Artificial Intelligence

大语言模型LLM之为完成指令任务进行微调 Fine-Tuning for Following Instructions of Large Language Models (LLMs)

2025-01-25 18:00:00

作为LLM应用的一种，我们可以将预训练好的模型进一步微调，使其具有多用途性并能完成输入的指令 As an application of LLMs, we can further fine-tune the pre-trained model to make it more versatile and capable of completing given instructions.

大语言模型LLM之为分类任务进行微调 Fine-Tuning for Classification Tasks of Large Language Models (LLMs)

2025-01-19 15:30:00

作为LLM应用的一种，我们可以将预训练好的模型进一步微调，使其能完成专业的垃圾邮件分类任务 As one application of LLMs, we can further fine-tune a pre-trained model to enable it to perform specialized tasks, such as spam classification.

大语言模型LLM之使用无标签数据预训练 Pretraining On Unlabeled Data of Large Language Models (LLMs)

2025-01-13 21:00:00

构建损失函数实现对GPT模型的评估，实现预训练过程，文本生成的多样性策略以及加载开源的GPT2权重 Construct a loss function to evaluate the GPT model, implement the pre-training process, develop strategies for text generation diversity, and load open-source GPT-2 weights.

大语言模型LLM之构建GPT骨架 Building A GPT Framework of Large Language Models (LLMs)

2025-01-05 18:00:00

实现一个包含了多头注意力和前馈网络的Transformer块，并将其与token化与logits映射连接起来，实现可预测词功能 Implement a Transformer block that includes multi-head attention and feed-forward networks, and connect it with tokenization and logits mapping to enable word prediction functionality.

大语言模型LLM之自注意力机制 Self-Attention Mechanism in Large Language Models (LLMs)

2024-12-28 16:30:00

注意力机制是一种用于提高模型性能的技术，主要用于选择性地关注输入数据中的重要部分，从而增强模型对关键信息的理解能力。 The attention mechanism is a technique designed to improve model performance. It selectively focuses on important parts of the input data, enhancing the model's ability to understand and process key information.

大语言模型LLM之文本数据处理 Text Data Processing for Large Language Models (LLMs)

2024-12-22 15:30:00

有了LLM的基础之后，若要实现自己的LLM，首先需要对文本进行一定处理，将其转成计算机能理解的嵌入向量数据 After understanding the fundamentals of LLMs, if you want to build your own LLM, the first step is to process the text data and convert it into embedding vector data that can be understood by computers.