logical-reasoning / llama-factory /config /qwen2_72b_lora_sft_4bit-p1.yaml

Commit History

updated logging/save/eval steps
dc1ac39

dh-mc commited on

Update qwen2_72b_lora_sft_4bit-p1.yaml
c817aef

dh-mc commited on

reduce b-->4
5f306e7

dh-mc commited on

reduce batch size to 8
103dbc8

dh-mc commited on

reduce batch size to 16
af72c0c

dh-mc commited on

fixed templates
685f9c5

dh-mc commited on

fixed dataset
25508d8

dh-mc commited on

upcast_layernorm: true for 4bit qlora
de6246f

dh-mc commited on

tune-mgtv-qwen2_72b.sh
09e6035

dh-mc commited on