Skip to content

Commit

Permalink
OAI: Fix model parameter placement
Browse files Browse the repository at this point in the history
Accidentally edited the Model Card parameters vs the model load request
ones.

Signed-off-by: kingbri <bdashore3@proton.me>
  • Loading branch information
bdashore3 committed Dec 17, 2023
1 parent 1d0bdfa commit 70fbee3
Showing 1 changed file with 2 additions and 2 deletions.
4 changes: 2 additions & 2 deletions OAI/types/model.py
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@
class ModelCardParameters(BaseModel):
max_seq_len: Optional[int] = 4096
rope_scale: Optional[float] = 1.0
rope_alpha: Optional[float] = None
rope_alpha: Optional[float] = 1.0
prompt_template: Optional[str] = None
cache_mode: Optional[str] = "FP16"
draft: Optional['ModelCard'] = None
Expand Down Expand Up @@ -35,7 +35,7 @@ class ModelLoadRequest(BaseModel):
gpu_split_auto: Optional[bool] = True
gpu_split: Optional[List[float]] = Field(default_factory=list)
rope_scale: Optional[float] = 1.0
rope_alpha: Optional[float] = 1.0
rope_alpha: Optional[float] = None
no_flash_attention: Optional[bool] = False
# low_mem: Optional[bool] = False
cache_mode: Optional[str] = "FP16"
Expand Down

0 comments on commit 70fbee3

Please sign in to comment.