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

[Core][Model] Terratorch backend integration (#23513)

Browse Source 
Signed-off-by: Michele Gazzetti <michele.gazzetti1@ibm.com>
Signed-off-by: Christian Pinto <christian.pinto@ibm.com>
Co-authored-by: Christian Pinto <christian.pinto@ibm.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
This commit is contained in:
mgazz
2025-09-04 08:22:41 +01:00
committed by GitHub
parent e7fc70016f
commit 51d5e9be7d
23 changed files with 305 additions and 208 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
vllm/config/__init__.py
View File

@@ -171,6 +171,7 @@ class ModelImpl(str, enum.Enum):
AUTO = "auto"
VLLM = "vllm"
TRANSFORMERS = "transformers"
TERRATORCH = "terratorch"
def get_attr_docs(cls: type[Any]) -> dict[str, str]:
@@ -496,7 +497,9 @@ class ModelConfig:
back to the Transformers implementation if no vLLM implementation is
available.\n
- "vllm" will use the vLLM model implementation.\n
- "transformers" will use the Transformers model implementation."""
- "transformers" will use the Transformers model implementation.\n
- "terratorch" will use the TerraTorch model implementation.
"""
override_attention_dtype: Optional[str] = None
"""Override dtype for attention"""
logits_processors: Optional[list[Union[str, type[LogitsProcessor]]]] = None