vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-10 09:04:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

emulated oai image generation

Browse source
This commit is contained in:
Concedo 2025-07-02 16:01:56 +08:00
parent cdda9d16e0
commit f407aa3b8a
1 changed files with 20 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

21
koboldcpp.py
View file

@ -1639,6 +1639,18 @@ def sd_load_model(model_filename,vae_filename,lora_filename,t5xxl_filename,clipl
ret = handle.sd_load_model(inputs)
return ret
def sd_oai_tranform_params(genparams):
size = genparams.get('size', "512x512")
if size and size!="":
pattern = r'^\D*(\d+)x(\d+)$'
match = re.fullmatch(pattern, size)
if match:
width = int(match.group(1))
height = int(match.group(2))
genparams["width"] = width
genparams["height"] = height
return genparams
def sd_comfyui_tranform_params(genparams):
promptobj = genparams.get('prompt', None)
if promptobj and isinstance(promptobj, dict):
@ -3700,6 +3712,7 @@ Change Mode<br>
api_format = 0 #1=basic,2=kai,3=oai,4=oai-chat,5=interrogate,6=ollama,7=ollamachat
is_imggen = False
is_comfyui_imggen = False
is_oai_imggen = False
is_transcribe = False
is_tts = False
is_embeddings = False
@ -3785,10 +3798,12 @@ Change Mode<br>
api_format = 6
elif self.path.endswith('/api/chat'): #ollama
api_format = 7
elif self.path=="/prompt" or self.path.endswith('/sdapi/v1/txt2img') or self.path.endswith('/sdapi/v1/img2img'):
elif self.path=="/prompt" or self.path.endswith('/v1/images/generations') or self.path.endswith('/sdapi/v1/txt2img') or self.path.endswith('/sdapi/v1/img2img'):
is_imggen = True
if self.path=="/prompt":
is_comfyui_imggen = True
elif self.path.endswith('/v1/images/generations'):
is_oai_imggen = True
elif self.path.endswith('/api/extra/transcribe') or self.path.endswith('/v1/audio/transcriptions'):
is_transcribe = True
elif self.path.endswith('/api/extra/tts') or self.path.endswith('/v1/audio/speech') or self.path.endswith('/tts_to_audio'):
@ -3898,6 +3913,8 @@ Change Mode<br>
if is_comfyui_imggen:
lastgeneratedcomfyimg = b''
genparams = sd_comfyui_tranform_params(genparams)
elif is_oai_imggen:
genparams = sd_oai_tranform_params(genparams)
gen = sd_generate(genparams)
genresp = None
if is_comfyui_imggen:
@ -3906,6 +3923,8 @@ Change Mode<br>
else:
lastgeneratedcomfyimg = b''
genresp = (json.dumps({"prompt_id": "12345678-0000-0000-0000-000000000001","number": 0,"node_errors":{}}).encode())
elif is_oai_imggen:
genresp = (json.dumps({"created":int(time.time()),"data":[{"b64_json":gen}],"background":"opaque","output_format":"png","size":"1024x1024","quality":"medium"}).encode())
else:
genresp = (json.dumps({"images":[gen],"parameters":{},"info":""}).encode())
self.send_response(200)