encoder (~8.7 GB just for that component, per the log: 8759.61 MB loaded) resident in VRAM. (My GPU is 8 GB and the 60 seconds audio clip took almost 16 minutes to generate. Not great.)

Keep the --cuda-device 0 start command, adding a launch flag like --lowvram , --novram or --cpu won't fix it itself, but it's worth testing — some ComfyUI memory paths behave differently when the mode is forced versus auto-detected, and it's possible the crash only happens in the "auto-decided to offload" path, not the "user explicitly requested CPU" path.

This part: missing device.type == "cuda" check before calling torch.cuda.get_device_capability is problematic for a GTX 1660, Turing-generation, it doesn't support flash attention efficiently anyway, so even if the crash were fixed, this node might just fall back to a slower attention path — but at least it would run instead of erroring out (as in my slow-motion experience).

This is very likely to affect anyone running MiniMax Music on a GPU with too little VRAM to hold the text encoder, so it's probably not unique to this setup. Bottom line: this model's text-encoder is fairly heavy for a 6 GB card, and the crash is happening because ComfyUI's CPU-fallback path for it isn't implemented for that case. But I bet, there are already people working on smaller nvfp4- or w4a8- versions.