Huawei H13-323_V1.0 Practice Exams
Last updated on Oct 22,2026- Exam Code: H13-323_V1.0
- Exam Name: HCIP-AI Solution Architect V1.0
- Certification Provider: Huawei
- Latest update: Oct 22,2026
What are the built-in loss functions under tf.keras.losses? (Multiple choice)
- A . Squared Error Loss
- B . Binary Cross Loss
- C . Similarity
- D . Absolute percentage error
Which of the following are the basic disciplines of artificial intelligence? (Multiple choice)
- A . Mathematics
- B . Finance
- C . Science of thinking
- D . Computer Science
Which of the following statements about image content review services is incorrect?
- A . politics is the detection result of politically sensitive figures
- B . terrorism is a result related to political violence and terrorism
- C . confidence represents the confidence level, ranging from 0 to 100
- D . label represents each detection result label
On-Device execution, that is, whole-image offloading execution, fully utilizes the computing power of the Intel Xeon chips, can greatly reduce the interaction overhead, and thus increase the accelerator occupancy rate. Which of the following descriptions about On-Device execution is incorrect?
- A . MindSpore achieves decentralized autonomous A11 Reduce through adaptive graph optimization driven by gradient data, with consistent gradient aggregation and full flow of computing and communication.
- B . Challenges of model execution under super chip computing power: memory wall problem, interaction overhead, and difficulty in providing big data. Part of the model is executed on the host, and part of the model is executed on the device. The interaction overhead is even much greater than the execution overhead, resulting in low accelerator utilization.
- C . MindSpore uses chip-oriented deep graph optimization technology to reduce synchronization waiting and maximize the parallelism of "data computing and communication". The training performance is better than the host side graph debugging method.
- D . Challenges of distributed gradient aggregation under super chip computing power: When ResNet50 single-select generation takes 20mms, it will generate central control synchronization overhead and frequent synchronization communication overhead. The traditional method requires 3 synchronizations to complete A11Reduce, while the data-driven method has independent A11Reduce and no control overhead.
Which of the following descriptions of the large model are correct? (Multiple choice)
- A . LLM may have hallucination problems when generating text, that is, sometimes it may be "serious nonsense"
- B . The cost and cycle of large-scale model training increase, making it difficult to include timely data in training, affecting the model’s ability to answer relevant questions
- C . LLM can process and answer all time-sensitive questions instantly, such as recommending currently popular movies
- D . Enterprises can ensure data security by uploading all data to the cloud and using online big models to handle all business calculations.
Which of the following attributes does TensorFlow 2.0 not support in creating tensors?
- A . create
- B . f11
- C . zeros
- D . constant
Federated learning uses different data sources to train models while ensuring data privacy and security, and further breaks through data bottlenecks.
- A . True
- B . False
HUAWEI What value does HiAI empower APP? (Multiple choice)
- A . Security
- B . Stability
- C . Real-time
- D . Anytime
What are the common clustering algorithms? (Multiple choice)
- A . Density Clustering
- B . Hierarchical Clustering
- C . Spectral clustering
- D . K-means
Machine learning is a technique that uses algorithms to learn from data.
- A . True
- B . False