Edit model card

hubert-base-common-language

This model is a fine-tuned version of facebook/hubert-base-ls960 on the common_language dataset. It achieves the following results on the evaluation set:

  • Loss: 1.3477
  • Accuracy: 0.7317

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 0.0001
  • train_batch_size: 1
  • eval_batch_size: 4
  • seed: 0
  • distributed_type: IPU
  • gradient_accumulation_steps: 32
  • total_train_batch_size: 128
  • total_eval_batch_size: 64
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.25
  • num_epochs: 10.0
  • training precision: Mixed Precision

Training results

Framework versions

  • Transformers 4.18.0.dev0
  • Pytorch 1.10.0+cpu
  • Datasets 2.0.0
  • Tokenizers 0.11.6
Downloads last month
24
Safetensors
Model size
94.6M params
Tensor type
FP16
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Dataset used to train Graphcore/hubert-base-common-language