使用Selenium下载PDF文件并保存到指定路径 参数: url: PDF文件的直接URL save_path: 保存文件的目录路径 filename: 保存后的文件名（包含.pdf扩展名） 返回: file_path: 最终保存的文件路径

Execute the tool to fetch stock information. Args: stock_code: The stock code to query Returns: Sto

执行并发任务 :param task_list: 任务列表，每个元素是元组(func, args, kwargs) func: 要执行的函数 args: 位置参数

从output中提取思维链和答案

Extract numerical or text answer from a response. Args: text: The text response to analyze

extract output between the last ```output\n...\n```

r"""Extract tool message.

获取股票列表的资金流向数据（按主力净流入排序） 参数: page_size: 每页显示数量，默认50 page_num: 页码，默认第1页 max_retries: 最大重试次数 retry_delay: 重试延迟时

r"""Find the modules in the expanded blocks to apply lora.

查找特定股票的所有报告

Args: input_ids: input ids. shape (batch_size, seq_length) attention_mask: attention_mask. shape (batch_size, seq_le

Args: input_ids: input ids. shape (batch_size, seq_length) attention_mask: attention_mask. shape (batch_size, seq_le

r"""Run forward pass and computes the log probabilities.

We assume: - The model takes input: (input_ids, attention_mask, position_ids). No rmpad for the input - The communication sha

获取中国国家统计局，所有统计分类

获取公司的所有三大财务报表 Args: stock_code (str): 股票代码 market (str): 股票市场，"HK"为港股，"A"为A股 period (str): 报告期间，可选"年度"或"中期"（仅港股适

Get all the parameters of the models in all pp ranks Returns: params: List[List[Dict[str, Tensor]]]: a list of parameters in all

获取公司的资产负债表 Args: stock_code (str): 股票代码 market (str): 股票市场，"HK"为港股，"A"为A股 period (str): 报告期间，可选"年度"或"中期"（仅港股适用），

r"""Compute the log probabilities of the given labels under the given logits. Returns: logps: A tensor of shape (batch_size,) containing

生成缓存键

获取公司的现金流量表 Args: stock_code (str): 股票代码 market (str): 股票市场，"HK"为港股，"A"为A股 period (str): 报告期间，可选"年度"或"中期"（仅港股适用），

Determine the directory name for this rank's checkpoint.

Create a schedule with a learning rate that decreases following the values of the cosine function between the initial lr set in the optimizer

获取恒生指数列表

r"""Convert dataset or dataset dict to dataset module.

获取公司的利润表 Args: stock_code (str): 股票代码 market (str): 股票市场，"HK"为港股，"A"为A股 period (str): 报告期间，可选"年度"或"中期"（仅港股适用），默认为

r"""Return the ollama modelfile. TODO: support function calling.

获取所有问题文本。 返回: 问题文本列表

获取上交所企业报告，code：股票号码；start：开始日期（默认365天前）；end：结束日期（默认今天）；type：报告类型，年报/业绩预测；title：报告标题

获取上交所企业报告，code：股票号码；start：开始日期（默认365天前）；end：结束日期（默认今天）；type：报告类型，年报/业绩预测；title：报告标题

get order of seq lengths to make partitions balanced, this is used in balacing sum of seqlength across dp ranks and microbatches Paramete

r"""Get the sequnce lengths in the current batch. e.g. ```python # input [ [1, 1, 2, 2, 2, 0], [1, 2, 2, 3, 3, 3],

Return a dictionary containing name to a shape and dtype.

The answer will be considered correct if: (a) it normalizes to the same string as the ground truth answer OR (b) sympy can simplify t

r"""Infer the optimal dtype according to the model_dtype and device compatibility.

加载AFAC数据集

r"""Load the training configuration from config path.

加载数据文件。 返回: 加载的DataFrame对象

load all model params to cuda

Load MCP server configuration from JSON file

Load MCP server configuration from JSON file