Skip to content

Exporting Alpamayo-R1-10B fails with AssertionError #166

Description

@FABallemand

Describe the bug

Exporting the officially supported Alpamayo-R1-10B checkpoint consistently fails with the following assertion during ONNX export:

Traceback (most recent call last):
  File "/workspace/TensorRT-Edge-LLM/host_venv/bin/tensorrt-edgellm-export", line 8, in <module>
    sys.exit(main())
             ^^^^^^
  File "/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/tensorrt_edgellm/scripts/export.py", line 2812, in main
    fn(
  File "/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/tensorrt_edgellm/scripts/export.py", line 2762, in <lambda>
    and _allow("action"), "action", lambda out: _export_action(
                                                ^^^^^^^^^^^^^^^
  File "/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/tensorrt_edgellm/scripts/export.py", line 2285, in _export_action
    export_action_onnx(
  File "/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/tensorrt_edgellm/onnx/export_encoder.py", line 470, in export_action_onnx
    _run_dynamo_export(model, dynamo_inputs, output_path, onnx_input_names,
  File "/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/tensorrt_edgellm/onnx/export_encoder.py", line 220, in _run_dynamo_export
    assert len(dynamo_inputs) == len(onnx_input_names) and \
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: dynamo_inputs: 78, onnx_input_names: 78, dynamic_shapes: 7

Since Alpamayo-R1-10B is listed as a supported checkpoint, I expected the export to complete successfully.

During export, the following warning is also emitted. I'm not sure whether it's related to the assertion failure:

XX:XX:XX  WARNING   tensorrt_edgellm.checkpoint.checkpoint_utils: AutoConfig.from_pretrained failed for Alpamayo-R1-10B (The checkpoint you are trying to load has model type `alpamayo_r1` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.

TensorRT Edge-LLM 0.9.1 uses transformers==5.9.0, which is newer than the minimum version (4.57.1) required by the Alpamayo-R1-10B checkpoint. Because of that, I'm not sure why AutoConfig.from_pretrained reports that the transformers version may be too old.

Using transformers 5.12.1 (which also satisfies the safetensors==0.7.0 TensorRT Edge-LLM requirement) does not solve the issue.

I also noticed this warning during export:

XX:XX:XX  WARNING   tensorrt_edgellm.scripts.export: [Visual] Neither preprocessor_config.json nor processor_config.json found at Alpamayo-R1-10B

Steps/Code to reproduce bug

Installation method:
As described in TensorRT Edge-LLM on Jetson, I used an NVIDIA Jetson Thor as both the host and target device. I installed TensorRT Edge-LLM v0.9.1 by following the Installation Guide. The installation completed successfully without any issues.

$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2025 NVIDIA Corporation
Built on Wed_Jul_16_07:31:19_PM_PDT_2025
Cuda compilation tools, release 13.0, V13.0.48
Build cuda_13.0.r13.0/compiler.36260728_0

$ nvidia-smi
Mon Jul 27 08:52:49 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.00                 Driver Version: 580.00         CUDA Version: 13.0     |
+-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA Thor                    Off |   00000000:01:00.0 Off |                  N/A |
| N/A   36C  N/A               3W /  N/A  | Not Supported          |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    0   N/A  N/A           15507      G   /usr/lib/xorg/Xorg                       92MiB |
|    0   N/A  N/A           15707      G   /usr/bin/gnome-shell                     55MiB |
+-----------------------------------------------------------------------------------------+

$ python3 --version
Python 3.12.3

$ cat /etc/nv_tegra_release
# R38 (release), REVISION: 4.0, GCID: 43443517, BOARD: generic, EABI: aarch64, DATE: Wed Dec 31 00:15:19 UTC 2025
# KERNEL_VARIANT: oot
TARGET_USERSPACE_LIB_DIR=nvidia
TARGET_USERSPACE_LIB_DIR_PATH=usr/lib/aarch64-linux-gnu/nvidia
INSTALL_TYPE=

Export command used:
I use the following command to export models:

$ hf download nvidia/Alpamayo-R1-10B --local-dir Alpamayo-R1-10B
$ tensorrt-edgellm-export \
    Alpamayo-R1-10B \
    Alpamayo-R1-10B/onnx \
    --max-kv-cache-capacity 4096
Details
13:04:45  INFO      tensorrt_edgellm.scripts.export: ============================================================
13:04:45  INFO      tensorrt_edgellm.scripts.export: Model type    : alpamayo_r1
13:04:45  INFO      tensorrt_edgellm.scripts.export: Checkpoint    : Alpamayo-R1-10B
13:04:45  INFO      tensorrt_edgellm.scripts.export: Output dir    : Alpamayo-R1-10B/onnx
13:04:45  INFO      tensorrt_edgellm.scripts.export:   thinker        : yes
13:04:45  INFO      tensorrt_edgellm.scripts.export:   mtp_draft      : no
13:04:45  INFO      tensorrt_edgellm.scripts.export:   mtp_draft      : no
13:04:45  INFO      tensorrt_edgellm.scripts.export:   dflash_draft   : no
13:04:45  INFO      tensorrt_edgellm.scripts.export:   talker         : no
13:04:45  INFO      tensorrt_edgellm.scripts.export:   code_predictor : no
13:04:45  INFO      tensorrt_edgellm.scripts.export:   visual         : yes
13:04:45  INFO      tensorrt_edgellm.scripts.export:   audio          : no
13:04:45  INFO      tensorrt_edgellm.scripts.export:   code2wav       : no
13:04:45  INFO      tensorrt_edgellm.scripts.export:   action         : yes
13:04:45  INFO      tensorrt_edgellm.scripts.export: FP8 embedding : no
13:04:45  INFO      tensorrt_edgellm.scripts.export: MTP capable   : no
13:04:45  INFO      tensorrt_edgellm.scripts.export: MTP export    : no
13:04:45  INFO      tensorrt_edgellm.scripts.export: Gemma4 MTP    : no
13:04:45  INFO      tensorrt_edgellm.scripts.export: DFlash base   : no
13:04:45  INFO      tensorrt_edgellm.scripts.export: DFlash draft  : no
13:04:45  INFO      tensorrt_edgellm.scripts.export: Reduced vocab : no
13:04:45  INFO      tensorrt_edgellm.scripts.export: External weights: no
13:04:45  INFO      tensorrt_edgellm.scripts.export: TP size       : 1
13:04:45  INFO      tensorrt_edgellm.scripts.export: ============================================================
13:04:45  INFO      tensorrt_edgellm.scripts.export: [LLM] Loading checkpoint from Alpamayo-R1-10B
13:04:45  WARNING   tensorrt_edgellm.checkpoint.checkpoint_utils: AutoConfig.from_pretrained failed for Alpamayo-R1-10B (The checkpoint you are trying to load has model type `alpamayo_r1` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.

You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git`); falling back to raw config.json.
13:04:45  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:04:45  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:04:45  WARNING   tensorrt_edgellm.checkpoint.checkpoint_utils: AutoConfig.from_pretrained failed for Alpamayo-R1-10B (The checkpoint you are trying to load has model type `alpamayo_r1` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.

You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git`); falling back to raw config.json.
13:04:45  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:04:45  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:04:51  INFO      tensorrt_edgellm.checkpoint.loader: Stripping key prefix 'vlm.model.language_model.' from checkpoint keys (inserting 'model.')
13:04:56  INFO      tensorrt_edgellm.checkpoint.loader: Loaded 399 tensors, skipped 767 from Alpamayo-R1-10B
13:04:56  INFO      tensorrt_edgellm.checkpoint.repacking: Repacked GPTQ weights
13:04:56  INFO      tensorrt_edgellm.scripts.export: [LLM] Exporting to Alpamayo-R1-10B/onnx/llm/model.onnx
13:04:56  INFO      tensorrt_edgellm.onnx.export: Exporting ONNX to Alpamayo-R1-10B/onnx/llm/model.onnx (opset 24, dynamo) ...
/usr/lib/python3.12/contextlib.py:144: UserWarning: The tensor attribute self._model.model.last_pre_norm_hidden_states was assigned during export. Such attributes must be registered as buffers using the `register_buffer` API (https://pytorch.org/docs/stable/generated/torch.nn.Module.html#torch.nn.Module.register_buffer).
  next(self.gen)
/usr/lib/python3.12/copyreg.py:99: FutureWarning: `isinstance(treespec, LeafSpec)` is deprecated, use `isinstance(treespec, TreeSpec) and treespec.is_leaf()` instead.
  return cls.__new__(cls, *args)
13:05:13  INFO      onnx_ir.passes.common.unused_removal: No unused functions to remove
13:05:14  INFO      onnx_ir.passes.common.unused_removal: Removed 27 unused nodes
13:05:14  INFO      onnx_ir.passes.common.unused_removal: No unused functions to remove
13:05:14  INFO      onnxscript.optimizer._constant_folding: Skipping constant folding for node 'node_Shape_1' because it is graph input to preserve graph signature
[...]
13:05:14  INFO      onnxscript.optimizer._constant_folding: Skipping constant folding for node 'n1' because it is graph input to preserve graph signature
13:05:18  INFO      onnx_ir.passes.common.unused_removal: Removed 688 unused nodes
13:05:18  INFO      onnxscript.rewriter: Applied 360 of general pattern rewrite rules.
13:05:18  INFO      onnx_ir.passes.common.unused_removal: No unused functions to remove
13:05:18  INFO      onnxscript.optimizer._constant_folding: Skipping constant folding for node 'node_Shape_1' because it is graph input to preserve graph signature
[...]
13:05:18  INFO      onnxscript.optimizer._constant_folding: Skipping constant folding for node 'n1' because it is graph input to preserve graph signature
13:05:20  INFO      onnx_ir.passes.common.unused_removal: Removed 360 unused nodes
13:05:20  INFO      onnx_ir.passes.common.unused_removal: No unused functions to remove
13:05:20  INFO      onnx_ir.passes.common.initializer_deduplication: Replaced initializer 'val_26' with existing initializer 'val_12'
[...]
13:05:20  INFO      onnx_ir.passes.common.initializer_deduplication: Replaced initializer 'axes' with existing initializer 'val_12'
/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/torch/onnx/_internal/exporter/_onnx_program.py:486: UserWarning: # The axis name: batch will not be used, since it shares the same shape constraints with another axis: batch.
  rename_mapping = _dynamic_shapes.create_rename_mapping(
/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/torch/onnx/_internal/exporter/_onnx_program.py:486: UserWarning: # The axis name: past_len will not be used, since it shares the same shape constraints with another axis: past_len.
  rename_mapping = _dynamic_shapes.create_rename_mapping(
/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/torch/onnx/_internal/exporter/_onnx_program.py:486: UserWarning: # The axis name: seq_len will not be used, since it shares the same shape constraints with another axis: seq_len.
  rename_mapping = _dynamic_shapes.create_rename_mapping(
13:08:22  INFO      tensorrt_edgellm.onnx.export: Export complete: Alpamayo-R1-10B/onnx/llm/model.onnx
13:08:22  INFO      tensorrt_edgellm.checkpoint.checkpoint_utils: Wrote config.json to /workspace/tensorrt-edgellm-workspace/VLA/Alpamayo-R1-10B/onnx/llm
13:08:23  INFO      tensorrt_edgellm.checkpoint.checkpoint_utils: Wrote embedding.safetensors ([155697, 4096])
13:08:25  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:25  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:25  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:25  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:25  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:25  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:25  INFO      httpx: HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-VL-8B-Instruct/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found"
13:08:26  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:26  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/chat_template.json "HTTP/1.1 307 Temporary Redirect"
13:08:26  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/chat_template.json "HTTP/1.1 200 OK"
13:08:26  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/chat_template.jinja "HTTP/1.1 404 Not Found"
13:08:26  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/audio_tokenizer_config.json "HTTP/1.1 404 Not Found"
13:08:26  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:26  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:26  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:26  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:26  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:26  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:27  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:08:27  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:08:27  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/tokenizer_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:27  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/tokenizer_config.json "HTTP/1.1 200 OK"
13:08:27  INFO      httpx: HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-VL-8B-Instruct/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found"
13:08:27  INFO      httpx: HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-VL-8B-Instruct/tree/main?recursive=true&expand=false "HTTP/1.1 200 OK"
13:08:28  INFO      httpx: HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-VL-8B-Instruct "HTTP/1.1 200 OK"
13:08:28  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:28  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/video_preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:28  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/video_preprocessor_config.json "HTTP/1.1 200 OK"
13:08:28  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:28  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:28  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:28  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/video_preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:28  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/video_preprocessor_config.json "HTTP/1.1 200 OK"
13:08:28  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:28  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:29  INFO      tensorrt_edgellm.checkpoint.checkpoint_utils: Saved Alpamayo tokenizer (155697 tokens) to /workspace/tensorrt-edgellm-workspace/VLA/Alpamayo-R1-10B/onnx/llm
13:08:29  WARNING   tensorrt_edgellm.checkpoint.checkpoint_utils: AutoConfig.from_pretrained failed for Alpamayo-R1-10B (The checkpoint you are trying to load has model type `alpamayo_r1` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.

You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git`); falling back to raw config.json.
13:08:29  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:08:29  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:08:29  WARNING   tensorrt_edgellm.checkpoint.checkpoint_utils: AutoConfig.from_pretrained failed for Alpamayo-R1-10B (The checkpoint you are trying to load has model type `alpamayo_r1` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.

You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git`); falling back to raw config.json.
13:08:29  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:08:29  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
[transformers] You are using a model of type `alpamayo_r1` to instantiate a model of type ``. This may be expected if you are loading a checkpoint that shares a subset of the architecture (e.g., loading a `sam2_video` checkpoint into `Sam2Model`), but is otherwise not supported and can yield errors. Please verify that the checkpoint is compatible with the model you are instantiating.
[transformers] You are using a model of type `alpamayo_r1` to instantiate a model of type ``. This may be expected if you are loading a checkpoint that shares a subset of the architecture (e.g., loading a `sam2_video` checkpoint into `Sam2Model`), but is otherwise not supported and can yield errors. Please verify that the checkpoint is compatible with the model you are instantiating.
[transformers] PreTrainedConfig got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 
[transformers] Unrecognized keys in `rope_parameters` for 'rope_type'='default': {'mrope_interleaved', 'mrope_section'}
[transformers] The tokenizer you are loading from '/workspace/tensorrt-edgellm-workspace/VLA/Alpamayo-R1-10B/onnx/llm' with an incorrect regex pattern: https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503/discussions/84#69121093e8b480e709447d5e. This will lead to incorrect tokenization. You should set the `fix_mistral_regex=True` flag when loading this tokenizer to fix this issue.
13:08:30  WARNING   tensorrt_edgellm.checkpoint.checkpoint_utils: AutoConfig.from_pretrained failed for Alpamayo-R1-10B (The checkpoint you are trying to load has model type `alpamayo_r1` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.

You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git`); falling back to raw config.json.
13:08:30  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:08:30  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:08:30  INFO      tensorrt_edgellm.chat_template: Detected Alpamayo 1 model, adding <|cot_start|> to generation prompt
13:08:30  WARNING   tensorrt_edgellm.checkpoint.checkpoint_utils: AutoConfig.from_pretrained failed for Alpamayo-R1-10B (The checkpoint you are trying to load has model type `alpamayo_r1` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.

You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git`); falling back to raw config.json.
13:08:30  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:08:30  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:08:30  WARNING   tensorrt_edgellm.checkpoint.checkpoint_utils: AutoConfig.from_pretrained failed for Alpamayo-R1-10B (The checkpoint you are trying to load has model type `alpamayo_r1` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.

You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git`); falling back to raw config.json.
13:08:30  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:08:30  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:08:30  WARNING   tensorrt_edgellm.checkpoint.checkpoint_utils: AutoConfig.from_pretrained failed for Alpamayo-R1-10B (The checkpoint you are trying to load has model type `alpamayo_r1` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.

You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git`); falling back to raw config.json.
13:08:31  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:08:31  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:08:31  INFO      tensorrt_edgellm.chat_template: Chat template saved to /workspace/tensorrt-edgellm-workspace/VLA/Alpamayo-R1-10B/onnx/llm/processed_chat_template.json
13:08:31  WARNING   tensorrt_edgellm.scripts.export: Could not find token ID for '<|image_pad|>' in Alpamayo-R1-10B
13:08:31  INFO      tensorrt_edgellm.scripts.export: [LLM] Patched multimodal token IDs: image_token_id=151655
13:08:31  INFO      tensorrt_edgellm.scripts.export: [LLM] Done: Alpamayo-R1-10B/onnx/llm/model.onnx
13:08:31  INFO      tensorrt_edgellm.scripts.export: Loading safetensors weights ...
13:08:31  INFO      tensorrt_edgellm.scripts.export:   Loading shard: model-00001-of-00005.safetensors
13:08:31  INFO      tensorrt_edgellm.scripts.export:   Loading shard: model-00002-of-00005.safetensors
13:08:31  INFO      tensorrt_edgellm.scripts.export:   Loading shard: model-00003-of-00005.safetensors
13:08:31  INFO      tensorrt_edgellm.scripts.export:   Loading shard: model-00004-of-00005.safetensors
13:08:31  INFO      tensorrt_edgellm.scripts.export:   Loading shard: model-00005-of-00005.safetensors
13:08:31  WARNING   tensorrt_edgellm.checkpoint.checkpoint_utils: AutoConfig.from_pretrained failed for Alpamayo-R1-10B (The checkpoint you are trying to load has model type `alpamayo_r1` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.

You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git`); falling back to raw config.json.
13:08:31  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:08:31  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:08:31  WARNING   tensorrt_edgellm.checkpoint.checkpoint_utils: AutoConfig.from_pretrained failed for Alpamayo-R1-10B (The checkpoint you are trying to load has model type `alpamayo_r1` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.

You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git`); falling back to raw config.json.
13:08:31  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:08:31  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:08:31  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:08:31  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:08:31  INFO      tensorrt_edgellm.scripts.export: [Visual] Exporting qwen3_vl visual encoder to Alpamayo-R1-10B/onnx/visual/model.onnx
13:08:31  INFO      tensorrt_edgellm.onnx.export_encoder: Building qwen3_vl visual model ...
13:08:31  INFO      tensorrt_edgellm.models.ops: Using attention plugin or non-fused attention
[...]
13:08:31  INFO      tensorrt_edgellm.models.ops: Using attention plugin or non-fused attention
13:08:32  INFO      tensorrt_edgellm.checkpoint.repacking: Repacked GPTQ weights
13:08:32  INFO      tensorrt_edgellm.onnx.export_encoder: Exporting ONNX to Alpamayo-R1-10B/onnx/visual/model.onnx (opset 24) ...
/usr/lib/python3.12/copyreg.py:99: FutureWarning: `isinstance(treespec, LeafSpec)` is deprecated, use `isinstance(treespec, TreeSpec) and treespec.is_leaf()` instead.
  return cls.__new__(cls, *args)
13:08:43  INFO      onnx_ir.passes.common.unused_removal: No unused functions to remove
13:08:43  INFO      onnx_ir.passes.common.unused_removal: Removed 31 unused nodes
13:08:43  INFO      onnx_ir.passes.common.unused_removal: No unused functions to remove
13:08:43  INFO      onnxscript.optimizer._constant_folding: Skipping constant folding for node 'node_Shape_2' because it is graph input to preserve graph signature
[...]
13:08:44  INFO      onnxscript.optimizer._constant_folding: Skipping constant folding for node 'n0_108' because it is graph input to preserve graph signature
13:08:44  INFO      onnxscript.rewriter.rules.common._collapse_slices: The value 'end' is less than the shape of the specified axis.
13:08:44  INFO      onnxscript.rewriter.rules.common._collapse_slices: The value 'start' is not 0.
[...]
13:08:46  INFO      onnxscript.rewriter.rules.common._collapse_slices: The value 'end' is less than the shape of the specified axis.
13:08:46  INFO      onnxscript.rewriter.rules.common._collapse_slices: The value 'start' is not 0.
13:08:46  INFO      onnx_ir.passes.common.unused_removal: Removed 689 unused nodes
13:08:46  INFO      onnxscript.rewriter: Applied 55 of general pattern rewrite rules.
13:08:46  INFO      onnx_ir.passes.common.unused_removal: No unused functions to remove
13:08:46  INFO      onnxscript.optimizer._constant_folding: Skipping constant folding for node 'node_Shape_2' because it is graph input to preserve graph signature
13:08:46  INFO      onnxscript.optimizer._constant_folding: Skipping constant folding for node 'node_view' because it is graph input to preserve graph signature
[...]
13:08:46  INFO      onnxscript.optimizer._constant_folding: Skipping constant folding for node 'n0_4' because it is graph input to preserve graph signature
[...]
13:08:46  INFO      onnxscript.optimizer._constant_folding: Skipping constant folding for node 'n0_108' because it is graph input to preserve graph signature
13:08:47  INFO      onnxscript.rewriter.rules.common._collapse_slices: The value 'end' is less than the shape of the specified axis.
13:08:47  INFO      onnxscript.rewriter.rules.common._collapse_slices: The value 'start' is not 0.
[...]
13:08:48  INFO      onnxscript.rewriter.rules.common._collapse_slices: The value 'end' is less than the shape of the specified axis.
13:08:48  INFO      onnxscript.rewriter.rules.common._collapse_slices: The value 'start' is not 0.
13:08:48  INFO      onnx_ir.passes.common.unused_removal: Removed 55 unused nodes
13:08:48  INFO      onnx_ir.passes.common.unused_removal: No unused functions to remove
13:08:48  INFO      onnx_ir.passes.common.initializer_deduplication: Replaced initializer 'val_39' with existing initializer 'val_36'
[...]
13:08:48  INFO      onnx_ir.passes.common.initializer_deduplication: Replaced initializer 'val_1730' with existing initializer 'val_37'
/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/torch/onnx/_internal/exporter/_onnx_program.py:486: UserWarning: # The axis name: total_tokens will not be used, since it shares the same shape constraints with another axis: total_tokens.
  rename_mapping = _dynamic_shapes.create_rename_mapping(
13:08:50  INFO      tensorrt_edgellm.onnx.export_encoder: Export complete: Alpamayo-R1-10B/onnx/visual/model.onnx
13:08:50  INFO      tensorrt_edgellm.scripts.export: [Visual] Done: Alpamayo-R1-10B/onnx/visual/model.onnx
13:08:50  WARNING   tensorrt_edgellm.scripts.export: [Visual] Neither preprocessor_config.json nor processor_config.json found at Alpamayo-R1-10B
13:08:50  INFO      tensorrt_edgellm.scripts.export: [Visual] Wrote config.json: Alpamayo-R1-10B/onnx/visual/config.json
13:08:50  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:50  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:50  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:50  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:51  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:51  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:51  INFO      httpx: HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-VL-8B-Instruct/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found"
13:08:51  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:51  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/chat_template.json "HTTP/1.1 307 Temporary Redirect"
13:08:51  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/chat_template.json "HTTP/1.1 200 OK"
13:08:51  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/chat_template.jinja "HTTP/1.1 404 Not Found"
13:08:51  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/audio_tokenizer_config.json "HTTP/1.1 404 Not Found"
13:08:51  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:51  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:51  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:51  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:52  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:52  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:52  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/config.json "HTTP/1.1 307 Temporary Redirect"
13:08:52  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/config.json "HTTP/1.1 200 OK"
13:08:52  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/tokenizer_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:52  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/tokenizer_config.json "HTTP/1.1 200 OK"
13:08:52  INFO      httpx: HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-VL-8B-Instruct/tree/main/additional_chat_templates?recursive=false&expand=false "HTTP/1.1 404 Not Found"
13:08:52  INFO      httpx: HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-VL-8B-Instruct/tree/main?recursive=true&expand=false "HTTP/1.1 200 OK"
13:08:53  INFO      httpx: HTTP Request: GET https://huggingface.co/api/models/Qwen/Qwen3-VL-8B-Instruct "HTTP/1.1 200 OK"
13:08:53  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:53  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/video_preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:53  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/video_preprocessor_config.json "HTTP/1.1 200 OK"
13:08:53  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:53  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:53  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/processor_config.json "HTTP/1.1 404 Not Found"
13:08:53  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/video_preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:53  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/video_preprocessor_config.json "HTTP/1.1 200 OK"
13:08:53  INFO      httpx: HTTP Request: HEAD https://huggingface.co/Qwen/Qwen3-VL-8B-Instruct/resolve/main/preprocessor_config.json "HTTP/1.1 307 Temporary Redirect"
13:08:53  INFO      httpx: HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/Qwen/Qwen3-VL-8B-Instruct/0c351dd01ed87e9c1b53cbc748cba10e6187ff3b/preprocessor_config.json "HTTP/1.1 200 OK"
13:08:53  INFO      tensorrt_edgellm.scripts.export: [Visual] Saved Alpamayo processor sidecar files to Alpamayo-R1-10B/onnx/visual
13:08:54  INFO      tensorrt_edgellm.scripts.export: [Action] Building ActionConfig from checkpoint ...
13:08:54  INFO      tensorrt_edgellm.scripts.export: [Action] Expert: 36 layers, 16 heads, hidden=2048
13:08:54  INFO      tensorrt_edgellm.scripts.export: [Action] Exporting to Alpamayo-R1-10B/onnx/action/model.onnx
13:08:54  INFO      tensorrt_edgellm.onnx.export_encoder: Building Alpamayo action expert ...
13:09:04  INFO      tensorrt_edgellm.models.alpamayo.modeling_alpamayo_action: Loaded 412 action expert tensors
[torch.onnx] Obtain model graph for `_Wrapper([...]` with `torch.export.export(..., strict=False)`...
[torch.onnx] Obtain model graph for `_Wrapper([...]` with `torch.export.export(..., strict=False)`... ✅
[torch.onnx] Run decompositions...
[torch.onnx] Run decompositions... ✅
[torch.onnx] Translate the graph into ONNX...
[torch.onnx] Translate the graph into ONNX... ✅
[torch.onnx] Optimize the ONNX graph...
[torch.onnx] Optimize the ONNX graph... ✅
[torch.onnx] Obtain model graph for `Qwen3VLVisualModel([...]` with `torch.export.export(..., strict=False)`...
[torch.onnx] Obtain model graph for `Qwen3VLVisualModel([...]` with `torch.export.export(..., strict=False)`... ✅
[torch.onnx] Run decompositions...
[torch.onnx] Run decompositions... ✅
[torch.onnx] Translate the graph into ONNX...
[torch.onnx] Translate the graph into ONNX... ✅
[torch.onnx] Optimize the ONNX graph...
[torch.onnx] Optimize the ONNX graph... ✅
Traceback (most recent call last):
  File "/workspace/TensorRT-Edge-LLM/host_venv/bin/tensorrt-edgellm-export", line 8, in <module>
    sys.exit(main())
             ^^^^^^
  File "/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/tensorrt_edgellm/scripts/export.py", line 2812, in main
    fn(
  File "/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/tensorrt_edgellm/scripts/export.py", line 2762, in <lambda>
    and _allow("action"), "action", lambda out: _export_action(
                                                ^^^^^^^^^^^^^^^
  File "/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/tensorrt_edgellm/scripts/export.py", line 2285, in _export_action
    export_action_onnx(
  File "/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/tensorrt_edgellm/onnx/export_encoder.py", line 470, in export_action_onnx
    _run_dynamo_export(model, dynamo_inputs, output_path, onnx_input_names,
  File "/workspace/TensorRT-Edge-LLM/host_venv/lib/python3.12/site-packages/tensorrt_edgellm/onnx/export_encoder.py", line 220, in _run_dynamo_export
    assert len(dynamo_inputs) == len(onnx_input_names) and \
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: dynamo_inputs: 78, onnx_input_names: 78, dynamic_shapes: 7

System information (Nvidia Jetson Thor)

  • Container used: nvcr.io/nvidia/pytorch:25.12-py3
  • OS: Ubuntu 24.04.3 LTS
  • CPU architecture: aarch64
  • GPU name: Nvidia Jetson T5000
  • GPU memory size: 128 GB
  • Number of GPUs: 1
  • Library versions:
    • Python: 3.12.3
    • TensorRT Edge-LLM version: 0.9.1
    • CUDA: 13.1
    • PyTorch: 2.12.0+cu130
    • Transformers: 5.9.0
    • ModelOpt: 0.44.0
    • ONNX: 1.19.0

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions