Implement lora support (#24)

* Model: Implement basic lora support

* Add ability to load loras from config on launch
* Supports loading multiple loras and lora scaling
* Add function to unload loras

* Colab: Update for basic lora support

* Model: Test vram alloc after lora load, add docs

* Git: Add loras folder to .gitignore

* API: Add basic lora-related endpoints

* Add /loras/ endpoint for querying available loras
* Add /model/lora endpoint for querying currently loaded loras
* Add /model/lora/load endpoint for loading loras
* Add /model/lora/unload endpoint for unloading loras
* Move lora config-checking logic to main.py for better compat with API endpoints

* Revert bad CRLF line ending changes

* API: Add basic lora-related endpoints (fixed)

* Add /loras/ endpoint for querying available loras
* Add /model/lora endpoint for querying currently loaded loras
* Add /model/lora/load endpoint for loading loras
* Add /model/lora/unload endpoint for unloading loras
* Move lora config-checking logic to main.py for better compat with API endpoints

* Model: Unload loras first when unloading model

* API + Models: Cleanup lora endpoints and functions

Condenses down endpoint and model load code. Also makes the routes
behave the same way as model routes to help not confuse the end user.

Signed-off-by: kingbri <bdashore3@proton.me>

* Loras: Optimize load endpoint

Return successes and failures along with consolidating the request
to the rewritten load_loras function.

Signed-off-by: kingbri <bdashore3@proton.me>

---------

Co-authored-by: kingbri <bdashore3@proton.me>
Co-authored-by: DocShotgun <126566557+DocShotgun@users.noreply.github.com>

OAI/types/lora.py

@@ -0,0 +1,25 @@
+from pydantic import BaseModel, Field;
+from time import time
+from typing import Optional, List
+
+class LoraCard(BaseModel):
+    id: str = "test"
+    object: str = "lora"
+    created: int = Field(default_factory=lambda: int(time()))
+    owned_by: str = "tabbyAPI"
+    scaling: Optional[float] = None
+
+class LoraList(BaseModel):
+    object: str = "list"
+    data: List[LoraCard] = Field(default_factory=list)
+
+class LoraLoadInfo(BaseModel):
+    name: str
+    scaling: Optional[float] = 1.0
+
+class LoraLoadRequest(BaseModel):
+    loras: List[LoraLoadInfo]
+
+class LoraLoadResponse(BaseModel):
+    success: List[str] = Field(default_factory=list)
+    failure: List[str] = Field(default_factory=list)

OAI/utils.py

@@ -8,9 +8,10 @@ from OAI.types.chat_completion import (
     ChatCompletionStreamChoice
 )
 from OAI.types.common import UsageStats
+from OAI.types.lora import LoraList, LoraCard
 from OAI.types.model import ModelList, ModelCard
 from packaging import version
-from typing import Optional, List
+from typing import Optional, List, Dict
 
 # Check fastchat
 try:
@@ -100,6 +101,15 @@ def get_model_list(model_path: pathlib.Path, draft_model_path: Optional[str]):
 
     return model_card_list
 
+def get_lora_list(lora_path: pathlib.Path):
+    lora_list = LoraList()
+    for path in lora_path.iterdir():
+        if path.is_dir():
+            lora_card = LoraCard(id = path.name)
+            lora_list.data.append(lora_card)
+
+    return lora_list
+
 def get_chat_completion_prompt(model_path: str, messages: List[ChatCompletionMessage]):
 
     # Check if fastchat is available