list dequeの要素の合計を知りたい

dequeはprint結果が嘘をついてなければ一次元です
出ているエラーの意味が分かりません　単一の要素の整数ということは
np.sumはlistの要素の合計ではなく入れた一つの数字をそのまま返すだけなんですか？

list dequeの要素の合計をだす関数はありませんか？わかる人教えてください
もしくは正しいsumの使い方がわかる人教えてください

追加してるところ
＃torchからnumpyに型を変えてます
TD　print結果[0.04558122]
self.buffer.append(TD[0])
試したコード
p=memory_TDerror.buffer/np.sum(memory_TDerror.buffer)
idx=np.random.choice(np.arange(len(memory_TDerror.buffer)), size=batch_size,replace=False,p=p)

ただのsumでも同様のエラー


memory_TDerror.bufferのプリント結果
deque([0.048867036, 0.04883557, 0.048770107, 0.048671108, 0.04853942, 0.048376095, 0.04818251, 0.047960237, 0.0477111, 0.04743718, 0.04714063, 0.04682391, 0.046489548, 0.046140272, 0.0457788, 0.045408025, 0.045030855, 0.044650193, 0.04426902, 0.043890193, 0.043516655, 0.043151148, 0.04279647, 0.042455144, 0.042129703, 0.041822523, 0.041535787, 0.041271534, 0.041031655, 0.040817846, 0.040631555, 0.040474117, 0.040346622, 0.040249933, 0.040184755, 0.04015153, 0.0401505, 0.040181704, 0.04024497, 0.040339865, 0.040465776, 0.040621877, 0.040807124, 0.04102028, 0.041259866, 0.041524258, 0.041811615, 0.042119924, 0.042447004, 0.042790532, 0.04314801, 0.0435168, 0.043894205, 0.044277422, 0.04466352, 0.04504957, 0.045432527, 0.04580952, 0.046177484, 0.046533555, 0.046874855, 0.04719862, 0.047502182, 0.047783088, 0.04803901, 0.048267793, 0.04846752, 0.048636504, 0.04877331, 0.04887679, 0.048946057, 0.048980456, 0.048979715, 0.048943806, 0.04887302, 0.048767887, 0.048629317, 0.04845845, 0.048256636, 0.04802562, 0.04776726, 0.04748366, 0.047177166, 0.04685022, 0.046505485, 0.04614572, 0.04577385, 0.04539271, 0.04500537, 0.044614784, 0.04422401, 0.043836035, 0.043453738, 0.043080032, 0.0427177, 0.042369395, 0.042037617], maxlen=10000)

でてるエラー
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-4-f40d86d018aa> in <module>
    427                                                              multireward_steps=multireward_steps,
    428                                                              tau=tau,
--> 429                                                              mord_update=mord_update)  
    430 
    431         if done or t==max_number_of_steps - 1:

<ipython-input-4-f40d86d018aa> in Double_R2D2_IQN_pioritized_Nstep_NAF_replay(self, batch_size, gamma, step, state_size, action_size, multireward_steps, tau, mord_update)
    227         trin_x=list(range(batch_size))
    228         weights = torch.ones(batch_size,device='cuda:0')
--> 229         p=memory_TDerror.buffer/np.sum(memory_TDerror.buffer)
    230         print(memory_TDerror.buffer)
    231         idx=np.random.choice(np.arange(len(memory_TDerror.buffer)), size=batch_size,

~\Anaconda3\envs\pyflan\lib\site-packages\torch\tensor.py in __rdiv__(self, other)
    405     def __rdiv__(self, other):
    406         if self.dtype.is_floating_point:
--> 407             return self.reciprocal() * other
    408         else:
    409             return (self.double().reciprocal() * other).type_as(self)

TypeError: only integer tensors of a single element can be converted to an index