File size: 477 Bytes
d1308f2 | 1 2 3 4 5 6 7 8 9 10 11 12 13 | default_stage:
default_modifiers:
GPTQModifier:
targets: [Linear]
ignore: [lm_head, 're:.*embed_tokens.*', 're:model\.layers\.78\..*', 're:.*\.mlp\.gate$',
're:.*self_attn.*', 're:.*indexer.*', 're:.*shared_experts.*', 're:^model\.layers\.[0-2]\.mlp\.(gate_proj|up_proj|down_proj)$']
scheme: W4A16_ASYM
bypass_divisibility_checks: false
block_size: 128
dampening_frac: 0.01
actorder: static
offload_hessians: false
|