Errors about loading a pretrained model

[wuliting-wlt]

I meet two problems

1. I follow the words on readme.md as follows to load a pre-trained model:

   import fairseq
   import hubert_pretraining, hubert
   ckpt_path = "/../../large_lrs3_iter5.pt"
   models, cfg, task = fairseq.checkpoint_utils.load_model_ensemble_and_task([ckpt_path])
   At this point, an error occurred:
   omegaconf.errors.ConfigAttributeError: Key 'required_seq_len_multiple' not in 'AVHubertConfig'
   full_key: required_seq_len_multiple
   object_type=AVHubertConfig

2. I want to Finetune an AV-HuBERT model with Seq2Seq ,follow the readme.md:
   $ cd avhubert
   $ fairseq-hydra-train --config-dir /path/to/conf/finetune/ --config-name large_lrs3_433h.yaml
   task.data=/path/to/data task.label_dir=/path/to/label
   task.tokenizer_bpe_model=/path/to/tokenizer model.w2v_path=/path/to/checkpoint/large_lrs3_iter5.pt
   hydra.run.dir=/path/to/experiment/finetune/ common.user_dir=pwd
   The error is as follows:
   omegaconf.errors.ConfigKeyError: Key 'input_modality' not in 'AVHubertPretrainingConfig'
   full_key: input_modality
   object_type=AVHubertPretrainingConfig

I don't know what caused these errors