Select the options below that make the statements below corr…

Select the options below that make the statements below correct: Models based on the encoder/decoder Seq2Seq architecture are [ans1]. Transformers are [ans2] models that [ans3] about sequence order. BERT models are [ans4], BART models are [ans5], and GPT models are [ans6]. In the original transformer architecture, [ans7] must be divisible by [ans8].

Consider a recurrent network with a single layer GRU module…

Consider a recurrent network with a single layer GRU module with an input size of 5 and a hidden size of 7, processing sequences of 20 time steps. Fill in the blanks in the following table, breaking down the number of weights and biases per layer. GRU Layer Weights Biases #1 [ans1] [ans2]

Given the network below, answer the following questions. T…

Given the network below, answer the following questions. The images processed by this network are [ans1]. The first convolutional layer has what kind of padding? [ans2] The second convolutional layer has what kind of padding? [ans3] The first convolutional layer has a stride of [ans4]. The second convolutional layer has a stride of [ans5]. This network is suitable for what computer vision task? [ans6]

In order to hit your earnings goals, you expect that you wil…

In order to hit your earnings goals, you expect that you will need to raise an additional $600,000 in a Series B offering three years from now.  If those investors expect to earn a 25% return on their investment, what percentage of the firm’s equity will they need to own?