Fix Compute Tab for a local provider

api/pyproject.toml

@@ -19,7 +19,6 @@ dependencies = [
   "boto3==1.40.70",
   "datasets==3.6.0",
   "diffusers==0.36.0",
-  "do-i-have-the-vram==0.1.0",
   "evaluate==0.4.3",
   "fastapi-users[sqlalchemy,oauth]==15.0.1",
   "hf_xet==1.1.4",

api/transformerlab/routers/model.py

@@ -26,8 +26,6 @@
 from transformerlab.models import filesystemmodel
 import transformerlab.services.job_service as job_service
 from transformerlab.services.job_service import job_update_status
-from transformerlab.services import vram_service
-from transformerlab.schemas.vram import VramEstimateResponse
 from lab.dirs import get_workspace_dir
 from lab.model import Model as ModelService
 from lab import storage
@@ -218,34 +216,6 @@ async def model_gallery(model_id: str):
     return await get_model_details_from_gallery(model_id)
 
 
-@router.get("/model/vram_estimate", response_model=VramEstimateResponse)
-async def model_vram_estimate(
-    model_id: str,
-    dtype: str = "float16",
-    batch: int = 1,
-    seq_len: int = 4096,
-    no_kv: bool = False,
-    filename: str | None = None,
-    x_team_id: str | None = Header(None, alias="X-Team-Id"),
-    user: User = Depends(current_active_user),
-) -> VramEstimateResponse:
-    model_id = model_id.replace("~~~", "/")
-    user_id = str(user.id) if user else None
-    return await vram_service.estimate_vram(
-        model_id=model_id,
-        dtype=dtype,
-        batch=batch,
-        seq_len=seq_len,
-        no_kv=no_kv,
-        filename=filename,
-        user_id=user_id,
-        team_id=x_team_id,
-    )
-
-
-# Should this be a POST request?
-
-
 @router.get("/model/upload_to_huggingface", summary="Given a model ID, upload it to Hugging Face.")
 async def upload_model_to_huggingface(
     model_id: str, model_name: str = "transformerlab-model", organization_name: str = "", model_card_data: str = "{}"

api/transformerlab/routers/serverinfo.py

@@ -165,9 +165,6 @@ async def get_macmon_data():
 @router.get("/info")
 async def get_computer_information():
     # start with our static system information and add current performance details
-    if os.environ.get("MULTIUSER", "true").lower() == "true":
-        return system_info
-
     r = system_info
 
     # Get the current disk usage if its a mac