Spaces:
Runtime error
Runtime error
Update run.py
Browse files
run.py
CHANGED
|
@@ -342,8 +342,8 @@ def response(message, history):
|
|
| 342 |
|
| 343 |
# Request Response from LLM:
|
| 344 |
system2=None # system2 can be used as fictive first words of the AI, which are not displayed or stored
|
| 345 |
-
print("RAG: "+rag)
|
| 346 |
-
print("System: "+system+"\n\nMessage: "+message)
|
| 347 |
prompt=extend_prompt(
|
| 348 |
message, # current message of the user
|
| 349 |
history, # complete history
|
|
@@ -352,7 +352,7 @@ def response(message, history):
|
|
| 352 |
system2, # fictive first words of the AI (neither displayed nor stored)
|
| 353 |
historylimit=historylimit # number of past messages to consider for response to current message
|
| 354 |
)
|
| 355 |
-
print(prompt)
|
| 356 |
|
| 357 |
## Request response from model
|
| 358 |
#------------------------------
|
|
@@ -375,10 +375,10 @@ def response(message, history):
|
|
| 375 |
)
|
| 376 |
stream = client.text_generation(prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
|
| 377 |
response = ""
|
| 378 |
-
print("User: "+message+"\nAI: ")
|
| 379 |
for text in stream:
|
| 380 |
part=text.token.text
|
| 381 |
-
print(part, end="", flush=True)
|
| 382 |
response += part
|
| 383 |
yield response
|
| 384 |
if((myType=="1a")): #add RAG-results to chat-output if appropriate
|
|
@@ -406,7 +406,7 @@ def response(message, history):
|
|
| 406 |
response="" #+"("+myType+")\n"
|
| 407 |
buffer=""
|
| 408 |
#print("URL: "+url)
|
| 409 |
-
print("User: "+message+"\nAI: ")
|
| 410 |
for text in requests.post(url, json=body, stream=True): #-H 'accept: application/json' -H 'Content-Type: application/json'
|
| 411 |
if buffer is None: buffer=""
|
| 412 |
buffer=str("".join(buffer))
|
|
@@ -423,7 +423,7 @@ def response(message, history):
|
|
| 423 |
if(part.lstrip('\n\r').startswith("data: ")): part=part.lstrip('\n\r').replace("data: ", "")
|
| 424 |
try:
|
| 425 |
part = str(json.loads(part)["choices"][0]["text"])
|
| 426 |
-
print(part, end="", flush=True)
|
| 427 |
response=response+part
|
| 428 |
buffer="" # reset buffer
|
| 429 |
except Exception as e:
|
|
|
|
| 342 |
|
| 343 |
# Request Response from LLM:
|
| 344 |
system2=None # system2 can be used as fictive first words of the AI, which are not displayed or stored
|
| 345 |
+
#print("RAG: "+rag)
|
| 346 |
+
#print("System: "+system+"\n\nMessage: "+message)
|
| 347 |
prompt=extend_prompt(
|
| 348 |
message, # current message of the user
|
| 349 |
history, # complete history
|
|
|
|
| 352 |
system2, # fictive first words of the AI (neither displayed nor stored)
|
| 353 |
historylimit=historylimit # number of past messages to consider for response to current message
|
| 354 |
)
|
| 355 |
+
print("\n\nPrompt:\n"+prompt)
|
| 356 |
|
| 357 |
## Request response from model
|
| 358 |
#------------------------------
|
|
|
|
| 375 |
)
|
| 376 |
stream = client.text_generation(prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
|
| 377 |
response = ""
|
| 378 |
+
#print("User: "+message+"\nAI: ")
|
| 379 |
for text in stream:
|
| 380 |
part=text.token.text
|
| 381 |
+
#print(part, end="", flush=True)
|
| 382 |
response += part
|
| 383 |
yield response
|
| 384 |
if((myType=="1a")): #add RAG-results to chat-output if appropriate
|
|
|
|
| 406 |
response="" #+"("+myType+")\n"
|
| 407 |
buffer=""
|
| 408 |
#print("URL: "+url)
|
| 409 |
+
#print("User: "+message+"\nAI: ")
|
| 410 |
for text in requests.post(url, json=body, stream=True): #-H 'accept: application/json' -H 'Content-Type: application/json'
|
| 411 |
if buffer is None: buffer=""
|
| 412 |
buffer=str("".join(buffer))
|
|
|
|
| 423 |
if(part.lstrip('\n\r').startswith("data: ")): part=part.lstrip('\n\r').replace("data: ", "")
|
| 424 |
try:
|
| 425 |
part = str(json.loads(part)["choices"][0]["text"])
|
| 426 |
+
#print(part, end="", flush=True)
|
| 427 |
response=response+part
|
| 428 |
buffer="" # reset buffer
|
| 429 |
except Exception as e:
|