mirror of
https://github.com/vllm-project/vllm.git
synced 2025-10-20 14:53:52 +08:00
[Bugfix] Fix Phi3.5 mini and MoE LoRA inference (#8571)
This commit is contained in:
@ -50,7 +50,7 @@ _GENERATION_MODELS = {
|
||||
"OrionForCausalLM": ("orion", "OrionForCausalLM"),
|
||||
"PersimmonForCausalLM": ("persimmon", "PersimmonForCausalLM"),
|
||||
"PhiForCausalLM": ("phi", "PhiForCausalLM"),
|
||||
"Phi3ForCausalLM": ("llama", "LlamaForCausalLM"),
|
||||
"Phi3ForCausalLM": ("phi3", "Phi3ForCausalLM"),
|
||||
"PhiMoEForCausalLM": ("phimoe", "PhiMoEForCausalLM"),
|
||||
"Qwen2ForCausalLM": ("qwen2", "Qwen2ForCausalLM"),
|
||||
"Qwen2MoeForCausalLM": ("qwen2_moe", "Qwen2MoeForCausalLM"),
|
||||
|
17
vllm/model_executor/models/phi3.py
Normal file
17
vllm/model_executor/models/phi3.py
Normal file
@ -0,0 +1,17 @@
|
||||
# coding=utf-8
|
||||
# Adapted from llama.py
|
||||
"""Inference-only Phi3 model code inherit from Llama.py"""
|
||||
|
||||
from vllm.model_executor.models.llama import LlamaForCausalLM
|
||||
|
||||
|
||||
class Phi3ForCausalLM(LlamaForCausalLM):
|
||||
|
||||
packed_modules_mapping = {
|
||||
"qkv_proj": [
|
||||
"qkv_proj",
|
||||
],
|
||||
"gate_up_proj": [
|
||||
"gate_up_proj",
|
||||
],
|
||||
}
|
@ -491,6 +491,10 @@ class PhiMoEForCausalLM(nn.Module, SupportsLoRA):
|
||||
"o_proj",
|
||||
"embed_tokens",
|
||||
"lm_head",
|
||||
"w1",
|
||||
"w2",
|
||||
"w3",
|
||||
"gate",
|
||||
]
|
||||
embedding_modules = {
|
||||
"embed_tokens": "input_embeddings",
|
||||
|
Reference in New Issue
Block a user