File size: 13,357 Bytes
7db0ae4 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 |
import os, types
from enum import Enum
import json
import requests
import time
from typing import Callable, Optional, Any
import litellm
from litellm.utils import ModelResponse, EmbeddingResponse, get_secret, Usage
import sys
from copy import deepcopy
import httpx
from .prompt_templates.factory import prompt_factory, custom_prompt
class SagemakerError(Exception):
def __init__(self, status_code, message):
self.status_code = status_code
self.message = message
self.request = httpx.Request(
method="POST", url="https://us-west-2.console.aws.amazon.com/sagemaker"
)
self.response = httpx.Response(status_code=status_code, request=self.request)
super().__init__(
self.message
) # Call the base class constructor with the parameters it needs
class SagemakerConfig:
"""
Reference: https://d-uuwbxj1u4cnu.studio.us-west-2.sagemaker.aws/jupyter/default/lab/workspaces/auto-q/tree/DemoNotebooks/meta-textgeneration-llama-2-7b-SDK_1.ipynb
"""
max_new_tokens: Optional[int] = None
top_p: Optional[float] = None
temperature: Optional[float] = None
return_full_text: Optional[bool] = None
def __init__(
self,
max_new_tokens: Optional[int] = None,
top_p: Optional[float] = None,
temperature: Optional[float] = None,
return_full_text: Optional[bool] = None,
) -> None:
locals_ = locals()
for key, value in locals_.items():
if key != "self" and value is not None:
setattr(self.__class__, key, value)
@classmethod
def get_config(cls):
return {
k: v
for k, v in cls.__dict__.items()
if not k.startswith("__")
and not isinstance(
v,
(
types.FunctionType,
types.BuiltinFunctionType,
classmethod,
staticmethod,
),
)
and v is not None
}
"""
SAGEMAKER AUTH Keys/Vars
os.environ['AWS_ACCESS_KEY_ID'] = ""
os.environ['AWS_SECRET_ACCESS_KEY'] = ""
"""
# set os.environ['AWS_REGION_NAME'] = <your-region_name>
def completion(
model: str,
messages: list,
model_response: ModelResponse,
print_verbose: Callable,
encoding,
logging_obj,
custom_prompt_dict={},
hf_model_name=None,
optional_params=None,
litellm_params=None,
logger_fn=None,
):
import boto3
# pop aws_secret_access_key, aws_access_key_id, aws_region_name from kwargs, since completion calls fail with them
aws_secret_access_key = optional_params.pop("aws_secret_access_key", None)
aws_access_key_id = optional_params.pop("aws_access_key_id", None)
aws_region_name = optional_params.pop("aws_region_name", None)
if aws_access_key_id != None:
# uses auth params passed to completion
# aws_access_key_id is not None, assume user is trying to auth using litellm.completion
client = boto3.client(
service_name="sagemaker-runtime",
aws_access_key_id=aws_access_key_id,
aws_secret_access_key=aws_secret_access_key,
region_name=aws_region_name,
)
else:
# aws_access_key_id is None, assume user is trying to auth using env variables
# boto3 automaticaly reads env variables
# we need to read region name from env
# I assume majority of users use .env for auth
region_name = (
get_secret("AWS_REGION_NAME")
or "us-west-2" # default to us-west-2 if user not specified
)
client = boto3.client(
service_name="sagemaker-runtime",
region_name=region_name,
)
# pop streaming if it's in the optional params as 'stream' raises an error with sagemaker
inference_params = deepcopy(optional_params)
inference_params.pop("stream", None)
## Load Config
config = litellm.SagemakerConfig.get_config()
for k, v in config.items():
if (
k not in inference_params
): # completion(top_k=3) > sagemaker_config(top_k=3) <- allows for dynamic variables to be passed in
inference_params[k] = v
model = model
if model in custom_prompt_dict:
# check if the model has a registered custom prompt
model_prompt_details = custom_prompt_dict[model]
prompt = custom_prompt(
role_dict=model_prompt_details.get("roles", None),
initial_prompt_value=model_prompt_details.get("initial_prompt_value", ""),
final_prompt_value=model_prompt_details.get("final_prompt_value", ""),
messages=messages,
)
else:
if hf_model_name is None:
if "llama-2" in model.lower(): # llama-2 model
if "chat" in model.lower(): # apply llama2 chat template
hf_model_name = "meta-llama/Llama-2-7b-chat-hf"
else: # apply regular llama2 template
hf_model_name = "meta-llama/Llama-2-7b"
hf_model_name = (
hf_model_name or model
) # pass in hf model name for pulling it's prompt template - (e.g. `hf_model_name="meta-llama/Llama-2-7b-chat-hf` applies the llama2 chat template to the prompt)
prompt = prompt_factory(model=hf_model_name, messages=messages)
data = json.dumps({"inputs": prompt, "parameters": inference_params}).encode(
"utf-8"
)
## LOGGING
request_str = f"""
response = client.invoke_endpoint(
EndpointName={model},
ContentType="application/json",
Body={data},
CustomAttributes="accept_eula=true",
)
""" # type: ignore
logging_obj.pre_call(
input=prompt,
api_key="",
additional_args={
"complete_input_dict": data,
"request_str": request_str,
"hf_model_name": hf_model_name,
},
)
## COMPLETION CALL
try:
response = client.invoke_endpoint(
EndpointName=model,
ContentType="application/json",
Body=data,
CustomAttributes="accept_eula=true",
)
except Exception as e:
raise SagemakerError(status_code=500, message=f"{str(e)}")
response = response["Body"].read().decode("utf8")
## LOGGING
logging_obj.post_call(
input=prompt,
api_key="",
original_response=response,
additional_args={"complete_input_dict": data},
)
print_verbose(f"raw model_response: {response}")
## RESPONSE OBJECT
completion_response = json.loads(response)
try:
completion_response_choices = completion_response[0]
completion_output = ""
if "generation" in completion_response_choices:
completion_output += completion_response_choices["generation"]
elif "generated_text" in completion_response_choices:
completion_output += completion_response_choices["generated_text"]
# check if the prompt template is part of output, if so - filter it out
if completion_output.startswith(prompt) and "<s>" in prompt:
completion_output = completion_output.replace(prompt, "", 1)
model_response["choices"][0]["message"]["content"] = completion_output
except:
raise SagemakerError(
message=f"LiteLLM Error: Unable to parse sagemaker RAW RESPONSE {json.dumps(completion_response)}",
status_code=500,
)
## CALCULATING USAGE - baseten charges on time, not tokens - have some mapping of cost here.
prompt_tokens = len(encoding.encode(prompt))
completion_tokens = len(
encoding.encode(model_response["choices"][0]["message"].get("content", ""))
)
model_response["created"] = int(time.time())
model_response["model"] = model
usage = Usage(
prompt_tokens=prompt_tokens,
completion_tokens=completion_tokens,
total_tokens=prompt_tokens + completion_tokens,
)
model_response.usage = usage
return model_response
# async def acompletion(
# client: Any,
# model_response: ModelResponse,
# model: str,
# logging_obj: Any,
# data: dict,
# hf_model_name: str,
# ):
# """
# Use boto3 create_invocation_async endpoint
# """
# ## LOGGING
# request_str = f"""
# response = client.invoke_endpoint(
# EndpointName={model},
# ContentType="application/json",
# Body={data},
# CustomAttributes="accept_eula=true",
# )
# """ # type: ignore
# logging_obj.pre_call(
# input=data["prompt"],
# api_key="",
# additional_args={
# "complete_input_dict": data,
# "request_str": request_str,
# "hf_model_name": hf_model_name,
# },
# )
# ## COMPLETION CALL
# try:
# response = client.invoke_endpoint(
# EndpointName=model,
# ContentType="application/json",
# Body=data,
# CustomAttributes="accept_eula=true",
# )
# except Exception as e:
# raise SagemakerError(status_code=500, message=f"{str(e)}")
def embedding(
model: str,
input: list,
model_response: EmbeddingResponse,
print_verbose: Callable,
encoding,
logging_obj,
custom_prompt_dict={},
optional_params=None,
litellm_params=None,
logger_fn=None,
):
"""
Supports Huggingface Jumpstart embeddings like GPT-6B
"""
### BOTO3 INIT
import boto3
# pop aws_secret_access_key, aws_access_key_id, aws_region_name from kwargs, since completion calls fail with them
aws_secret_access_key = optional_params.pop("aws_secret_access_key", None)
aws_access_key_id = optional_params.pop("aws_access_key_id", None)
aws_region_name = optional_params.pop("aws_region_name", None)
if aws_access_key_id != None:
# uses auth params passed to completion
# aws_access_key_id is not None, assume user is trying to auth using litellm.completion
client = boto3.client(
service_name="sagemaker-runtime",
aws_access_key_id=aws_access_key_id,
aws_secret_access_key=aws_secret_access_key,
region_name=aws_region_name,
)
else:
# aws_access_key_id is None, assume user is trying to auth using env variables
# boto3 automaticaly reads env variables
# we need to read region name from env
# I assume majority of users use .env for auth
region_name = (
get_secret("AWS_REGION_NAME")
or "us-west-2" # default to us-west-2 if user not specified
)
client = boto3.client(
service_name="sagemaker-runtime",
region_name=region_name,
)
# pop streaming if it's in the optional params as 'stream' raises an error with sagemaker
inference_params = deepcopy(optional_params)
inference_params.pop("stream", None)
## Load Config
config = litellm.SagemakerConfig.get_config()
for k, v in config.items():
if (
k not in inference_params
): # completion(top_k=3) > sagemaker_config(top_k=3) <- allows for dynamic variables to be passed in
inference_params[k] = v
#### HF EMBEDDING LOGIC
data = json.dumps({"text_inputs": input}).encode("utf-8")
## LOGGING
request_str = f"""
response = client.invoke_endpoint(
EndpointName={model},
ContentType="application/json",
Body={data},
CustomAttributes="accept_eula=true",
)""" # type: ignore
logging_obj.pre_call(
input=input,
api_key="",
additional_args={"complete_input_dict": data, "request_str": request_str},
)
## EMBEDDING CALL
try:
response = client.invoke_endpoint(
EndpointName=model,
ContentType="application/json",
Body=data,
CustomAttributes="accept_eula=true",
)
except Exception as e:
raise SagemakerError(status_code=500, message=f"{str(e)}")
response = json.loads(response["Body"].read().decode("utf8"))
## LOGGING
logging_obj.post_call(
input=input,
api_key="",
original_response=response,
additional_args={"complete_input_dict": data},
)
print_verbose(f"raw model_response: {response}")
if "embedding" not in response:
raise SagemakerError(status_code=500, message="embedding not found in response")
embeddings = response["embedding"]
if not isinstance(embeddings, list):
raise SagemakerError(
status_code=422, message=f"Response not in expected format - {embeddings}"
)
output_data = []
for idx, embedding in enumerate(embeddings):
output_data.append(
{"object": "embedding", "index": idx, "embedding": embedding}
)
model_response["object"] = "list"
model_response["data"] = output_data
model_response["model"] = model
input_tokens = 0
for text in input:
input_tokens += len(encoding.encode(text))
model_response["usage"] = Usage(
prompt_tokens=input_tokens, completion_tokens=0, total_tokens=input_tokens
)
return model_response
|