Convert padding structure and return sorted tensor with indices.

Convert Hydra configs and other special types to regular Python types.

Deprecated. Please use copy_to_local instead.

This function should return a class instance that delegates the calls to every cls in cls_dict

Args: model_name: override_config_kwargs: Returns:

name: name of the parameter param: training parameters infer_params (List[torch.Tensor]): a list of parameters all-gathered f

User passes in dp data. The data is dispatched to all tp/pp ranks with the same dp

treat pp as dp.

处理消息列表，将视频和图片转换为 base64 格式。 Returns: 处理后的消息列表（base64 格式）

The training loop of PPO. The driver process only need to call the compute functions of the worker group through RPC to construct the

Patches flash attention forward to handle 3D position ids in mrope. (3, batch_size, seq_length)

Format search results to OpenAI message format.

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

Gather the label from data. The value in label should be [0, vocab_size) Args: data: (..., vocab_size) label (torch.IntTensor) :

生成多模态消息内容。 1. 以 Item (整个视频或图片) 为单位进行 Top-K 筛选 2. 保留视频的所有帧 3. 视频内的所有帧共享该视频分到的总像素额度

Tracker file rescords the latest chckpoint during training to restart from.

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

获取文件类型图标

Return a local temp path that joins cache_dir and basename of hdfs_path Args: hdfs_path: cache_dir: Returns:

Get the number of gpus in this cluster.

Note that here we only asks execute_all and execute_rank_zero to be implemented Leave the choice of how these two functions handle argument '

Gets the position ids for Qwen2-VL, it should be generated before sharding the sequence. The batch dim has been removed and the input_ids sho

Get ulysses sequence parallel rank.

初始化 session state

Init resource pool and worker group

注入自定义 CSS 样式 - Kimi 风格明亮浅色主题

Log samples to swanlab as text

Log samples to wandb as a table

A memory efficient implementation of logprobs_from_logits