Spaces:
Runtime error
Runtime error
File size: 2,237 Bytes
9045a87 308dd86 57817e9 b91d77f 57817e9 8db8595 57817e9 b91d77f 8db8595 00d1f95 8db8595 308dd86 57817e9 308dd86 9045a87 57817e9 |
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 56 57 58 |
import gradio as gr
import requests
import random
import json
def response(message, history, model):
url="https://afischer1985-OpenHermes-2-GGUF-API.hf.space/v1/completions"
prompt_type="ChatML"
endstr="<|im_end|>"
print(model)
if(model=="WizardLM-13B"):
url="https://wizardlm-13b-v1-2-q4-0-gguf.hf.space/v1/completions"
prompt_type="Alpaca"
if(model=="SauerkrautLM-7B"):
url="https://SauerkrautLM-GGUF-API.hf.space/v1/completions"
prompt_type="Alpaca"
if(model=="OpenHermes2-7B"):
url="https://AFischer1985-CollectiveCognition-GGUF-API.hf.space/v1/completions"
prompt_type="ChatML"
if(model=="CollectiveCognition-7B"):
url="https://AFischer1985-CollectiveCognition-GGUF-API.hf.space/v1/completions"
prompt_type="ChatML"
print(prompt_type)
if(prompt_type=="Alpaca"):
body={"prompt":"###Frage: "+message+" ###Antwort:","max_tokens":1000,"stop":"###","stream":True}
if(prompt_type!="Alpaca"):
body={"prompt":"""<|im_start|>system
You are a helpful AI-Assistant.<|im_end|>
<|im_start|>user
"""+message+"""<|im_end|>
<|im_start|>assistant
""","max_tokens":1000,"stop":"<|im_end|>","stream":True}
response=""
buffer=""
print(str(body))
print("User: "+message+"\nAI: ")
for text in requests.post(url, json=body, stream=True):
#print("*** Raw String: "+str(text)+"\n***\n")
text=text.decode('utf-8')
if(text.startswith(": ping -")==False):buffer=str(buffer)+str(text)
#if(text.startswith(": ping -")): print("\n*** PING!\n***\n")
#print("\n*** Buffer: "+str(buffer)+"\n***\n")
buffer=buffer.split('"finish_reason"')
if(len(buffer)==1):
buffer="".join(buffer)
pass
if(len(buffer)==2):
part=buffer[0]+'"finish_reason": null}]}'
if(part.startswith("data: ")):part=part.replace("data: ", "")
try:
part = str(json.loads(part)["choices"][0]["text"])
print(part, end="", flush=True)
response=response+part
buffer="" # reset buffer
except:
pass
yield response
gr.ChatInterface(response,additional_inputs=[gr.Dropdown(["CollectiveCognition-7B", "OpenHermes2-7B"],value="OpenHermes2-7B",label="Model")]).queue().launch(share=True) |