HI,I fail to set up Tabular model for inference.
learn = load_learner(path/model.pkl') #previously has been exported with .export()
test_df = inderence_dataset_prep() #custom func to return inference dataset
test_dl = learn.dls.test_dl(test_df)
learn.get_preds(dl=test_dl)
this returns error:
…
AttributeError: ‘DataFrame’ object has no attribute ‘with_cont’
Thank you!
Does your inderence_dataset_prep return a DataFrame?
Also what version of fastai are you using?
Can you also try doing the following:
to = learn.dls.train_ds.new(test_df)
to.process()
dl = TabDataLoader(to)
learn.get_preds(dl=dl)
Hi, Zachary!
my fastai version is ‘2.0.13’
yes, inderence_dataset_prep returns DataFrame
Zachary, now it’s better:
error:
AssertionError: nan values in price but not in setup training set
Do I have to specify procs with to again?
If so, does it forget my categories dictionary and normalization together with fill-missing policy?
My procs where procs=[Categorify, FillMissing, Normalize].
When specifying cpu=False
learn = load_learner(models_path/'model.pkl', cpu=False)
learn.get_preds(dl=dl)
gives error:
TypeError: object of type 'NoneType' has no len()
How to run the model in GPU mode?
What’s happening here is fastai will preprocess and denote what classes are missing and what are not. So in your training set you initially used, if you expect values to be missing you should make a dummy row with np.nan for the columns that at one point may be missing. Note: this could also affect how inference goes since those values won’t ever “really” be trained on (one sample doesn’t do much at all), but that is why you have the error
I am having the exact same issue. Code and traceback below:
Code:
import pandas as pd
import numpy as np
from fastai.tabular.all import *
train_features = pd.read_csv('train_features.csv')
train_targets = pd.read_csv('train_targets_scored.csv')
test_features = pd.read_csv('test_features.csv')
cat_names = ['cp_type', 'cp_time', 'cp_dose']
cont_names = [x for x in train_features.columns if x not in cat_names]
procs = [Categorify, FillMissing, Normalize]
splits = RandomSplitter(valid_pct=0.2)(range_of(train_features))
data = pd.concat([train_features, train_targets], axis=1)
to = TabularPandas(data, procs=[Categorify, FillMissing,Normalize],
cat_names = cat_names,
cont_names = cont_names,
y_names=train_targets.columns.to_list(),
y_block=MultiCategoryBlock(),
splits=splits)
trn_dl = TabDataLoader(to.train, bs=64, shuffle=True, drop_last=True)
val_dl = TabDataLoader(to.valid, bs=128)
dls = DataLoaders(trn_dl, val_dl)
learn = tabular_learner(dls, layers=[300, 250], n_out=206)
learn.fit_one_cycle(1, 0.00363078061491251)
All of the above runs no problem, but getting predictions on my test set (not part of to, the TabularPandas object) is giving me issues. I tried two things:
dl = learn.dls.test_dl(test_features)
learn.get_preds(dl=dl)
This gives the following traceback:
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-7-68608cc7e448> in <module>()
1 dl = learn.dls.test_dl(test_features)
2
----> 3 learn.get_preds(dl=dl)
14 frames
/usr/local/lib/python3.6/dist-packages/fastai/learner.py in get_preds(self, ds_idx, dl, with_input, with_decoded, with_loss, act, inner, reorder, cbs, **kwargs)
233 if with_loss: ctx_mgrs.append(self.loss_not_reduced())
234 with ContextManagers(ctx_mgrs):
--> 235 self._do_epoch_validate(dl=dl)
236 if act is None: act = getattr(self.loss_func, 'activation', noop)
237 res = cb.all_tensors()
/usr/local/lib/python3.6/dist-packages/fastai/learner.py in _do_epoch_validate(self, ds_idx, dl)
186 if dl is None: dl = self.dls[ds_idx]
187 self.dl = dl
--> 188 with torch.no_grad(): self._with_events(self.all_batches, 'validate', CancelValidException)
189
190 def _do_epoch(self):
/usr/local/lib/python3.6/dist-packages/fastai/learner.py in _with_events(self, f, event_type, ex, final)
153
154 def _with_events(self, f, event_type, ex, final=noop):
--> 155 try: self(f'before_{event_type}') ;f()
156 except ex: self(f'after_cancel_{event_type}')
157 finally: self(f'after_{event_type}') ;final()
/usr/local/lib/python3.6/dist-packages/fastai/learner.py in all_batches(self)
159 def all_batches(self):
160 self.n_iter = len(self.dl)
--> 161 for o in enumerate(self.dl): self.one_batch(*o)
162
163 def _do_one_batch(self):
/usr/local/lib/python3.6/dist-packages/fastai/data/load.py in __iter__(self)
102 for b in _loaders[self.fake_l.num_workers==0](self.fake_l):
103 if self.device is not None: b = to_device(b, self.device)
--> 104 yield self.after_batch(b)
105 self.after_iter()
106 if hasattr(self, 'it'): del(self.it)
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in __call__(self, o)
196 self.fs.append(t)
197
--> 198 def __call__(self, o): return compose_tfms(o, tfms=self.fs, split_idx=self.split_idx)
199 def __repr__(self): return f"Pipeline: {' -> '.join([f.name for f in self.fs if f.name != 'noop'])}"
200 def __getitem__(self,i): return self.fs[i]
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in compose_tfms(x, tfms, is_enc, reverse, **kwargs)
148 for f in tfms:
149 if not is_enc: f = f.decode
--> 150 x = f(x, **kwargs)
151 return x
152
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in __call__(self, x, **kwargs)
111 "A transform that always take tuples as items"
112 _retain = True
--> 113 def __call__(self, x, **kwargs): return self._call1(x, '__call__', **kwargs)
114 def decode(self, x, **kwargs): return self._call1(x, 'decode', **kwargs)
115 def _call1(self, x, name, **kwargs):
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in _call1(self, x, name, **kwargs)
114 def decode(self, x, **kwargs): return self._call1(x, 'decode', **kwargs)
115 def _call1(self, x, name, **kwargs):
--> 116 if not _is_tuple(x): return getattr(super(), name)(x, **kwargs)
117 y = getattr(super(), name)(list(x), **kwargs)
118 if not self._retain: return y
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in __call__(self, x, **kwargs)
71 @property
72 def name(self): return getattr(self, '_name', _get_name(self))
---> 73 def __call__(self, x, **kwargs): return self._call('encodes', x, **kwargs)
74 def decode (self, x, **kwargs): return self._call('decodes', x, **kwargs)
75 def __repr__(self): return f'{self.name}:\nencodes: {self.encodes}decodes: {self.decodes}'
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in _call(self, fn, x, split_idx, **kwargs)
81 def _call(self, fn, x, split_idx=None, **kwargs):
82 if split_idx!=self.split_idx and self.split_idx is not None: return x
---> 83 return self._do_call(getattr(self, fn), x, **kwargs)
84
85 def _do_call(self, f, x, **kwargs):
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in _do_call(self, f, x, **kwargs)
87 if f is None: return x
88 ret = f.returns_none(x) if hasattr(f,'returns_none') else None
---> 89 return retain_type(f(x, **kwargs), x, ret)
90 res = tuple(self._do_call(f, x_, **kwargs) for x_ in x)
91 return retain_type(res, x)
/usr/local/lib/python3.6/dist-packages/fastcore/dispatch.py in __call__(self, *args, **kwargs)
110 if not f: return args[0]
111 if self.inst is not None: f = MethodType(f, self.inst)
--> 112 return f(*args, **kwargs)
113
114 def __get__(self, inst, owner):
/usr/local/lib/python3.6/dist-packages/fastai/tabular/core.py in encodes(self, to)
321
322 def encodes(self, to):
--> 323 if not to.with_cont: res = (tensor(to.cats).long(),)
324 else: res = (tensor(to.cats).long(),tensor(to.conts).float())
325 ys = [n for n in to.y_names if n in to.items.columns]
/usr/local/lib/python3.6/dist-packages/pandas/core/generic.py in __getattr__(self, name)
5272 if self._info_axis._can_hold_identifiers_and_holds_name(name):
5273 return self[name]
-> 5274 return object.__getattribute__(self, name)
5275
5276 def __setattr__(self, name: str, value) -> None:
AttributeError: 'DataFrame' object has no attribute 'with_cont'
One thing I noticed here is that in the second to last frame of the traceback, the line highlighted, if not to.with_cont: res = (tensor(to.cats).long(),) might be the cause because when I run to.with_cont it outputs True, so I don’t think this line should run.
to = learn.dls.train_ds.new(test_features)
to.process()
dl = TabDataLoader(to)
learn.get_preds(dl=dl)
This gives the following traceback:
---------------------------------------------------------------------------
KeyError Traceback (most recent call last)
<ipython-input-8-adcc56cec048> in <module>()
1 to = learn.dls.train_ds.new(test_features)
----> 2 to.process()
3 dl = TabDataLoader(top_k_accuracy)
4 learn.get_preds(dl=dl)
8 frames
/usr/local/lib/python3.6/dist-packages/fastai/tabular/core.py in process(self)
176 def show(self, max_n=10, **kwargs): display_df(self.new(self.all_cols[:max_n]).decode().items)
177 def setup(self): self.procs.setup(self)
--> 178 def process(self): self.procs(self)
179 def loc(self): return self.items.loc
180 def iloc(self): return _TabIloc(self)
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in __call__(self, o)
196 self.fs.append(t)
197
--> 198 def __call__(self, o): return compose_tfms(o, tfms=self.fs, split_idx=self.split_idx)
199 def __repr__(self): return f"Pipeline: {' -> '.join([f.name for f in self.fs if f.name != 'noop'])}"
200 def __getitem__(self,i): return self.fs[i]
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in compose_tfms(x, tfms, is_enc, reverse, **kwargs)
148 for f in tfms:
149 if not is_enc: f = f.decode
--> 150 x = f(x, **kwargs)
151 return x
152
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in __call__(self, x, **kwargs)
71 @property
72 def name(self): return getattr(self, '_name', _get_name(self))
---> 73 def __call__(self, x, **kwargs): return self._call('encodes', x, **kwargs)
74 def decode (self, x, **kwargs): return self._call('decodes', x, **kwargs)
75 def __repr__(self): return f'{self.name}:\nencodes: {self.encodes}decodes: {self.decodes}'
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in _call(self, fn, x, split_idx, **kwargs)
81 def _call(self, fn, x, split_idx=None, **kwargs):
82 if split_idx!=self.split_idx and self.split_idx is not None: return x
---> 83 return self._do_call(getattr(self, fn), x, **kwargs)
84
85 def _do_call(self, f, x, **kwargs):
/usr/local/lib/python3.6/dist-packages/fastcore/transform.py in _do_call(self, f, x, **kwargs)
87 if f is None: return x
88 ret = f.returns_none(x) if hasattr(f,'returns_none') else None
---> 89 return retain_type(f(x, **kwargs), x, ret)
90 res = tuple(self._do_call(f, x_, **kwargs) for x_ in x)
91 return retain_type(res, x)
/usr/local/lib/python3.6/dist-packages/fastcore/dispatch.py in __call__(self, *args, **kwargs)
110 if not f: return args[0]
111 if self.inst is not None: f = MethodType(f, self.inst)
--> 112 return f(*args, **kwargs)
113
114 def __get__(self, inst, owner):
/usr/local/lib/python3.6/dist-packages/fastai/data/transforms.py in encodes(self, o)
257 self.vocab = CategoryMap(list(vals), add_na=self.add_na)
258
--> 259 def encodes(self, o): return TensorMultiCategory([self.vocab.o2i[o_] for o_ in o])
260 def decodes(self, o): return MultiCategory ([self.vocab [o_] for o_ in o])
261
/usr/local/lib/python3.6/dist-packages/fastai/data/transforms.py in <listcomp>(.0)
257 self.vocab = CategoryMap(list(vals), add_na=self.add_na)
258
--> 259 def encodes(self, o): return TensorMultiCategory([self.vocab.o2i[o_] for o_ in o])
260 def decodes(self, o): return MultiCategory ([self.vocab [o_] for o_ in o])
261
KeyError: 'cp_type'
cp_type is one of the categorical variables in cat_names above.
No y_names is only the dataframe columns for train_targets, not data
I found a couple of kernels from the Lish-MOA Kaggle competition using FastAI, and I solved my issues by removing the y_block=MultiCategoryBlock() parameter for TabularPandas.