blob: 42091b96efee76829b033409f694c724e64953c9 (
plain)
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
|
from typing import Callable
from abc import ABC, abstractmethod
def make_chat_template(
prompt: str,
response_prefix: str = "",
is_chat: bool = True,
tokenizer: Callable = None
) -> str:
if is_chat:
prompt = tokenizer.apply_chat_template(
[
{"role": "user", "content": prompt},
],
tokenize = False,
add_generation_prompt = True
) + response_prefix
if tokenizer.bos_token and prompt.startswith(tokenizer.bos_token):
prompt = prompt[len(tokenizer.bos_token):]
return prompt
else:
return prompt
class Generator(ABC):
model_name: str = None
def __init__(self, model_name: str) -> None:
"""
:param stop_words: list
list of stop words if the generation uses a stopping criteria during generation
:param requires_execution: bool
wheter the task requires code execution during evaluation or not
"""
self.model_name = model_name
def fewshot_examples(self):
"""Loads and returns the few-shot examples for the task if they exist."""
pass
@abstractmethod
def set_stop(self):
"""
Set the stop tokens for the model
"""
pass
@abstractmethod
def generate(self):
"""Builds the prompt for the LM to generate from.
:param doc: dict[str: str]
sample from the test dataset
"""
pass
|
