# Failed to convert a NumPy array to a Tensor (Unsupported object type dict)

**URL:** <https://discuss.pytorch.org/t/failed-to-convert-a-numpy-array-to-a-tensor-unsupported-object-type-dict/125487>\
**Category:** tensorboard\
**Created:** [June 30, 2021, 7:28pm UTC](https://discuss.pytorch.org/t/failed-to-convert-a-numpy-array-to-a-tensor-unsupported-object-type-dict/125487 "2021-06-30T19:28:24Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![samm](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/samm/32/36169_2.png) [@samm](https://discuss.pytorch.org/u/samm)\
**Post date:** [June 30, 2021, 7:28pm UTC](https://discuss.pytorch.org/t/failed-to-convert-a-numpy-array-to-a-tensor-unsupported-object-type-dict/125487/1 "2021-06-30T19:28:24Z")

</div>

> history = model.fit\_generator(train\_generator, epochs=epochs, steps\_per\_epoch=train\_steps, verbose=1, callbacks=[checkpoint], validation\_data=val\_generator, validation\_steps=val\_steps)
> 
> def create\_sequences(tokenizer, max\_length, desc\_list, photo, vocab\_size):  
> X1, X2, y = list(), list(), list()  
> for desc in desc\_list:  
> seq = tokenizer.texts\_to\_sequences([desc])[0]  
> for i in range(1, len(seq)):  
> in\_seq, out\_seq = seq[:i], seq[i] in\_seq = pad\_sequences([in\_seq], maxlen=max\_length)[0]  
> out\_seq = to\_categorical([out\_seq], num\_classes=vocab\_size)[0]  
> X1.append(photo)  
> X2.append(in\_seq)  
> y.append(out\_seq)  
> return array(X1), array(X2), array(y)
> 
> def data\_generator(descriptions, photos, tokenizer, max\_length, imgsIds, vocab\_size):  
> while 1:  
> for ind in range(len(imgsIds)):  
> photo = photos[ind]  
> key = imgsIds[ind]  
> desc\_list = descriptions[str(key)]  
> in\_img, in\_seq, out\_word = create\_sequences( tokenizer, max\_length, desc\_list, photo, vocab\_size) yield [in\_img, in\_seq], out\_word

i got

> 82783/82783 [==============================] - ETA: 0s - loss: 4.0302Traceback (most recent call last):  
> File “rnn\_flickr\_fit.py”, line 281, in   
> main(sys.argv)  
> File “rnn\_flickr\_fit.py”, line 278, in main  
> fit\_model(train, train\_descriptions, train\_rnn\_input, val, val\_descriptions, val\_rnn\_input)  
> File “rnn\_flickr\_fit.py”, line 265, in fit\_model  
> history = model.fit(train\_generator, epochs=epochs, steps\_per\_epoch=train\_steps, verbose=1, callbacks=[checkpoint], validation\_data=val\_generator, validation\_steps=val\_steps)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/keras/engine/training.py”, line 1213, in fit  
> steps\_per\_execution=self.\_steps\_per\_execution)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/keras/engine/data\_adapter.py”, line 1364, in get\_data\_handler  
> return DataHandler(\*args, \*\*kwargs)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/keras/engine/data\_adapter.py”, line 1166, in **init**  
> model=model)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/keras/engine/data\_adapter.py”, line 811, in **init**  
> peek = \_process\_tensorlike(peek)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/keras/engine/data\_adapter.py”, line 1046, in \_process\_tensorlike  
> inputs = nest.map\_structure(\_convert\_numpy\_and\_scipy, inputs)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/util/nest.py”, line 867, in map\_structure  
> structure[0], [func(\*x) for x in entries],  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/util/nest.py”, line 867, in   
> structure[0], [func(\*x) for x in entries],  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/keras/engine/data\_adapter.py”, line 1041, in \_convert\_numpy\_and\_scipy  
> return ops.convert\_to\_tensor\_v2\_with\_dispatch(x, dtype=dtype)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/util/dispatch.py”, line 206, in wrapper  
> return target(\*args, \*\*kwargs)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/framework/ops.py”, line 1431, in convert\_to\_tensor\_v2\_with\_dispatch  
> value, dtype=dtype, dtype\_hint=dtype\_hint, name=name)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/framework/ops.py”, line 1441, in convert\_to\_tensor\_v2  
> as\_ref=False)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/profiler/trace.py”, line 163, in wrapped  
> return func(\*args, \*\*kwargs)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/framework/ops.py”, line 1566, in convert\_to\_tensor  
> ret = conversion\_func(value, dtype=dtype, name=name, as\_ref=as\_ref)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/framework/tensor\_conversion\_registry.py”, line 52, in \_default\_conversion\_function  
> return constant\_op.constant(value, dtype, name=name)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/framework/constant\_op.py”, line 265, in constant  
> allow\_broadcast=True)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/framework/constant\_op.py”, line 276, in \_constant\_impl  
> return \_constant\_eager\_impl(ctx, value, dtype, shape, verify\_shape)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/framework/constant\_op.py”, line 301, in \_constant\_eager\_impl  
> t = convert\_to\_eager\_tensor(value, ctx, dtype)  
> File “/home/user/.local/lib/python3.6/site-packages/tensorflow/python/framework/constant\_op.py”, line 98, in convert\_to\_eager\_tensor  
> return ops.EagerTensor(value, ctx.device\_name, dtype)  
> ValueError: Failed to convert a NumPy array to a Tensor (Unsupported object type dict).  
> 2021-06-30 20:08:00.122545: W tensorflow/core/kernels/data/generator\_dataset\_op.cc:107] Error occurred when finalizing GeneratorDataset iterator: Failed precondition: Python interpreter state is not initialized. The process may be terminated.  
> [[{{node PyFunc}}]]

---

<div class="post-metadata">

**Author:** ![ptrblck](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/ptrblck/32/1823_2.png) [@ptrblck](https://discuss.pytorch.org/u/ptrblck)\
**Post date:** [June 30, 2021, 7:51pm UTC](https://discuss.pytorch.org/t/failed-to-convert-a-numpy-array-to-a-tensor-unsupported-object-type-dict/125487/2 "2021-06-30T19:51:48Z")

</div>

It seems you are using TensorFlow, so you might want to post the question in their discussion board or StackOverflow. 😉

Based on the error message it seems that `dict`s are not supported as inputs to the conversion method.

---

<div class="post-metadata">

**Author:** ![samm](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/samm/32/36169_2.png) [@samm](https://discuss.pytorch.org/u/samm)\
**Post date:** [July 3, 2021, 1:17pm UTC](https://discuss.pytorch.org/t/failed-to-convert-a-numpy-array-to-a-tensor-unsupported-object-type-dict/125487/3 "2021-07-03T13:17:27Z")

</div>

thanks a lot … i did