convincing_LLM / README.md
librarian-bot's picture
Librarian Bot: Add base_model information to model
73639b2
|
raw
history blame
511 Bytes
metadata
library_name: peft
base_model: meta-llama/Llama-2-7b-chat-hf

Training procedure

The following bitsandbytes quantization config was used during training:

  • quant_method: bitsandbytes
  • load_in_8bit: False
  • load_in_4bit: True
  • 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: nf4
  • bnb_4bit_use_double_quant: False
  • bnb_4bit_compute_dtype: float16

Framework versions

  • PEFT 0.6.0.dev0