Plim/xls-r-300m-fr

---

Model description

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the MOZILLA-FOUNDATION/COMMON_VOICE_7_0 - FR dataset.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 7.5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2000
• num_epochs: 2.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.495	0.16	500	3.3883	1.0
2.9095	0.32	1000	2.9152	1.0000
1.8434	0.49	1500	1.0473	0.7446
1.4298	0.65	2000	0.5729	0.5130
1.1937	0.81	2500	0.3795	0.3450
1.1248	0.97	3000	0.3321	0.3052
1.0835	1.13	3500	0.3038	0.2805
1.0479	1.3	4000	0.2910	0.2689
1.0413	1.46	4500	0.2798	0.2593
1.014	1.62	5000	0.2727	0.2512
1.004	1.78	5500	0.2646	0.2471
0.9949	1.94	6000	0.2619	0.2457

It achieves the best result on STEP 6000 on the validation set:

• Loss: 0.2619
• Wer: 0.2457

Framework versions

• Transformers 4.17.0.dev0
• Pytorch 1.10.2+cu102
• Datasets 1.18.2.dev0
• Tokenizers 0.11.0

Evaluation Commands

1. To evaluate on mozilla-foundation/common_voice_7 with split test

python eval.py --model_id Plim/xls-r-300m-fr --dataset mozilla-foundation/common_voice_7_0 --config fr --split test

2. To evaluate on speech-recognition-community-v2/dev_data

python eval.py --model_id Plim/xls-r-300m-fr --dataset speech-recognition-community-v2/dev_data --config fr --split validation --chunk_length_s 5.0 --stride_length_s 1.0