AI/dify
1
0
Fork 0
mirror of https://git.mirrors.martin98.com/https://github.com/langgenius/dify.git synced 2025-08-11 03:19:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

chore: make prompt generator max tokens configurable (#6693)

Browse Source
This commit is contained in:
Sangmin Ahn 2024-07-26 11:20:23 +09:00 committed by GitHub
parent bd97ce9489
commit ecb9c311b5
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 8 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
api/.env.example
View File

@ -183,6 +183,7 @@ UPLOAD_IMAGE_FILE_SIZE_LIMIT=10
# Model Configuration # Model Configuration
MULTIMODAL_SEND_IMAGE_FORMAT=base64 MULTIMODAL_SEND_IMAGE_FORMAT=base64
PROMPT_GENERATION_MAX_TOKENS=512
# Mail configuration, support: resend, smtp # Mail configuration, support: resend, smtp
MAIL_TYPE= MAIL_TYPE=

6
api/controllers/console/app/generator.py
View File

@ -1,3 +1,5 @@
import os
from flask_login import current_user from flask_login import current_user
from flask_restful import Resource, reqparse from flask_restful import Resource, reqparse
@ -28,13 +30,15 @@ class RuleGenerateApi(Resource):
args = parser.parse_args() args = parser.parse_args()
account = current_user account = current_user
PROMPT_GENERATION_MAX_TOKENS = int(os.getenv('PROMPT_GENERATION_MAX_TOKENS', '512'))
try: try:
rules = LLMGenerator.generate_rule_config( rules = LLMGenerator.generate_rule_config(
tenant_id=account.current_tenant_id, tenant_id=account.current_tenant_id,
instruction=args['instruction'], instruction=args['instruction'],
model_config=args['model_config'], model_config=args['model_config'],
no_variable=args['no_variable'] no_variable=args['no_variable'],
rule_config_max_tokens=PROMPT_GENERATION_MAX_TOKENS
) )
except ProviderTokenNotInitError as ex: except ProviderTokenNotInitError as ex:
raise ProviderNotInitializeError(ex.description) raise ProviderNotInitializeError(ex.description)

4
api/core/llm_generator/llm_generator.py
View File

@ -118,7 +118,7 @@ class LLMGenerator:
return questions return questions
@classmethod @classmethod
def generate_rule_config(cls, tenant_id: str, instruction: str, model_config: dict, no_variable: bool) -> dict: def generate_rule_config(cls, tenant_id: str, instruction: str, model_config: dict, no_variable: bool, rule_config_max_tokens: int = 512) -> dict:
output_parser = RuleConfigGeneratorOutputParser() output_parser = RuleConfigGeneratorOutputParser()
error = "" error = ""
@ -130,7 +130,7 @@ class LLMGenerator:
"error": "" "error": ""
} }
model_parameters = { model_parameters = {
"max_tokens": 512, "max_tokens": rule_config_max_tokens,
"temperature": 0.01 "temperature": 0.01
} }