rixty/KoboldAI-Client
1
1
Fork 0
mirror of https://github.com/KoboldAI/KoboldAI-Client.git synced 2025-06-05 21:59:24 +02:00
Code Issues Projects Releases Wiki Activity

Added support for Author's Note

Browse Source 
Increased input textarea height
Removed generator options from save/load system
Set output length slider to use steps of 2
This commit is contained in:
KoboldAI Dev
2021-05-05 03:04:06 -04:00
parent 0ce77f4875
commit 229b10cb91
5 changed files with 254 additions and 73 deletions
Download Patch File Download Diff File Expand all files Collapse all files

142
aiserver.py
View File

@ -40,20 +40,22 @@ modellist = [
# Variables
class vars:
lastact = "" # The last action submitted to the generator
model = ''
model = ""
noai = False # Runs the script without starting up the transformers pipeline
aibusy = False # Stops submissions while the AI is working
max_length = 500 # Maximum number of tokens to submit per action
genamt = 60 # Amount of text for each action to generate
rep_pen = 1.0 # Generator repetition_penalty
temp = 0.9 # Generator temperature
top_p = 1.0 # Generator top_p
rep_pen = 1.0 # Default generator repetition_penalty
temp = 0.9 # Default generator temperature
top_p = 1.0 # Default generator top_p
gamestarted = False
prompt = ""
memory = ""
authornote = ""
andepth = 3 # How far back in history to append author's note
actions = []
mode = "play" # Whether the interface is in play, memory, or edit mode
editln = 0 # Which line was last selected in Edit Mode
editln = 0 # Which line was last selected in Edit Mode
url = "https://api.inferkit.com/v1/models/standard/generate" # InferKit API URL
apikey = "" # API key to use for InferKit API calls
savedir = getcwd()+"\stories\\newstory.json"
@ -303,6 +305,13 @@ def get_message(msg):
elif(msg['cmd'] == 'setoutput'):
vars.genamt = int(msg['data'])
emit('from_server', {'cmd': 'setlabeloutput', 'data': msg['data']})
# Author's Note field update
elif(msg['cmd'] == 'anote'):
anotesubmit(msg['data'])
# Author's Note depth update
elif(msg['cmd'] == 'anotedepth'):
vars.andepth = int(msg['data'])
emit('from_server', {'cmd': 'setlabelanotedepth', 'data': msg['data']})
#==================================================================#
#
@ -333,9 +342,20 @@ def actionsubmit(data):
# Take submitted text and build the text to be given to generator
#==================================================================#
def calcsubmit(txt):
vars.lastact = txt # Store most recent action in memory (is this still needed?)
anotetxt = "" # Placeholder for Author's Note text
lnanote = 0 # Placeholder for Author's Note length
forceanote = False # In case we don't have enough actions to hit A.N. depth
anoteadded = False # In case our budget runs out before we hit A.N. depth
actionlen = len(vars.actions)
# Build Author's Note if set
if(vars.authornote != ""):
anotetxt = "\n[Author's note: "+vars.authornote+"]\n"
# For all transformers models
if(vars.model != "InferKit"):
vars.lastact = txt # Store most recent action in memory (is this still needed?)
anotetkns = [] # Placeholder for Author's Note tokens
# Calculate token budget
prompttkns = tokenizer.encode(vars.prompt)
@ -344,17 +364,28 @@ def calcsubmit(txt):
memtokens = tokenizer.encode(vars.memory)
lnmem = len(memtokens)
budget = vars.max_length - lnprompt - lnmem - vars.genamt
if(anotetxt != ""):
anotetkns = tokenizer.encode(anotetxt)
lnanote = len(anotetkns)
if(len(vars.actions) == 0):
budget = vars.max_length - lnprompt - lnmem - lnanote - vars.genamt
if(actionlen == 0):
# First/Prompt action
subtxt = vars.memory + vars.prompt
lnsub = len(memtokens+prompttkns)
subtxt = vars.memory + anotetxt + vars.prompt
lnsub = lnmem + lnprompt + lnanote
generate(subtxt, lnsub+1, lnsub+vars.genamt)
else:
tokens = []
# Check if we have the action depth to hit our A.N. depth
if(anotetxt != "" and actionlen < vars.andepth):
forceanote = True
# Get most recent action tokens up to our budget
tokens = []
for n in range(len(vars.actions)):
for n in range(actionlen):
if(budget <= 0):
break
acttkns = tokenizer.encode(vars.actions[(-1-n)])
@ -366,9 +397,22 @@ def calcsubmit(txt):
count = budget * -1
tokens = acttkns[count:] + tokens
break
# Add mmory & prompt tokens to beginning of bundle
tokens = memtokens + prompttkns + tokens
# Inject Author's Note if we've reached the desired depth
if(n == vars.andepth-1):
if(anotetxt != ""):
tokens = anotetkns + tokens # A.N. len already taken from bdgt
anoteadded = True
# Did we get to add the A.N.? If not, do it here
if(anotetxt != ""):
if((not anoteadded) or forceanote):
tokens = memtokens + anotetkns + prompttkns + tokens
else:
tokens = memtokens + prompttkns + tokens
else:
# Prepend Memory and Prompt before action tokens
tokens = memtokens + prompttkns + tokens
# Send completed bundle to generator
ln = len(tokens)
@ -379,9 +423,15 @@ def calcsubmit(txt):
)
# For InferKit web API
else:
budget = vars.max_length - len(vars.prompt) - len(vars.memory) - 1
# Check if we have the action depth to hit our A.N. depth
if(anotetxt != "" and actionlen < vars.andepth):
forceanote = True
budget = vars.max_length - len(vars.prompt) - len(anotetxt) - len(vars.memory) - 1
subtxt = ""
for n in range(len(vars.actions)):
for n in range(actionlen):
if(budget <= 0):
break
actlen = len(vars.actions[(-1-n)])
@ -392,12 +442,26 @@ def calcsubmit(txt):
count = budget * -1
subtxt = vars.actions[(-1-n)][count:] + subtxt
break
# Inject Author's Note if we've reached the desired depth
if(n == vars.andepth-1):
if(anotetxt != ""):
subtxt = anotetxt + subtxt # A.N. len already taken from bdgt
anoteadded = True
# Add mmory & prompt tokens to beginning of bundle
# Format memory for inclusion (adding newline separator)
memsub = ""
if(vars.memory != ""):
subtxt = vars.memory + "\n" + vars.prompt + subtxt
memsub = vars.memory + "\n"
# Did we get to add the A.N.? If not, do it here
if(anotetxt != ""):
if((not anoteadded) or forceanote):
subtxt = memsub + anotetxt + vars.prompt + subtxt
else:
subtxt = memsub + vars.prompt + subtxt
else:
subtxt = vars.prompt + subtxt
subtxt = memsub + vars.prompt + subtxt
# Send it!
ikrequest(subtxt)
@ -407,7 +471,12 @@ def calcsubmit(txt):
#==================================================================#
def generate(txt, min, max):
print("{0}Min:{1}, Max:{2}, Txt:{3}{4}".format(colors.WARNING, min, max, txt, colors.ENDC))
# Clear CUDA cache if using GPU
if(vars.hascuda and vars.usegpu):
torch.cuda.empty_cache()
# Submit input text to generator
genout = generator(
txt,
do_sample=True,
@ -421,6 +490,10 @@ def generate(txt, min, max):
refresh_story()
emit('from_server', {'cmd': 'texteffect', 'data': len(vars.actions)})
# Clear CUDA cache again if using GPU
if(vars.hascuda and vars.usegpu):
torch.cuda.empty_cache()
set_aibusy(0)
#==================================================================#
@ -460,6 +533,7 @@ def refresh_settings():
emit('from_server', {'cmd': 'updatetopp', 'data': vars.top_p})
emit('from_server', {'cmd': 'updatereppen', 'data': vars.rep_pen})
emit('from_server', {'cmd': 'updateoutlen', 'data': vars.genamt})
emit('from_server', {'cmd': 'updatanotedepth', 'data': vars.andepth})
#==================================================================#
# Sets the logical and display states for the AI Busy condition
@ -521,6 +595,7 @@ def togglememorymode():
vars.mode = "memory"
emit('from_server', {'cmd': 'memmode', 'data': 'true'})
emit('from_server', {'cmd': 'setinputtext', 'data': vars.memory})
emit('from_server', {'cmd': 'setanote', 'data': vars.authornote})
elif(vars.mode == "memory"):
vars.mode = "play"
emit('from_server', {'cmd': 'memmode', 'data': 'false'})
@ -534,6 +609,17 @@ def memsubmit(data):
vars.memory = data
vars.mode = "play"
emit('from_server', {'cmd': 'memmode', 'data': 'false'})
# Ask for contents of Author's Note field
emit('from_server', {'cmd': 'getanote', 'data': ''})
#==================================================================#
# Commit changes to Author's Note
#==================================================================#
def anotesubmit(data):
# Maybe check for length at some point
# For now just send it to storage
vars.authornote = data
#==================================================================#
# Assembles game data into a request to InferKit API
@ -611,11 +697,12 @@ def saveRequest():
js = {}
#js["maxlegth"] = vars.max_length # This causes problems when switching to/from InfraKit
#js["genamt"] = vars.genamt
js["rep_pen"] = vars.rep_pen
js["temp"] = vars.temp
#js["rep_pen"] = vars.rep_pen
#js["temp"] = vars.temp
js["gamestarted"] = vars.gamestarted
js["prompt"] = vars.prompt
js["memory"] = vars.memory
js["authorsnote"] = vars.authornote
js["actions"] = vars.actions
js["savedir"] = path
# Write it
@ -638,13 +725,18 @@ def loadRequest():
# Copy file contents to vars
#vars.max_length = js["maxlegth"] # This causes problems when switching to/from InfraKit
#vars.genamt = js["genamt"]
vars.rep_pen = js["rep_pen"]
vars.temp = js["temp"]
#vars.rep_pen = js["rep_pen"]
#vars.temp = js["temp"]
vars.gamestarted = js["gamestarted"]
vars.prompt = js["prompt"]
vars.memory = js["memory"]
vars.actions = js["actions"]
vars.savedir = js["savedir"]
# Try not to break older save files
if("authorsnote" in js):
vars.authornote = js["authorsnote"]
file.close()
# Refresh game screen
refresh_story()