lowhipa-base-thchs30

This Whisper-for-IPA (WhIPA) model adapter is a PEFT LoRA fine-tuned version of openai/whisper-base on a subset (1k samples) of the Mandarin THCHS-30 database (https://arxiv.org/pdf/1512.01882) with IPA transcriptions by Taubert (2023, https://zenodo.org/records/7528596).

Model description

For deployment and description, please refer to https://github.com/jshrdt/whipa.

from transformers import WhisperForConditionalGeneration, WhisperTokenizer, WhisperProcessor
from peft import PeftModel

tokenizer = WhisperTokenizer.from_pretrained("openai/whisper-base", task="transcribe")
tokenizer.add_special_tokens({"additional_special_tokens": ["<|ip|>"] + tokenizer.all_special_tokens})

base_model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base")
base_model.generation_config.lang_to_id["<|ip|>"] = tokenizer.convert_tokens_to_ids(["<|ip|>"])[0]
base_model.resize_token_embeddings(len(tokenizer))

whipa_model = PeftModel.from_pretrained(base_model, "jshrdt/lowhipa-base-thchs30")

whipa_model.generation_config.language = "<|ip|>"
whipa_model.generation_config.task = "transcribe"

whipa_processor = WhisperProcessor.from_pretrained("openai/whisper-base", task="transcribe")

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: 0.001
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• lr_scheduler_warmup_steps: 100
• training_steps: 630
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.7877	2.0323	126	0.5588
0.3438	4.0645	252	0.3379
0.2765	6.0968	378	0.3056
0.2425	8.1290	504	0.2966
0.2195	10.1613	630	0.2911

Framework versions

• PEFT 0.15.1
• Transformers 4.48.3
• Pytorch 2.6.0+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0