p04798526
/
LLaMA-Factory-Mirror
1
0
Fork 1
Code Issues Pull Requests Packages Projects Releases Wiki Activity

fix #2282 and update tool prompt

This commit is contained in:
hiyouga 2024-01-22 22:27:30 +08:00
parent 6fc2d5cc03
commit b2fb0eca56
5 changed files with 25 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
data/alpaca_data_zh_51k.json
View File

@ -11192,7 +11192,7 @@
{
"instruction": "从给定数据中找出两个事件的概率",
"input": "事件A：会下雨\n事件B：会下雪\n数据：\n- 明日天气预报：下雨的概率为80％，下雪的概率为20％",
"输出": "事件A和事件B同时发生的概率为0.16或16％。这是通过将事件A的概率（0.8）和事件B的概率（0.2）相乘来计算的。"
"output": "事件A和事件B同时发生的概率为0.16或16％。这是通过将事件A的概率（0.8）和事件B的概率（0.2）相乘来计算的。"
},
{
"instruction": "对于给定情境给予建议",

13
src/llmtuner/data/aligner.py
View File

@ -15,7 +15,7 @@ def convert_alpaca(examples: Dict[str, List[Any]], dataset_attr: "DatasetAttr")
outputs = {"prompt": [], "response": [], "system": [], "tools": []}
for i in range(len(examples[dataset_attr.prompt])):
prompt = []
if dataset_attr.history:
if dataset_attr.history and isinstance(examples[dataset_attr.history][i], list):
for old_prompt, old_response in examples[dataset_attr.history][i]:
prompt.append({"role": Role.USER, "content": old_prompt})
prompt.append({"role": Role.ASSISTANT, "content": old_response})
@ -25,13 +25,10 @@ def convert_alpaca(examples: Dict[str, List[Any]], dataset_attr: "DatasetAttr")
instruction += "\n" + examples[dataset_attr.query][i]
prompt.append({"role": Role.USER, "content": instruction})
if dataset_attr.response:
if isinstance(examples[dataset_attr.response][i], list):
response = [
{"role": Role.ASSISTANT, "content": content} for content in examples[dataset_attr.response][i]
]
else:
response = [{"role": Role.ASSISTANT, "content": examples[dataset_attr.response][i]}]
if dataset_attr.response and isinstance(examples[dataset_attr.response][i], list):
response = [{"role": Role.ASSISTANT, "content": content} for content in examples[dataset_attr.response][i]]
elif dataset_attr.response and isinstance(examples[dataset_attr.response][i], str):
response = [{"role": Role.ASSISTANT, "content": examples[dataset_attr.response][i]}]
else:
response = []

13
src/llmtuner/data/formatter.py
View File

@ -15,10 +15,10 @@ JSON_FORMAT_PROMPT = (
TOOL_SYSTEM_PROMPT = (
"You have access to the following tools:\n{tool_text}"
"Use the following format to answer the question:\n"
"Use the following format if using a tool:\n"
"```\n"
"Action: the action to take, should be one of [{tool_names}] if using a tool.\n"
"Action Input: the input to the action{format_prompt}.\n"
"Action: tool name (one of [{tool_names}]).\n"
"Action Input: the input to the tool{format_prompt}.\n"
"```\n"
)
@ -95,12 +95,15 @@ class StringFormatter(Formatter):
for slot in self.slots:
if isinstance(slot, str):
for name, value in kwargs.items():
if not isinstance(value, str):
raise RuntimeError("Expected a string, got {}".format(value))
slot = slot.replace("{{" + name + "}}", value, 1)
elements.append(slot)
elif isinstance(slot, (dict, set)):
elements.append(slot)
else:
raise ValueError("Input must be string, set[str] or dict[str, str], got {}".format(type(slot)))
raise RuntimeError("Input must be string, set[str] or dict[str, str], got {}".format(type(slot)))
return elements
@ -124,7 +127,7 @@ class FunctionFormatter(Formatter):
elif isinstance(slot, (dict, set)):
elements.append(slot)
else:
raise ValueError("Input must be string, set[str] or dict[str, str], got {}".format(type(slot)))
raise RuntimeError("Input must be string, set[str] or dict[str, str], got {}".format(type(slot)))
return elements

15
src/llmtuner/data/preprocess.py
View File

@ -4,6 +4,7 @@ from typing import TYPE_CHECKING, Any, Callable, Dict, List, Literal, Tuple
from ..extras.constants import IGNORE_INDEX
from ..extras.logging import get_logger
from .utils import Role
if TYPE_CHECKING:
@ -51,7 +52,7 @@ def preprocess_supervised_dataset(
model_inputs = {"input_ids": [], "attention_mask": [], "labels": []}
for i in range(len(examples["prompt"])):
if len(examples["prompt"][i]) == 0 or len(examples["response"][i]) != 1:
if len(examples["prompt"][i]) % 2 != 1 or len(examples["response"][i]) != 1:
continue
messages = examples["prompt"][i] + examples["response"][i]
@ -93,7 +94,7 @@ def preprocess_packed_supervised_dataset(
model_inputs = {"input_ids": [], "attention_mask": [], "labels": []}
input_ids, labels = [], []
for i in range(len(examples["prompt"])):
if len(examples["prompt"][i]) == 0 or len(examples["response"][i]) != 1:
if len(examples["prompt"][i]) % 2 != 1 or len(examples["response"][i]) != 1:
continue
messages = examples["prompt"][i] + examples["response"][i]
@ -137,10 +138,14 @@ def preprocess_unsupervised_dataset(
model_inputs = {"input_ids": [], "attention_mask": [], "labels": []}
for i in range(len(examples["prompt"])):
if len(examples["prompt"][i]) == 0 or len(examples["response"][i]) != 1:
if len(examples["prompt"][i]) % 2 != 1:
continue
messages = examples["prompt"][i] + examples["response"][i]
if len(examples["response"][i]) == 1:
messages = examples["prompt"][i] + examples["response"][i]
else:
messages = examples["prompt"][i] + [{"role": Role.ASSISTANT, "content": ""}]
input_ids, labels = template.encode_oneturn(
tokenizer, messages, examples["system"][i], examples["tools"][i], data_args.cutoff_len
)
@ -164,7 +169,7 @@ def preprocess_pairwise_dataset(
# build input pairs with format `<bos> X`, `Y1 <eos>` and `Y2 <eos>`
model_inputs = {"prompt_ids": [], "chosen_ids": [], "rejected_ids": []}
for i in range(len(examples["prompt"])):
if len(examples["prompt"][i]) == 0 or len(examples["response"][i]) < 2:
if len(examples["prompt"][i]) % 2 != 1 or len(examples["response"][i]) < 2:
continue
chosen_messages = examples["prompt"][i] + [examples["response"][i][0]]

2
src/llmtuner/hparams/parser.py
View File

@ -60,7 +60,7 @@ def _verify_model_args(model_args: "ModelArguments", finetuning_args: "Finetunin
if finetuning_args.finetuning_type != "lora":
raise ValueError("Quantization is only compatible with the LoRA method.")
if finetuning_args.create_new_adapter:
if model_args.adapter_name_or_path is not None and finetuning_args.create_new_adapter:
raise ValueError("Cannot create new adapter upon a quantized model.")
if model_args.adapter_name_or_path is not None and len(model_args.adapter_name_or_path) != 1: