mpt.py

from .base import BaseAWQForCausalLM

class MptAWQForCausalLM(BaseAWQForCausalLM):
    layer_type = "MPTBlock"

    def get_model_layers(self, model):
        return model.transformer.blocks
    
    def get_layers_for_scaling(self, module, input_feat, module_kwargs):
        layers = []

        # attention input
        layers.append(dict(
            prev_op=module.norm_1,
            layers=[module.attn.Wqkv],
            inp=input_feat['attn.Wqkv'],
            module2inspect=module.attn,
            kwargs=module_kwargs
        ))

        # attention output
        layers.append(dict(
            prev_op=module.attn.Wqkv,
            layers=[module.attn.out_proj],
            inp=input_feat['attn.out_proj']
        ))

        # linear 1
        layers.append(dict(
            prev_op=module.norm_2,
            layers=[module.ffn.up_proj],
            inp=input_feat['ffn.up_proj'],
            module2inspect=module.ffn
        ))

        # linear 2
        layers.append(dict(
            prev_op=module.ffn.act,
            layers=[module.ffn.down_proj],
            inp=input_feat['ffn.down_proj']
        ))

        return layers
    
    def get_act_for_scaling(self, module):
        return dict(
            scale_name="ffn.act",
            scale_layer=module.ffn.act,
            scale_shape=module.ffn.up_proj.out_features
        )
    
    def move_embed(self, model, device):
        model.transformer.wte = model.transformer.wte.to(device)
        model.transformer.emb_drop = model.transformer.emb_drop.to(device)