xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-03-25 17:21:24 +08:00
Code Issues Packages Projects Releases Wiki Activity

remove dtype_original_type

Browse Source 
Signed-off-by: Xingyu Liu <charlotteliu12x@gmail.com>
This commit is contained in:
Xingyu Liu 2025-12-05 10:17:49 -08:00
parent 9b19e3b94f
commit 0a4f4724ef
2 changed files with 17 additions and 42 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
tests/config/model_arch_groundtruth.json
View File

@ -14,8 +14,7 @@
"num_experts": 0,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"mosaicml/mpt-7b": {
"architectures": [
@ -32,8 +31,7 @@
"num_experts": 0,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"databricks/dbrx-instruct": {
"architectures": [
@ -50,8 +48,7 @@
"num_experts": 0,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"tiiuae/falcon-7b": {
"architectures": [
@ -68,8 +65,7 @@
"num_experts": 0,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"tiiuae/falcon-40b": {
"architectures": [
@ -86,8 +82,7 @@
"num_experts": 0,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"luccafong/deepseek_mtp_main_random": {
"architectures": [
@ -104,8 +99,7 @@
"num_experts": 72,
"is_deepseek_mla": true,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"luccafong/deepseek_mtp_draft_random": {
"architectures": [
@ -122,8 +116,7 @@
"num_experts": 72,
"is_deepseek_mla": true,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"Qwen/Qwen3-Next-80B-A3B-Instruct": {
"architectures": [
@ -140,8 +133,7 @@
"num_experts": 512,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"tiny-random/qwen3-next-moe": {
"architectures": [
@ -158,8 +150,7 @@
"num_experts": 32,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"zai-org/GLM-4.5": {
"architectures": [
@ -176,8 +167,7 @@
"num_experts": 160,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"baidu/ERNIE-4.5-21B-A3B-PT": {
"architectures": [
@ -194,8 +184,7 @@
"num_experts": 64,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"lmsys/gpt-oss-20b-bf16": {
"architectures": [
@ -212,8 +201,7 @@
"num_experts": 32,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"deepseek-ai/DeepSeek-V3.2-Exp": {
"architectures": [
@ -230,8 +218,7 @@
"num_experts": 256,
"is_deepseek_mla": true,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"meta-llama/Llama-4-Scout-17B-16E-Instruct": {
"architectures": [
@ -248,8 +235,7 @@
"num_experts": 16,
"is_deepseek_mla": false,
"is_multimodal_model": true,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"nvidia/Llama-3_3-Nemotron-Super-49B-v1": {
"architectures": [
@ -266,8 +252,7 @@
"num_experts": 0,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"XiaomiMiMo/MiMo-7B-RL": {
"architectures": [
@ -284,8 +269,7 @@
"num_experts": 0,
"is_deepseek_mla": false,
"is_multimodal_model": false,
"dtype": "torch.bfloat16",
"dtype_original_type": "torch.dtype"
"dtype": "torch.bfloat16"
},
"meituan-longcat/LongCat-Flash-Chat": {
"architectures": [
@ -302,7 +286,6 @@
"num_experts": 512,
"is_deepseek_mla": true,
"is_multimodal_model": false,
"dtype": "torch.float32",
"dtype_original_type": "torch.dtype"
"dtype": "torch.float32"
}
}

8
tests/config/test_model_arch_config.py
View File

@ -3,8 +3,6 @@
import json
from pathlib import Path
import torch
from vllm.config import ModelConfig
@ -67,12 +65,6 @@ def test_model_arch_config():
dtype = model_arch_config.torch_dtype
assert str(dtype) == expected["dtype"]
if expected["dtype_original_type"] == "str":
assert isinstance(dtype, str)
elif expected["dtype_original_type"] == "torch.dtype":
assert isinstance(dtype, torch.dtype)
else:
raise ValueError(f"Unknown dtype_original_type: {expected['dtype']}")
# Test that model_config methods return expected values
assert model_config.architectures == expected["architectures"]