mirror of
https://github.com/mudler/LocalAI.git
synced 2025-06-12 20:18:11 +00:00
Bump vLLM version + more options when loading models in vLLM (#1782)
* Bump vLLM version to 0.3.2 * Add vLLM model loading options * Remove transformers-exllama * Fix install exllama
This commit is contained in:
@ -126,6 +126,11 @@ message ModelOptions {
|
||||
|
||||
// vllm
|
||||
string Quantization = 40;
|
||||
float GPUMemoryUtilization = 50;
|
||||
bool TrustRemoteCode = 51;
|
||||
bool EnforceEager = 52;
|
||||
int32 SwapSpace = 53;
|
||||
int32 MaxModelLen = 54;
|
||||
|
||||
string MMProj = 41;
|
||||
|
||||
|
Reference in New Issue
Block a user