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

[data] add min resolution option (#6975)

Browse Source 
Former-commit-id: 76bd9a98a2fb00f1a1d881e6e1364c02fd36d327
This commit is contained in:
hoshi-hiyouga
2025-02-18 01:40:46 +08:00
committed by GitHub
parent f2fd9d1b25
commit c09b648934
9 changed files with 59 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
src/llamafactory/api/chat.py
View File

@@ -22,6 +22,8 @@ from typing import TYPE_CHECKING, AsyncGenerator, Dict, List, Optional, Tuple
from ..data import Role as DataRole
from ..extras import logging
from ..extras.constants import IMAGE_PLACEHOLDER
from ..extras.misc import is_env_enabled
from ..extras.packages import is_fastapi_available, is_pillow_available, is_requests_available
from .common import dictify, jsonify
from .protocol import (
@@ -70,7 +72,8 @@ ROLE_MAPPING = {
def _process_request(
request: "ChatCompletionRequest",
) -> Tuple[List[Dict[str, str]], Optional[str], Optional[str], Optional[List["ImageInput"]]]:
logger.info_rank0(f"==== request ====\n{json.dumps(dictify(request), indent=2, ensure_ascii=False)}")
if is_env_enabled("API_VERBOSE", "1"):
logger.info_rank0(f"==== request ====\n{json.dumps(dictify(request), indent=2, ensure_ascii=False)}")
if len(request.messages) == 0:
raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail="Invalid length")
@@ -99,10 +102,12 @@ def _process_request(
content = json.dumps(tool_calls, ensure_ascii=False)
input_messages.append({"role": ROLE_MAPPING[Role.FUNCTION], "content": content})
elif isinstance(message.content, list):
text_content = ""
for input_item in message.content:
if input_item.type == "text":
input_messages.append({"role": ROLE_MAPPING[message.role], "content": input_item.text})
text_content += input_item.text
else:
text_content += IMAGE_PLACEHOLDER
image_url = input_item.image_url.url
if re.match(r"^data:image\/(png|jpg|jpeg|gif|bmp);base64,(.+)$", image_url): # base64 image
image_stream = io.BytesIO(base64.b64decode(image_url.split(",", maxsplit=1)[1]))
@@ -112,6 +117,8 @@ def _process_request(
image_stream = requests.get(image_url, stream=True).raw
images.append(Image.open(image_stream).convert("RGB"))
input_messages.append({"role": ROLE_MAPPING[message.role], "content": text_content})
else:
input_messages.append({"role": ROLE_MAPPING[message.role], "content": message.content})