feat: Organize Inference Files - Part 1 (#216)

This is the first part towards the new inference pipeline - organizing
inference files into their own folder.

This way we can easier trigger build pipelines (edits to
presets/models/inference) as well as maintain/add new types of inference

pkg/inference/preset-inference-types.go

@@ -48,7 +48,6 @@ var (
 	}
 
 	DefaultAccelerateParams = map[string]string{
-		"config_file":   DefaultConfigFile,
 		"num_processes": DefaultNumProcesses,
 		"num_machines":  DefaultNumMachines,
 		"machine_rank":  DefaultMachineRank,

presets/models/llama2chat/inference-api.py → presets/models/inference/llama2-chat/inference-api.py

@@ -1,21 +1,21 @@
 # Copyright (c) Microsoft Corporation.
 # Licensed under the MIT license.
-from fastapi import FastAPI, HTTPException
-import uvicorn
-from pydantic import BaseModel
-from typing import Optional
+import argparse
+import functools
 import multiprocessing
 import multiprocessing.pool
+import os
+import signal
+import sys
 import threading
-import functools
+from typing import Optional
 
-from llama import Llama
 import torch
-import sys
-import signal
-import os
 import torch.distributed as dist
-import argparse
+import uvicorn
+from fastapi import FastAPI, HTTPException
+from llama import Llama
+from pydantic import BaseModel
 
 # Setup argparse
 parser = argparse.ArgumentParser(description="Llama API server.")

presets/models/llama2/inference-api.py → presets/models/inference/llama2-completion/inference-api.py

@@ -1,21 +1,21 @@
 # Copyright (c) Microsoft Corporation.
 # Licensed under the MIT license.
-from fastapi import FastAPI, HTTPException
-import uvicorn
-from pydantic import BaseModel
-from typing import Optional
+import argparse
+import functools
 import multiprocessing
 import multiprocessing.pool
+import os
+import signal
+import sys
 import threading
-import functools
+from typing import Optional
 
-from llama import Llama
 import torch
-import sys
-import signal
-import os
 import torch.distributed as dist
-import argparse
+import uvicorn
+from fastapi import FastAPI, HTTPException
+from llama import Llama
+from pydantic import BaseModel
 
 # Setup argparse
 parser = argparse.ArgumentParser(description="Llama API server.")