File size: 288 Bytes
3bf916d |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- load_in_8bit: True
- llm_int8_threshold: 6.0
- llm_int8_skip_modules: None
- llm_int8_enable_fp32_cpu_offload: False
### Framework versions
- PEFT 0.4.0.dev0
|