Closed
Description
Describe the bug/ 问题描述 (Mandatory / 必填)
A clear and concise description of what the bug is.
Traceback (most recent call last):
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/threading.py", line 973, in _bootstrap_inner
self.run()
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/threading.py", line 910, in run
self._target(*self._args, **self._kwargs)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/core/utils/_contextlib.py", line 117, in decorate_context
return func(*args, **kwargs)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/transformers/generation/utils.py", line 2024, in generate
result = self._sample(
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/transformers/generation/utils.py", line 3029, in _sample
model_inputs = self.prepare_inputs_for_generation(input_ids, **model_kwargs)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/transformers/models/baichuan/modeling_baichuan.py", line 1858, in prepare_inputs_for_generation
position_ids = position_ids.masked_fill(attention_mask == 0, 1)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindspore/common/tensor.py", line 2682, in masked_fill
value = tensor_operator_registry.get("scalar_to_tensor")(value, self.dtype)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindspore/common/_stub_tensor.py", line 99, in dtype
self.stub_dtype = self.stub.get_dtype()
TypeError: For primitive[CumSum], the input argument[x] must be a type of {Tensor[Float16], Tensor[Float32], Tensor[Float64], Tensor[Int32], Tensor[Int8], Tensor[UInt8]}, but got Tensor[Int64].
----------------------------------------------------
- C++ Call Stack: (For framework developers)
----------------------------------------------------
mindspore/core/utils/check_convert_utils.cc:1031 CheckTensorSubClass
- Hardware Environment(
Ascend/GPU/CPU) / 硬件环境:
Please delete the backend not involved / 请删除不涉及的后端:
Ascend
-
Software Environment / 软件环境 (Mandatory / 必填):
-
mindnlp是今日最新源码
-- MindSpore version (e.g., 1.7.0.Bxxx) : 2.3.1
-- Python version (e.g., Python 3.7.5) :Python 3.9.10
-- OS platform and distribution (e.g., Linux Ubuntu 16.04): 4.19.90-vhulk2211.3.0.h1543.eulerosv2r10.aarch6
-- GCC/Compiler version (if compiled from source): 7.3.0 -
Excute Mode / 执行模式 (Mandatory / 必填)(
PyNative/Graph):
Please delete the mode not involved / 请删除不涉及的模式:
/mode pynative
To Reproduce / 重现步骤 (Mandatory / 必填)
代码:
import mindspore as ms
from mindnlp.transformers import BaiChuanForCausalLM, BaiChuanTokenizer
from mindnlp.transformers.generation.utils import GenerationConfig
# ms.set_context(pynative_synchronize=True)
model = BaiChuanForCausalLM.from_pretrained("baichuan-inc/Baichuan-13B-Chat",mirror='modelscope',ms_dtype=ms.float16, size='13b').eval()
tokenizer = BaiChuanTokenizer.from_pretrained("baichuan-inc/Baichuan-13B-Chat",mirror='modelscope')
model.generation_config = GenerationConfig.from_pretrained("baichuan-inc/Baichuan-13B-Chat")
messages = []
messages.append({"role": "user", "content": "北京有啥好玩的地方？"})
response = model.chat(tokenizer, messages)
print(response)
Expected behavior / 预期结果 (Mandatory / 必填)
能够推理。
Screenshots/ 日志 / 截图 (Mandatory / 必填)
Additional context / 备注 (Optional / 选填)
我尝试修改后：
即将
long修改成int然后就变了新的错误：
Exception in thread Thread-7:
Traceback (most recent call last):
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/threading.py", line 973, in _bootstrap_inner
self.run()
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/threading.py", line 910, in run
self._target(*self._args, **self._kwargs)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/core/utils/_contextlib.py", line 117, in decorate_context
return func(*args, **kwargs)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/transformers/generation/utils.py", line 2024, in generate
result = self._sample(
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/transformers/generation/utils.py", line 3036, in _sample
outputs = self(**model_inputs, return_dict=True)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/core/nn/modules/module.py", line 391, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/core/nn/modules/module.py", line 405, in _call_impl
return forward_call(*args, **kwargs)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/transformers/models/baichuan/modeling_baichuan.py", line 1787, in forward
outputs = self.model(
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/core/nn/modules/module.py", line 391, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/core/nn/modules/module.py", line 405, in _call_impl
return forward_call(*args, **kwargs)
File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.9/site-packages/mindnlp/transformers/models/baichuan/modeling_baichuan.py", line 1560, in forward
bsz = inputs_embeds.size(0)
TypeError: 'int' object is not callable
就不会改了。