python dataloaders/save_train_dataset.py 数据预处理出现错误

[Johnren1111]

(semtalk) johnren@johnren-BATTLE-AX-Z790M-PLUS-D5:~/Desktop/SemTalk$ python dataloaders/save_train_dataset.py
funasr version: 1.3.0.
WARNING:root:trust_remote_code: False
Warning, miss key in ckpt: modality_encoders.AUDIO.decoder.blocks.0.0.weight, /home/johnren/Desktop/SemTalk/weights/emotion2vec_plus_large/model.pt
Warning, miss key in ckpt: modality_encoders.AUDIO.decoder.blocks.0.0.bias, /home/johnren/Desktop/SemTalk/weights/emotion2vec_plus_large/model.pt
Warning, miss key in ckpt: modality_encoders.AUDIO.decoder.blocks.1.0.weight, /home/johnren/Desktop/SemTalk/weights/emotion2vec_plus_large/model.pt
Warning, miss key in ckpt: modality_encoders.AUDIO.decoder.blocks.1.0.bias, /home/johnren/Desktop/SemTalk/weights/emotion2vec_plus_large/model.pt
Warning, miss key in ckpt: modality_encoders.AUDIO.decoder.blocks.2.0.weight, /home/johnren/Desktop/SemTalk/weights/emotion2vec_plus_large/model.pt
Warning, miss key in ckpt: modality_encoders.AUDIO.decoder.blocks.2.0.bias, /home/johnren/Desktop/SemTalk/weights/emotion2vec_plus_large/model.pt
Warning, miss key in ckpt: modality_encoders.AUDIO.decoder.blocks.3.0.weight, /home/johnren/Desktop/SemTalk/weights/emotion2vec_plus_large/model.pt
Warning, miss key in ckpt: modality_encoders.AUDIO.decoder.blocks.3.0.bias, /home/johnren/Desktop/SemTalk/weights/emotion2vec_plus_large/model.pt
Warning, miss key in ckpt: modality_encoders.AUDIO.decoder.proj.weight, /home/johnren/Desktop/SemTalk/weights/emotion2vec_plus_large/model.pt
Warning, miss key in ckpt: modality_encoders.AUDIO.decoder.proj.bias, /home/johnren/Desktop/SemTalk/weights/emotion2vec_plus_large/model.pt
Loading the Wav2Vec2 Processor...
Loading the HuBERT Model...
Some weights of the model checkpoint at ./facebook/hubert-large-ls960-ft were not used when initializing HubertModel: ['lm_head.weight', 'lm_head.bias']

• This IS expected if you are initializing HubertModel from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).
• This IS NOT expected if you are initializing HubertModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).
  2026-03-15 22:19:26.024 | INFO | dataloaders.beat2_add_hubert:build_cache:354 - Audio bit rate: 16000
  2026-03-15 22:19:26.024 | INFO | dataloaders.beat2_add_hubert:build_cache:355 - Reading data './BEAT2/beat_english_v2.0.0/'...
  2026-03-15 22:19:26.024 | INFO | dataloaders.beat2_add_hubert:build_cache:356 - Creating the dataset cache...
  2026-03-15 22:19:26.024 | INFO | dataloaders.beat2_add_hubert:build_cache:361 - Found the cache ./datasets/beat2_cache_2_eval/train/smplxflame_30_cache
  Successfully loaded self-pretrained checkpoints for vq_face
  Successfully loaded self-pretrained checkpoints for vq_upper
  Successfully loaded self-pretrained checkpoints for vq_hands
  Successfully loaded self-pretrained checkpoints for vq_lower
  Processing: 0it [00:00, ?it/s]s]
  Writing LMDB: 0it [00:00, ?it/s]
  Done. Wrote 0 samples to ./datasets/beat2_semtalk_train_eval/

Verifying written LMDB by reading one sample ...
Traceback (most recent call last):
File "dataloaders/save_train_dataset.py", line 441, in
main()
File "dataloaders/save_train_dataset.py", line 405, in main
test_function_data = test_ds[0] # <- 你要的变量名
File "dataloaders/save_train_dataset.py", line 431, in getitem
raise IndexError(f"Key {key!r} not found")
IndexError: Key b'0000000000' not found

[Johnren1111]

数据都已经下载好了，是这个数据预处理部分数据加载中出现的数据长度为0，麻烦您解答这个问题！谢谢～