About image size #5
Description
If I change image size from 1024 to 2048, it has error:
Traceback (most recent call last):
File "/root/autodl-tmp/code/SegEarth-R1/segearth_r1/train/train.py", line 474, in
train()
File "/root/autodl-tmp/code/SegEarth-R1/segearth_r1/train/train.py", line 451, in train
trainer.train()
File "/root/miniconda3/lib/python3.10/site-packages/transformers/trainer.py", line 1537, in train
return inner_training_loop(
File "/root/miniconda3/lib/python3.10/site-packages/transformers/trainer.py", line 1854, in _inner_training_loop
tr_loss_step = self.training_step(model, inputs)
File "/root/miniconda3/lib/python3.10/site-packages/transformers/trainer.py", line 2735, in training_step
loss = self.compute_loss(model, inputs)
File "/root/autodl-tmp/code/SegEarth-R1/segearth_r1/train/llava_trainer.py", line 280, in compute_loss
outputs = model(**inputs)
File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/root/miniconda3/lib/python3.10/site-packages/deepspeed/utils/nvtx.py", line 15, in wrapped_fn
ret_val = func(*args, **kwargs)
File "/root/miniconda3/lib/python3.10/site-packages/deepspeed/runtime/engine.py", line 1735, in forward
loss = self.module(*inputs, **kwargs)
File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/root/autodl-tmp/code/SegEarth-R1/segearth_r1/model/language_model/llava_phi.py", line 710, in forward
input_ids, attention_mask, past_key_values, inputs_embeds, labels, seg_query_mask, refer_embedding_indices = self.prepare_inputs_labels_for_multimodal(
File "/root/autodl-tmp/code/SegEarth-R1/segearth_r1/model/language_model/llava_phi.py", line 507, in prepare_inputs_labels_for_multimodal
image_features = self.encode_images(images) # image_features: [batch_size, 256, 2048]
File "/root/autodl-tmp/code/SegEarth-R1/segearth_r1/model/language_model/llava_phi.py", line 257, in encode_images
image_features = self.get_model().mm_projector(image_features[-1])
File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/root/autodl-tmp/code/SegEarth-R1/segearth_r1/model/multimodal_projector/builder.py", line 392, in forward
x = self.norm1(x)
File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(args, **kwargs)
File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/normalization.py", line 196, in forward
return F.layer_norm(
File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/functional.py", line 2543, in layer_norm
return torch.layer_norm(input, normalized_shape, weight, bias, eps, torch.backends.cudnn.enabled)
RuntimeError: Given normalized_shape=[2048, 16, 16], expected input with shape [, 2048, 16, 16], but got input of size[6, 2048, 32, 32]
Please give me some idea about this. It is strange!