README.md

metadata

library_name: peft
base_model: openai/whisper-large-v2

Training procedure

The following bitsandbytes quantization config was used during training:

• quant_method: bitsandbytes
• load_in_8bit: True
• load_in_4bit: False
• llm_int8_threshold: 6.0
• llm_int8_skip_modules: None
• llm_int8_enable_fp32_cpu_offload: False
• llm_int8_has_fp16_weight: False
• bnb_4bit_quant_type: fp4
• bnb_4bit_use_double_quant: False
• bnb_4bit_compute_dtype: float32

WER: 'eval/wer': 18.944399677679293, 'eval/normalized_wer': 13.092652542720156

The following bitsandbytes quantization config was used during training:

• load_in_8bit: True
• load_in_4bit: False
• llm_int8_threshold: 6.0
• llm_int8_skip_modules: None
• llm_int8_enable_fp32_cpu_offload: False
• llm_int8_has_fp16_weight: False
• bnb_4bit_quant_type: fp4
• bnb_4bit_use_double_quant: False
• bnb_4bit_compute_dtype: float32

Framework versions

• PEFT 0.5.0

• PEFT 0.5.0.dev0 bitsandbytes quantization config was used during training:

• load_in_8bit: True

• load_in_4bit: False

• llm_int8_threshold: 6.0

• llm_int8_skip_modules: None

• llm_int8_enable_fp32_cpu_offload: False

• llm_int8_has_fp16_weight: False

• bnb_4bit_quant_type: fp4

• bnb_4bit_use_double_quant: False

• bnb_4bit_compute_dtype: float32

Framework versions

• PEFT 0.5.0

• PEFT 0.5.0.dev0