hello can you share this model's bpe file in order to finetune this model?
i encounter some error when i try to finetune this model.
this the script and error info:
./pruned_transducer_stateless7/finetune.py \
--world-size 1
--num-epochs 20
--start-epoch 1
--exp-dir pruned_transducer_stateless7/exp_ali_giga_finetune
--subset S
--use-fp16 1
--base-lr 0.005
--lr-epochs 100
--lr-batches 100000
--do-finetune True
--use-mux False
--finetune-ckpt ./icefall-asr-alimeeting-pruned-transducer-stateless7/exp/pretrained.pt
--max 500
2024-09-02 19:12:29,787 INFO [finetune.py:1056] Training started
2024-09-02 19:12:29,789 INFO [finetune.py:1066] Device: cuda:0
2024-09-02 19:12:29,792 INFO [finetune.py:1075] {'frame_shift_ms': 10.0, 'allowed_excess_duration_ratio': 0.1, 'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 50, 'reset_interval': 200, 'valid_interval': 3000, 'feature_dim': 80, 'subsampling_factor': 4, 'warm_step': 2000, 'env_info': {'k2-version': '1.24.4', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': 'ff1d435a8d3c4eaa15828a84a7240678a70539a7', 'k2-git-date': 'Fri Feb 23 01:48:38 2024', 'lhotse-version': '1.27.0.dev+git.66b95ba.clean', 'torch-version': '1.12.1+cu113', 'torch-cuda-available': True, 'torch-cuda-version': '11.3', 'python-version': '3.10', 'icefall-git-branch': 'master', 'icefall-git-sha1': '59529722-dirty', 'icefall-git-date': 'Sat Aug 17 13:24:38 2024', 'icefall-path': '/home/project/icefall/icefall', 'k2-path': '/root/miniconda3/envs/icefall/lib/python3.10/site-packages/k2/init.py', 'lhotse-path': '/root/miniconda3/envs/icefall/lib/python3.10/site-packages/lhotse/init.py', 'hostname': 'ubuntuai', 'IP address': '127.0.1.1'}, 'world_size': 1, 'master_port': 12354, 'tensorboard': True, 'num_epochs': 20, 'start_epoch': 1, 'start_batch': 0, 'exp_dir': PosixPath('pruned_transducer_stateless7/exp_ali_giga_finetune'), 'bpe_model': 'data/lang_bpe_500/bpe.model', 'base_lr': 0.005, 'lr_batches': 100000.0, 'lr_epochs': 100.0, 'context_size': 2, 'prune_range': 5, 'lm_scale': 0.25, 'am_scale': 0.0, 'simple_loss_scale': 0.5, 'seed': 42, 'print_diagnostics': False, 'inf_check': False, 'save_every_n': 2000, 'keep_last_k': 30, 'average_period': 200, 'use_fp16': True, 'num_encoder_layers': '2,4,3,2,4', 'feedforward_dims': '1024,1024,2048,2048,1024', 'nhead': '8,8,8,8,8', 'encoder_dims': '384,384,384,384,384', 'attention_dims': '192,192,192,192,192', 'encoder_unmasked_dims': '256,256,256,256,256', 'zipformer_downsampling_factors': '1,2,4,8,2', 'cnn_module_kernels': '31,31,31,31,31', 'decoder_dim': 512, 'joiner_dim': 512, 'do_finetune': True, 'use_mux': False, 'init_modules': None, 'finetune_ckpt': './icefall-asr-alimeeting-pruned-transducer-stateless7/exp/pretrained.pt', 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 500, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'subset': 'S', 'small_dev': False, 'blank_id': 0, 'vocab_size': 500}
2024-09-02 19:12:29,792 INFO [finetune.py:1077] About to create model
2024-09-02 19:12:30,302 INFO [zipformer.py:178] At encoder stack 4, which has downsampling_factor=2, we will combine the outputs of layers 1 and 3, with downsampling_factors=2 and 8.
2024-09-02 19:12:30,314 INFO [finetune.py:1081] Number of model parameters: 70369391
2024-09-02 19:12:30,823 INFO [finetune.py:637] Loading checkpoint from ./icefall-asr-alimeeting-pruned-transducer-stateless7/exp/pretrained.pt
before strict
start strict
RuntimeError=============
Traceback (most recent call last):
File "/home/project/icefall/icefall/egs/librispeech/ASR/./pruned_transducer_stateless7/finetune.py", line 1368, in
main()
File "/home/project/icefall/icefall/egs/librispeech/ASR/./pruned_transducer_stateless7/finetune.py", line 1361, in main
run(rank=0, world_size=1, args=args)
File "/home/project/icefall/icefall/egs/librispeech/ASR/./pruned_transducer_stateless7/finetune.py", line 1092, in run
checkpoints = load_model_params(
File "/home/project/icefall/icefall/egs/librispeech/ASR/./pruned_transducer_stateless7/finetune.py", line 653, in load_model_params
model.load_state_dict(checkpoint["model"], strict=strict)
File "/root/miniconda3/envs/icefall/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1608, in load_state_dict
raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for Transducer:
size mismatch for decoder.embedding.weight: copying a param with shape torch.Size([3290, 512]) from checkpoint, the shape in current model is torch.Size([500, 512]).
size mismatch for joiner.output_linear.weight: copying a param with shape torch.Size([3290, 512]) from checkpoint, the shape in current model is torch.Size([500, 512]).
size mismatch for joiner.output_linear.bias: copying a param with shape torch.Size([3290]) from checkpoint, the shape in current model is torch.Size([500]).
size mismatch for simple_am_proj.weight: copying a param with shape torch.Size([3290, 384]) from checkpoint, the shape in current model is torch.Size([500, 384]).
size mismatch for simple_am_proj.bias: copying a param with shape torch.Size([3290]) from checkpoint, the shape in current model is torch.Size([500]).
size mismatch for simple_lm_proj.weight: copying a param with shape torch.Size([3290, 512]) from checkpoint, the shape in current model is torch.Size([500, 512]).
size mismatch for simple_lm_proj.bias: copying a param with shape torch.Size([3290]) from checkpoint, the shape in current model is torch.Size([500]).