File size: 1,055 Bytes
3d37533 14d2dc2 f813c3d 3d37533 14d2dc2 3d37533 14d2dc2 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 |
---
language:
- hi
- en
license: cc-by-4.0
library_name: peft
datasets:
- iamshnoo/alpaca-cleaned-hindi
base_model: meta-llama/Llama-2-13b-hf
---
This represents the PEFT weights only. The base model is LLaMA 2. Instruction finetuning was done using 4 bit QLoRA on a single A100 GPU with the PEFT config as given below. The dataset used for this instruction finetuning process is a translated version of the cleaned alpaca dataset (translated using NLLB-1.3B).
Do note that this model might have inferior performance on some language specific tasks compared to full finetuning or a different base model trained with more language specific data.
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- 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: True
- bnb_4bit_compute_dtype: bfloat16
### Framework versions
- PEFT 0.4.0 |