biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Fix Mistral model (#1220)

Browse Source
This commit is contained in:
Woosuk Kwon
2023-09-28 10:44:05 -07:00
committed by GitHub
parent bb1ba58f06
commit a8e98aee0c
4 changed files with 27 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/model_executor/models/mistral.py
View File

@@ -29,7 +29,6 @@ from typing import List, Optional, Tuple
import torch
from torch import nn
from vllm.transformers_utils.configs.mistral import MistralConfig
from vllm.model_executor.input_metadata import InputMetadata
from vllm.model_executor.layers.activation import SiluAndMul
@@ -46,6 +45,7 @@ from vllm.model_executor.weight_utils import (
convert_pyslice_to_tensor, hf_model_weights_iterator,
load_tensor_parallel_weights, load_padded_tensor_parallel_vocab)
from vllm.sequence import SamplerOutput
from vllm.transformers_utils.configs.mistral import MistralConfig
KVCache = Tuple[torch.Tensor, torch.Tensor]