simcop2387/AutoGPTQ
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages 1 Wiki Activity

make compatible with older transformers version

Browse source
This commit is contained in:
PanQiWei 2023-05-15 13:26:18 +08:00
parent 262669112b
commit 07e06fa08c
1 changed files with 8 additions and 12 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

20
auto_gptq/modeling/llama.py
View file

@ -1,18 +1,14 @@
from logging import getLogger from logging import getLogger
from os.path import join, isfile
from typing import Optional, Union
import accelerate
import torch
import transformers
from transformers import AutoConfig, AutoModelForCausalLM
from ._const import *
from ._utils import *
from ._base import * from ._base import *
from ..nn_modules.fused_llama_attn import FusedLlamaAttentionForQuantizedModel from ..utils.import_utils import compare_transformers_version
from ..nn_modules.fused_llama_mlp import FusedLlamaMLPForQuantizedModel
if compare_transformers_version("v4.28.0", op="ge"):
from ..nn_modules.fused_llama_attn import FusedLlamaAttentionForQuantizedModel
from ..nn_modules.fused_llama_mlp import FusedLlamaMLPForQuantizedModel
else:
FusedLlamaAttentionForQuantizedModel = None
FusedLlamaMLPForQuantizedModel = None
logger = getLogger(__name__) logger = getLogger(__name__)