HIST model missing explicit pretrain toggle in initialization

[WindBT47]

In contrib\model\pytorch_hist.py
Description: Currently, the HIST model implementation in qlib lacks an explicit parameter to control the pre-training loading logic. This makes it difficult for users to skip the pre-training weight-loading phase even when a model path is specified or when they wish to train the architecture from scratch.

Proposed Changes:

init method: Added pretrain (defaulting to True for backward compatibility) to the parameters.

Logic Update: Wrapped the weight-loading and state_dict update logic inside if self.pretrain:.

# In __init__
def __init__(self, ..., pretrain=True, **kwargs):
    ...
    self.pretrain = pretrain

# In weight loading logic
if self.pretrain:
    if self.model_path is not None:
        self.logger.info("Loading pretrained model...")
        pretrained_model.load_state_dict(torch.load(self.model_path))
    
    # ... (state_dict update logic)
    self.logger.info("Loading pretrained model Done...")