search

NVlabs / EAGLE

EAGLE: Exploring The Design Space for Multimodal LLMs with Mixture of Encoders
https://arxiv.org/pdf/2408.15998
Apache License 2.0
525 stars 44 forks source link

NotImplementedError: Cannot copy out of meta tensor; no data! #20

Open clock-workorange opened 4 weeks ago

clock-workorange commented 4 weeks ago

Hi, I am getting this issue. I am running it on following system. I followed the instructions given in README.

Windows 11 Home Intel Core i9 32GB RAM ( I tried with Anaconda and python3.11.9 and pip both failed )

pip show accelerate:

Name: accelerate
Version: 0.34.2

Pytorch GPU CUDA version:

Is CUDA supported by this system True
CUDA version: 12.1
ID of current CUDA device: 0
Name of current CUDA device: NVIDIA GeForce RTX 4090 Laptop GPU

for pip version

pip install torch==2.3.1 torchvision==0.18.1 torchaudio==2.3.1 --index-url https://download.pytorch.org/whl/cu121

for conda version

conda install pytorch==2.3.1 torchvision==0.18.1 torchaudio==2.3.1 pytorch-cuda=12.1 -c pytorch -c nvidia

any idea how to solve this issue:

Some parameters are on the meta device because they were offloaded to the cpu.
C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\transformers\models\clip\modeling_clip.py:480: UserWarning: 1Torch was not compiled with flash attention. (Triggered internally at ..\aten\src\ATen\native\transformers\cuda\sdp_utils.cpp:455.)
  attn_output = torch.nn.functional.scaled_dot_product_attention(
Traceback (most recent call last):
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\predict_demo.py", line 66, in <module>
    output_ids = model.generate(
                 ^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\utils\_contextlib.py", line 115, in decorate_context
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\language_model\eagle_llama.py", line 140, in generate
    ) = self.prepare_inputs_labels_for_multimodal(
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\eagle_arch.py", line 221, in prepare_inputs_labels_for_multimodal
    image_features = self.encode_images(images)
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\eagle_arch.py", line 160, in encode_images
    image_features = self.get_model().get_vision_tower()(images)
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1532, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1541, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\hooks.py", line 170, in new_forward
    output = module._old_forward(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\multimodal_encoder\multi_backbone_channel_concatenation_encoder.py", line 116, in forward     
    feature = vision_tower(resized_x)
              ^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1532, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1541, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\hooks.py", line 170, in new_forward
    output = module._old_forward(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\multimodal_encoder\convnext_encoder.py", line 99, in forward
    image_features = self._forward_images(images)
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\multimodal_encoder\convnext_encoder.py", line 105, in _forward_images
    image_forward_outs = self.forward_features(images.to(device=self.device, dtype=self.dtype))
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\multimodal_encoder\convnext_encoder.py", line 86, in forward_features
    x = self.vision_tower.stem(x)
        ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1532, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1541, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\hooks.py", line 165, in new_forward
    args, kwargs = module._hf_hook.pre_forward(module, *args, **kwargs)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\hooks.py", line 364, in pre_forward
    return send_to_device(args, self.execution_device), send_to_device(
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\utils\operations.py", line 175, in send_to_device
    return honor_type(
           ^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\utils\operations.py", line 82, in honor_type
    return type(obj)(generator)
           ^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\utils\operations.py", line 176, in <genexpr>
    tensor, (send_to_device(t, device, non_blocking=non_blocking, skip_keys=skip_keys) for t in tensor)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\utils\operations.py", line 156, in send_to_device
    return tensor.to(device, non_blocking=non_blocking)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
NotImplementedError: Cannot copy out of meta tensor; no data!
98986oiuoy commented 2 days ago

same error

98986oiuoy commented 2 days ago

Hi, I am getting this issue. I am running it on following system. I followed the instructions given in README.

Windows 11 Home Intel Core i9 32GB RAM ( I tried with Anaconda and python3.11.9 and pip both failed )

pip show accelerate:

Name: accelerate
Version: 0.34.2

Pytorch GPU CUDA version:

Is CUDA supported by this system True
CUDA version: 12.1
ID of current CUDA device: 0
Name of current CUDA device: NVIDIA GeForce RTX 4090 Laptop GPU

for pip version

pip install torch==2.3.1 torchvision==0.18.1 torchaudio==2.3.1 --index-url https://download.pytorch.org/whl/cu121

for conda version

conda install pytorch==2.3.1 torchvision==0.18.1 torchaudio==2.3.1 pytorch-cuda=12.1 -c pytorch -c nvidia

any idea how to solve this issue:

Some parameters are on the meta device because they were offloaded to the cpu.
C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\transformers\models\clip\modeling_clip.py:480: UserWarning: 1Torch was not compiled with flash attention. (Triggered internally at ..\aten\src\ATen\native\transformers\cuda\sdp_utils.cpp:455.)
  attn_output = torch.nn.functional.scaled_dot_product_attention(
Traceback (most recent call last):
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\predict_demo.py", line 66, in <module>
    output_ids = model.generate(
                 ^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\utils\_contextlib.py", line 115, in decorate_context
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\language_model\eagle_llama.py", line 140, in generate
    ) = self.prepare_inputs_labels_for_multimodal(
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\eagle_arch.py", line 221, in prepare_inputs_labels_for_multimodal
    image_features = self.encode_images(images)
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\eagle_arch.py", line 160, in encode_images
    image_features = self.get_model().get_vision_tower()(images)
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1532, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1541, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\hooks.py", line 170, in new_forward
    output = module._old_forward(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\multimodal_encoder\multi_backbone_channel_concatenation_encoder.py", line 116, in forward     
    feature = vision_tower(resized_x)
              ^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1532, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1541, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\hooks.py", line 170, in new_forward
    output = module._old_forward(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\multimodal_encoder\convnext_encoder.py", line 99, in forward
    image_features = self._forward_images(images)
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\multimodal_encoder\convnext_encoder.py", line 105, in _forward_images
    image_forward_outs = self.forward_features(images.to(device=self.device, dtype=self.dtype))
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\Desktop\Automatizacion_Township\EAGLE\eagle\model\multimodal_encoder\convnext_encoder.py", line 86, in forward_features
    x = self.vision_tower.stem(x)
        ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1532, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\torch\nn\modules\module.py", line 1541, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\hooks.py", line 165, in new_forward
    args, kwargs = module._hf_hook.pre_forward(module, *args, **kwargs)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\hooks.py", line 364, in pre_forward
    return send_to_device(args, self.execution_device), send_to_device(
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\utils\operations.py", line 175, in send_to_device
    return honor_type(
           ^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\utils\operations.py", line 82, in honor_type
    return type(obj)(generator)
           ^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\utils\operations.py", line 176, in <genexpr>
    tensor, (send_to_device(t, device, non_blocking=non_blocking, skip_keys=skip_keys) for t in tensor)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\user\AppData\Local\Programs\Python\Python311\Lib\site-packages\accelerate\utils\operations.py", line 156, in send_to_device
    return tensor.to(device, non_blocking=non_blocking)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
NotImplementedError: Cannot copy out of meta tensor; no data!

hi bro have you sloved the problem?