espnet/meld_cls1_wavlm_base_plus

ESPnet2 CLS model

espnet/meld_cls1_wavlm_base_plus

This model was trained by itoten using meld recipe in espnet.

Demo: How to use in ESPnet2

Follow the ESPnet installation instructions if you haven't done that already.

cd espnet
git checkout b55d7896fb1d35f40fa9d21718953750071af8dc
pip install -e .
cd egs2/meld/cls1
./run.sh --skip_data_prep false --skip_train true --download_model espnet/meld_cls1_wavlm_base_plus

RESULTS

Environments

• date: Sat Aug 22 16:08:25 JST 2026
• python version: 3.10.14 (tags/v3.10.14-25-ge98930d7387-dirty:e98930d7387, May 24 2024, 23:30:09) [GCC 13.2.0]
• espnet2 version: espnet2 202604
• pytorch version: pytorch 2.11.0+cu130
• Git hash: 27b313e60734ae428b6d25482a423d7913df4770
  • Commit date: Thu Aug 13 11:36:45 2026 +0900

cls_20260822.155629

Split	mean_acc	mAP	mean_auc	n_labels	n_instances
cls_test	50.81	28.24	70.31	7.00	2608.00
cls_valid	48.19	30.05	69.63	7.00	1104.00

CLS config

expand

config: conf/train_cls_wavlm_transformer.yaml
print_config: false
log_level: INFO
drop_last_iter: false
dry_run: false
iterator_type: sequence
valid_iterator_type: null
output_dir: ./exp/cls_20260822.155629
ngpu: 1
seed: 0
num_workers: 1
num_att_plot: 0
dist_backend: nccl
dist_init_method: env://
dist_world_size: null
dist_rank: null
local_rank: 0
dist_master_addr: null
dist_master_port: null
dist_launcher: null
multiprocessing_distributed: false
unused_parameters: false
sharded_ddp: false
use_deepspeed: false
deepspeed_config: null
gradient_as_bucket_view: true
ddp_comm_hook: null
cudnn_enabled: true
cudnn_benchmark: false
cudnn_deterministic: true
use_tf32: false
collect_stats: false
write_collected_feats: false
max_epoch: 30
patience: 5
val_scheduler_criterion:
- valid
- loss
early_stopping_criterion:
- valid
- loss
- min
best_model_criterion:
-   - valid
    - acc
    - max
keep_nbest_models: 1
nbest_averaging_interval: 0
grad_clip: 5.0
grad_clip_type: 2.0
grad_noise: false
accum_grad: 1
no_forward_run: false
resume: true
train_dtype: float32
use_amp: false
log_interval: null
use_matplotlib: true
use_tensorboard: true
create_graph_in_tensorboard: false
use_wandb: false
wandb_project: null
wandb_id: null
wandb_entity: null
wandb_name: null
wandb_model_log_interval: -1
wandb_allow_val_change: true
detect_anomaly: false
use_adapter: false
adapter: lora
save_strategy: all
adapter_conf: {}
pretrain_path: null
init_param: []
ignore_init_mismatch: false
freeze_param:
- frontend.upstream
num_iters_per_epoch: null
batch_size: 32
valid_batch_size: null
batch_bins: 1000000
valid_batch_bins: null
category_sample_size: 10
upsampling_factor: 0.5
category_upsampling_factor: 0.5
dataset_upsampling_factor: 0.5
dataset_scaling_factor: 1.2
max_batch_size: null
min_batch_size: 1
train_shape_file:
- ./exp/cls_stats_16k/train/speech_shape
- ./exp/cls_stats_16k/train/label_shape
valid_shape_file:
- ./exp/cls_stats_16k/valid/speech_shape
- ./exp/cls_stats_16k/valid/label_shape
batch_type: folded
valid_batch_type: null
fold_length:
- 160000
- 2
sort_in_batch: descending
shuffle_within_batch: false
sort_batch: descending
multiple_iterator: false
chunk_length: 500
chunk_shift_ratio: 0.5
num_cache_chunks: 1024
chunk_excluded_key_prefixes: []
chunk_default_fs: null
chunk_max_abs_length: null
chunk_discard_short_samples: true
train_data_path_and_name_and_type:
-   - ./dump/train/wav.scp
    - speech
    - sound
-   - ./dump/train/text
    - label
    - text
valid_data_path_and_name_and_type:
-   - ./dump/valid/wav.scp
    - speech
    - sound
-   - ./dump/valid/text
    - label
    - text
multi_task_dataset: false
allow_variable_data_keys: false
max_cache_size: 0.0
max_cache_fd: 32
allow_multi_rates: false
valid_max_cache_size: null
exclude_weight_decay: false
exclude_weight_decay_conf: {}
optim: adam
optim_conf:
    lr: 0.001
scheduler: warmuplr
scheduler_conf:
    warmup_steps: 3180
token_list:
- neutral
- joy
- surprise
- anger
- sadness
- disgust
- fear
- <unk>
token_type: word
init: null
input_size: null
use_preprocessor: true
frontend: s3prl
frontend_conf:
    frontend_conf:
        upstream: wavlm_base_plus
    download_dir: ./hub
    multilayer_feature: true
    fs: 16k
specaug: null
specaug_conf: {}
normalize: utterance_mvn
normalize_conf: {}
preencoder: null
preencoder_conf: {}
encoder: transformer
encoder_conf:
    output_size: 128
    attention_heads: 4
    linear_units: 1024
    num_blocks: 4
    dropout_rate: 0.4
    input_layer: linear
decoder: linear
decoder_conf: {}
model: espnet
model_conf:
    classification_type: multi-class
    log_epoch_metrics: true
required:
- output_dir
- token_list
version: '202604'
distributed: false

Citing ESPnet

@inproceedings{watanabe2018espnet,
  author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
  title={{ESPnet}: End-to-End Speech Processing Toolkit},
  year={2018},
  booktitle={Proceedings of Interspeech},
  pages={2207--2211},
  doi={10.21437/Interspeech.2018-1456},
  url={http://dx.doi.org/10.21437/Interspeech.2018-1456}
}

or arXiv:

@misc{watanabe2018espnet,
  title={ESPnet: End-to-End Speech Processing Toolkit},
  author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
  year={2018},
  eprint={1804.00015},
  archivePrefix={arXiv},
  primaryClass={cs.CL}
}