Vaibhavs10 / fast-whisper-finetuning

Vaibhavs10/fast-whisper-finetuning Issues

RuntimeError: expected mat1 and mat2 to have the same dtype, but got: c10::Half != float
Updated 2 months ago
ValueError: A custom logits processor of type <class 'transformers.generation.logits_process.ForceTokensLogitsProcessor'>....
Updated 3 months ago
Increase speed of data loading
Closed 4 months ago
Merge and Unload Peft weights to base model
Closed 10 months ago3
How to invoke compute_metrics ?
Updated 5 months ago
Bitsandbytes error (cuda setup error) google colab
Updated 6 months ago2
Word Error Rate Increasing post training on whisper-large-v3
Updated 6 months ago
Appling SpecAugment while fine tuning.
Updated 7 months ago
TypeError: prepare_model_for_kbit_training() got an unexpected keyword argument 'output_embedding_layer_name'
Closed 7 months ago4
Upcoming release - To-Do
Updated 9 months ago
After using the 200h Portuguese data finetune, the recognition has become much worse?
Updated a year ago
AttributeError: 'NoneType' object has no attribute 'cget_col_row_stats'
Closed a year ago2
Max_new_token
Closed a year ago1
TypeError: prepare_model_for_int8_training() got an unexpected keyword argument 'output_embedding_layer_name'
Closed a year ago4