使用jittor加载模型的时候这个文件pytorch_model.bin.index.json是如何生成的？

def load_from_torch_shard_ckpt(model, ckpt_dir):
"""
Load sharded checkpoints directly from huggingface dir.
"""
with open(os.path.join(ckpt_dir, 'pytorch_model.bin.index.json')) as fp:
ckpt_index = json.load(fp)

total_size = ckpt_index['metadata']['total_size']
weight_map = ckpt_index['weight_map']

file_weight_map = {}
for key, value in weight_map.items():
    # key: param name; value: filename.
    if value not in file_weight_map:
        file_weight_map[value] = []
    file_weight_map[value].append(key)

load_from_map(model, ckpt_dir, file_weight_map)

Start with the load_from_torch_shard_ckpt function shown in the issue and trace how ckpt_dir is prepared before it is called. Determine which process creates pytorch_model.bin.index.json and how its metadata and weight_map relate to the shard files. Done means documenting that generation path clearly, including the relevant command or loading workflow if the repository contains one.