xgboost模型训练时报错

Terence888

1. #xgboost拟合训练数据，进行模型训练
   
2. import xgboost
   
3. xgb_reg = xgboost.XGBRegressor()
   
4. xgb_reg.fit(X_train, y_train)
   
5. 
   
6. #模型评估，对测试集进行预测
   
7. y_preds = xgb_reg.predict(X_test)
   
8. 
   
9. output = pd.DataFrame({'Id':test_data.Id,'SalePrice':y_preds})
   
10. output

复制代码

报错

1. ---------------------------------------------------------------------------
   
2. XGBoostError                              Traceback (most recent call last)
   
3. Cell In[40], line 4
   
4.       2 import xgboost
   
5.       3 xgb_reg = xgboost.XGBRegressor()
   
6. ----> 4 xgb_reg.fit(X_train, y_train)
   
7.       6 #模型评估，对测试集进行预测
   
8.       7 y_preds = xgb_reg.predict(X_test)
   
9. 
   
10. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:730, in require_keyword_args.<locals>.throw_if.<locals>.inner_f(*args, **kwargs)
    
11.     728 for k, arg in zip(sig.parameters, args):
    
12.     729     kwargs[k] = arg
    
13. --> 730 return func(**kwargs)
    
14. 
    
15. File /opt/conda/lib/python3.10/site-packages/xgboost/sklearn.py:1055, in XGBModel.fit(self, X, y, sample_weight, base_margin, eval_set, eval_metric, early_stopping_rounds, verbose, xgb_model, sample_weight_eval_set, base_margin_eval_set, feature_weights, callbacks)
    
16.    1053 with config_context(verbosity=self.verbosity):
    
17.    1054     evals_result: TrainingCallback.EvalsLog = {}
    
18. -> 1055     train_dmatrix, evals = _wrap_evaluation_matrices(
    
19.    1056         missing=self.missing,
    
20.    1057         X=X,
    
21.    1058         y=y,
    
22.    1059         group=None,
    
23.    1060         qid=None,
    
24.    1061         sample_weight=sample_weight,
    
25.    1062         base_margin=base_margin,
    
26.    1063         feature_weights=feature_weights,
    
27.    1064         eval_set=eval_set,
    
28.    1065         sample_weight_eval_set=sample_weight_eval_set,
    
29.    1066         base_margin_eval_set=base_margin_eval_set,
    
30.    1067         eval_group=None,
    
31.    1068         eval_qid=None,
    
32.    1069         create_dmatrix=self._create_dmatrix,
    
33.    1070         enable_categorical=self.enable_categorical,
    
34.    1071         feature_types=self.feature_types,
    
35.    1072     )
    
36.    1073     params = self.get_xgb_params()
    
37.    1075     if callable(self.objective):
    
38. 
    
39. File /opt/conda/lib/python3.10/site-packages/xgboost/sklearn.py:521, in _wrap_evaluation_matrices(missing, X, y, group, qid, sample_weight, base_margin, feature_weights, eval_set, sample_weight_eval_set, base_margin_eval_set, eval_group, eval_qid, create_dmatrix, enable_categorical, feature_types)
    
40.     501 def _wrap_evaluation_matrices(
    
41.     502     missing: float,
    
42.     503     X: Any,
    
43.    (...)
    
44.     517     feature_types: Optional[FeatureTypes],
    
45.     518 ) -> Tuple[Any, List[Tuple[Any, str]]]:
    
46.     519     """Convert array_like evaluation matrices into DMatrix.  Perform validation on the
    
47.     520     way."""
    
48. --> 521     train_dmatrix = create_dmatrix(
    
49.     522         data=X,
    
50.     523         label=y,
    
51.     524         group=group,
    
52.     525         qid=qid,
    
53.     526         weight=sample_weight,
    
54.     527         base_margin=base_margin,
    
55.     528         feature_weights=feature_weights,
    
56.     529         missing=missing,
    
57.     530         enable_categorical=enable_categorical,
    
58.     531         feature_types=feature_types,
    
59.     532         ref=None,
    
60.     533     )
    
61.     535     n_validation = 0 if eval_set is None else len(eval_set)
    
62.     537     def validate_or_none(meta: Optional[Sequence], name: str) -> Sequence:
    
63. 
    
64. File /opt/conda/lib/python3.10/site-packages/xgboost/sklearn.py:958, in XGBModel._create_dmatrix(self, ref, **kwargs)
    
65.     956 if _can_use_qdm(self.tree_method) and self.booster != "gblinear":
    
66.     957     try:
    
67. --> 958         return QuantileDMatrix(
    
68.     959             **kwargs, ref=ref, nthread=self.n_jobs, max_bin=self.max_bin
    
69.     960         )
    
70.     961     except TypeError:  # `QuantileDMatrix` supports lesser types than DMatrix
    
71.     962         pass
    
72. 
    
73. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:730, in require_keyword_args.<locals>.throw_if.<locals>.inner_f(*args, **kwargs)
    
74.     728 for k, arg in zip(sig.parameters, args):
    
75.     729     kwargs[k] = arg
    
76. --> 730 return func(**kwargs)
    
77. 
    
78. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:1529, in QuantileDMatrix.__init__(self, data, label, weight, base_margin, missing, silent, feature_names, feature_types, nthread, max_bin, ref, group, qid, label_lower_bound, label_upper_bound, feature_weights, enable_categorical, data_split_mode)
    
79.    1509     if any(
    
80.    1510         info is not None
    
81.    1511         for info in (
    
82.    (...)
    
83.    1522         )
    
84.    1523     ):
    
85.    1524         raise ValueError(
    
86.    1525             "If data iterator is used as input, data like label should be "
    
87.    1526             "specified as batch argument."
    
88.    1527         )
    
89. -> 1529 self._init(
    
90.    1530     data,
    
91.    1531     ref=ref,
    
92.    1532     label=label,
    
93.    1533     weight=weight,
    
94.    1534     base_margin=base_margin,
    
95.    1535     group=group,
    
96.    1536     qid=qid,
    
97.    1537     label_lower_bound=label_lower_bound,
    
98.    1538     label_upper_bound=label_upper_bound,
    
99.    1539     feature_weights=feature_weights,
    
100.    1540     feature_names=feature_names,
     
101.    1541     feature_types=feature_types,
     
102.    1542     enable_categorical=enable_categorical,
     
103.    1543 )
     
104. 
     
105. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:1588, in QuantileDMatrix._init(self, data, ref, enable_categorical, **meta)
     
106.    1576 config = make_jcargs(
     
107.    1577     nthread=self.nthread, missing=self.missing, max_bin=self.max_bin
     
108.    1578 )
     
109.    1579 ret = _LIB.XGQuantileDMatrixCreateFromCallback(
     
110.    1580     None,
     
111.    1581     it.proxy.handle,
     
112.    (...)
     
113.    1586     ctypes.byref(handle),
     
114.    1587 )
     
115. -> 1588 it.reraise()
     
116.    1589 # delay check_call to throw intermediate exception first
     
117.    1590 _check_call(ret)
     
118. 
     
119. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:576, in DataIter.reraise(self)
     
120.     574 exc = self._exception
     
121.     575 self._exception = None
     
122. --> 576 raise exc
     
123. 
     
124. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:557, in DataIter._handle_exception(self, fn, dft_ret)
     
125.     554     return dft_ret
     
126.     556 try:
     
127. --> 557     return fn()
     
128.     558 except Exception as e:  # pylint: disable=broad-except
     
129.     559     # Defer the exception in order to return 0 and stop the iteration.
     
130.     560     # Exception inside a ctype callback function has no effect except
     
131.     561     # for printing to stderr (doesn't stop the execution).
     
132.     562     tb = sys.exc_info()[2]
     
133. 
     
134. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:641, in DataIter._next_wrapper.<locals>.<lambda>()
     
135.     638     self._data_ref = ref
     
136.     640 # pylint: disable=not-callable
     
137. --> 641 return self._handle_exception(lambda: self.next(input_data), 0)
     
138. 
     
139. File /opt/conda/lib/python3.10/site-packages/xgboost/data.py:1280, in SingleBatchInternalIter.next(self, input_data)
     
140.    1278     return 0
     
141.    1279 self.it += 1
     
142. -> 1280 input_data(**self.kwargs)
     
143.    1281 return 1
     
144. 
     
145. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:730, in require_keyword_args.<locals>.throw_if.<locals>.inner_f(*args, **kwargs)
     
146.     728 for k, arg in zip(sig.parameters, args):
     
147.     729     kwargs[k] = arg
     
148. --> 730 return func(**kwargs)
     
149. 
     
150. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:633, in DataIter._next_wrapper.<locals>.input_data(data, feature_names, feature_types, **kwargs)
     
151.     631 self._temporary_data = (new, cat_codes, feature_names, feature_types)
     
152.     632 dispatch_proxy_set_data(self.proxy, new, cat_codes, self._allow_host)
     
153. --> 633 self.proxy.set_info(
     
154.     634     feature_names=feature_names,
     
155.     635     feature_types=feature_types,
     
156.     636     **kwargs,
     
157.     637 )
     
158.     638 self._data_ref = ref
     
159. 
     
160. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:730, in require_keyword_args.<locals>.throw_if.<locals>.inner_f(*args, **kwargs)
     
161.     728 for k, arg in zip(sig.parameters, args):
     
162.     729     kwargs[k] = arg
     
163. --> 730 return func(**kwargs)
     
164. 
     
165. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:932, in DMatrix.set_info(self, label, weight, base_margin, group, qid, label_lower_bound, label_upper_bound, feature_names, feature_types, feature_weights)
     
166.     929 from .data import dispatch_meta_backend
     
167.     931 if label is not None:
     
168. --> 932     self.set_label(label)
     
169.     933 if weight is not None:
     
170.     934     self.set_weight(weight)
     
171. 
     
172. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:1070, in DMatrix.set_label(self, label)
     
173.    1061 """Set label of dmatrix
     
174.    1062
     
175.    1063 Parameters
     
176.    (...)
     
177.    1066     The label information to be set into DMatrix
     
178.    1067 """
     
179.    1068 from .data import dispatch_meta_backend
     
180. -> 1070 dispatch_meta_backend(self, label, "label", "float")
     
181. 
     
182. File /opt/conda/lib/python3.10/site-packages/xgboost/data.py:1225, in dispatch_meta_backend(matrix, data, name, dtype)
     
183.    1223     return
     
184.    1224 if _is_pandas_series(data):
     
185. -> 1225     _meta_from_pandas_series(data, name, dtype, handle)
     
186.    1226     return
     
187.    1227 if _is_dlpack(data):
     
188. 
     
189. File /opt/conda/lib/python3.10/site-packages/xgboost/data.py:545, in _meta_from_pandas_series(data, name, dtype, handle)
     
190.     543     data = data.to_dense()  # type: ignore
     
191.     544 assert len(data.shape) == 1 or data.shape[1] == 0 or data.shape[1] == 1
     
192. --> 545 _meta_from_numpy(data, name, dtype, handle)
     
193. 
     
194. File /opt/conda/lib/python3.10/site-packages/xgboost/data.py:1159, in _meta_from_numpy(data, field, dtype, handle)
     
195.    1157     raise ValueError("Masked array is not supported.")
     
196.    1158 interface_str = _array_interface(data)
     
197. -> 1159 _check_call(_LIB.XGDMatrixSetInfoFromInterface(handle, c_str(field), interface_str))
     
198. 
     
199. File /opt/conda/lib/python3.10/site-packages/xgboost/core.py:282, in _check_call(ret)
     
200.     271 """Check the return value of C API call
     
201.     272
     
202.     273 This function will raise exception when error occurs.
     
203.    (...)
     
204.     279     return value from API calls
     
205.     280 """
     
206.     281 if ret != 0:
     
207. --> 282     raise XGBoostError(py_str(_LIB.XGBGetLastError()))
     
208. 
     
209. XGBoostError: [13:24:13] /workspace/src/data/data.cc:507: Check failed: valid: Label contains NaN, infinity or a value too large.
     
210. Stack trace:
     
211.   [bt] (0) /opt/conda/lib/python3.10/site-packages/xgboost/lib/libxgboost.so(+0x3588ca) [0x7a256777a8ca]
     
212.   [bt] (1) /opt/conda/lib/python3.10/site-packages/xgboost/lib/libxgboost.so(+0x38a21d) [0x7a25677ac21d]
     
213.   [bt] (2) /opt/conda/lib/python3.10/site-packages/xgboost/lib/libxgboost.so(+0x38ab51) [0x7a25677acb51]
     
214.   [bt] (3) /opt/conda/lib/python3.10/site-packages/xgboost/lib/libxgboost.so(XGDMatrixSetInfoFromInterface+0xb0) [0x7a25675803a0]
     
215.   [bt] (4) /opt/conda/lib/python3.10/lib-dynload/../../libffi.so.8(+0x6a4a) [0x7a260b0e4a4a]
     
216.   [bt] (5) /opt/conda/lib/python3.10/lib-dynload/../../libffi.so.8(+0x5fea) [0x7a260b0e3fea]
     
217.   [bt] (6) /opt/conda/lib/python3.10/lib-dynload/_ctypes.cpython-310-x86_64-linux-gnu.so(+0x12461) [0x7a260b0fc461]
     
218.   [bt] (7) /opt/conda/lib/python3.10/lib-dynload/_ctypes.cpython-310-x86_64-linux-gnu.so(+0x86eb) [0x7a260b0f26eb]
     
219.   [bt] (8) /opt/conda/bin/python3.10(_PyObject_MakeTpCall+0x26b) [0x5cd1861cfa6b]
     

复制代码

FishC

根据错误信息，`XGBoostError: Label contains NaN, infinity or a value too large.` 你可以推断出 `y_train` 包含无效值(`NaN`、无穷大或过大的值)。

解决方法如下：

1. **检查和处理NaN值**：
   确保 `y_train` 中没有缺失值。
  

1. import numpy as np
   
2.    y_train = np.where(np.isnan(y_train), 0, y_train)  # 用0替换NaN值，可以根据实际情况选择合适的替换值

复制代码

2. **检查和处理无穷大值**：
   确保 `y_train` 中没有无穷大的值。
  

1. y_train = np.where(np.isinf(y_train), 0, y_train)  # 用0替换无穷大值，可以根据实际情况选择合适的替换值

复制代码

3. **检查和处理异常大的值**：
   确保 `y_train` 中没有异常大的值。
  

1. y_train = np.where(y_train > 1e10, 1e10, y_train)  # 用1e10限制最大值，可以根据实际阈值选择

复制代码

总之，执行以下综合处理以确保数据有效：

1. import numpy as np
   
2. 
   
3. # 处理 NaN 值
   
4. y_train = np.where(np.isnan(y_train), 0, y_train)
   
5. 
   
6. # 处理无穷大值
   
7. y_train = np.where(np.isinf(y_train), 0, y_train)
   
8. 
   
9. # 处理异常大的值
   
10. y_train = np.where(y_train > 1e10, 1e10, y_train)
    
11. 
    
12. # 然后再进行模型训练
    
13. xgb_reg.fit(X_train, y_train)

复制代码

确保 `X_train` 和 `y_train` 的长度一致，并确认 `X_train` 中也没有无效数据。如果问题仍然存在，详细检查数据处理流程，确保所有步骤中数据都没有被意外修改。

以上回复来自 FishC AI 小助理，如未能正确解答您的问题，请继续追问。