Text Generation
Transformers
Safetensors
English
Japanese
llama
cybersecurity
conversational
text-generation-inference
Llama-Primus-Nemotron-70B-Instruct / mergekit_config.yml
Ubuntu
initial commit
a7558db
models:
- model: /mnt/shared/trendmicro/workspace/Llama-3.1-Nemotron-70B-Instruct-HF
parameters:
density: 0.63
weight: 0.4
epsilon: 0.1
- model: /mnt/shared/trendmicro/workspace/sft_weights/checkpoint-40
parameters:
density: 0.63
weight: 0.6
epsilon: 0.1
merge_method: della
base_model: /mnt/shared/trendmicro/workspace/Llama-3.1-70B
parameters:
epsilon: 0.1
int8_mask: true
dtype: bfloat16