vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fix for lite (+5 squashed commit)

Browse source 
Squashed commit:

[f9ce9855] catch more exceptions

[8cdaf149] tweaked horde worker timeouts, updated lite

[619ebef4] fixed abort no response if failed

[a54a66a2] fixed time overflow

[9affdc3e] updated lite
This commit is contained in:
Concedo 2023-10-15 16:50:03 +08:00
parent ee0681f0d9
commit 6f8fe88f10
2 changed files with 481 additions and 455 deletions
Download patch file Download diff file Expand all files Collapse all files

901
klite.embd
View file

File diff suppressed because one or more lines are too long

35
koboldcpp.py
View file

@ -679,7 +679,7 @@ class ServerRequestHandler(http.server.SimpleHTTPRequestHandler):
self.end_headers(content_type='application/json')
self.wfile.write(json.dumps({"value": count}).encode())
except ValueError as e:
except Exception as e:
utfprint("Count Tokens - Body Error: " + str(e))
self.send_response(400)
self.end_headers(content_type='application/json')
@ -690,19 +690,22 @@ class ServerRequestHandler(http.server.SimpleHTTPRequestHandler):
multiuserkey = ""
try:
tempbody = json.loads(body)
multiuserkey = tempbody.get('genkey', "")
except ValueError as e:
if isinstance(tempbody, dict):
multiuserkey = tempbody.get('genkey', "")
except Exception as e:
multiuserkey = ""
pass
if (multiuserkey=="" and requestsinqueue==0) or (multiuserkey!="" and multiuserkey==currentusergenkey):
ag = handle.abort_generate()
time.sleep(0.3) #short delay before replying
time.sleep(0.1) #short delay before replying
self.send_response(200)
self.end_headers(content_type='application/json')
self.wfile.write(json.dumps({"success": ("true" if ag else "false")}).encode())
print("\nGeneration Aborted")
else:
self.send_response(200)
self.end_headers(content_type='application/json')
self.wfile.write(json.dumps({"success": "false"}).encode())
return
@ -711,8 +714,9 @@ class ServerRequestHandler(http.server.SimpleHTTPRequestHandler):
multiuserkey = ""
try:
tempbody = json.loads(body)
multiuserkey = tempbody.get('genkey', "")
except ValueError as e:
if isinstance(tempbody, dict):
multiuserkey = tempbody.get('genkey', "")
except Exception as e:
multiuserkey = ""
pass
@ -765,7 +769,7 @@ class ServerRequestHandler(http.server.SimpleHTTPRequestHandler):
genparams = None
try:
genparams = json.loads(body)
except ValueError as e:
except Exception as e:
utfprint("Body Err: " + str(body))
return self.send_response(503)
@ -1501,11 +1505,11 @@ def run_horde_worker(args, api_key, worker_name):
session_jobs += 1
curtime = datetime.now()
elapsedtime=curtime-sessionstart
hrs = elapsedtime.seconds // 3600
hrs = int(elapsedtime.total_seconds()) // 3600
mins = elapsedtime.seconds // 60 % 60
secs = elapsedtime.seconds % 60
elapsedtimestr = f"{hrs:03d}h:{mins:02d}m:{secs:02d}s"
earnrate = session_kudos_earned/(elapsedtime.seconds/3600)
earnrate = session_kudos_earned/(elapsedtime.total_seconds()/3600)
print_with_time(f'Submitted {jobid} and earned {reward:.0f} kudos\n[Total:{session_kudos_earned:.0f} kudos, Time:{elapsedtimestr}, Jobs:{session_jobs}, EarnRate:{earnrate:.0f} kudos/hr]')
rewardcounter += 1
if rewardcounter > 50:
@ -1554,15 +1558,16 @@ def run_horde_worker(args, api_key, worker_name):
print_with_time(f"Embedded Horde Worker '{worker_name}' is started.")
break
while exitcounter < 35:
while exitcounter < 40:
currentjob_attempts = 0
current_generation = None
if punishcounter >= 10:
if punishcounter >= 8:
punishcounter = 0
print_with_time(f"Horde Worker Paused for 10 min - Too many errors. It will resume automatically.")
penaltymult = (1 + (exitcounter//10))
print_with_time(f"Horde Worker Paused for {penaltymult*10} min - Too many errors. It will resume automatically, but you should restart it.")
print_with_time(f"Caution: Too many failed jobs may lead to entering maintenance mode.")
time.sleep(600)
time.sleep(600 * penaltymult)
#first, make sure we are not generating
if modelbusy.locked():
@ -1583,8 +1588,8 @@ def run_horde_worker(args, api_key, worker_name):
if not pop:
exitcounter += 1
punishcounter += 1
print_with_time(f"Failed to fetch job from {cluster}. Waiting 5 seconds...")
time.sleep(5)
print_with_time(f"Failed to fetch job from {cluster}. Waiting 10 seconds...")
time.sleep(10)
continue
if not pop["id"]:
slp = (1 if sleepy_counter<10 else (2 if sleepy_counter<25 else 3))