dhmeltzer/Llama-2-13b-hf-ds_wiki_1024_full_r_64_alpha_16

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: True
• bnb_4bit_compute_dtype: bfloat16

Framework versions

• PEFT 0.4.0

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	46.47
ARC (25-shot)	59.04
HellaSwag (10-shot)	82.33
MMLU (5-shot)	55.36
TruthfulQA (0-shot)	35.75
Winogrande (5-shot)	76.32
GSM8K (5-shot)	10.01
DROP (3-shot)	6.48