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

server branch update with graphics

Browse source
This commit is contained in:
pudepiedj 2024-02-18 12:08:59 +00:00
parent 168a68395c
commit aed7507eb8
3 changed files with 669 additions and 1116 deletions
Download patch file Download diff file Expand all files Collapse all files

18
Llamaserver.py
View file

@ -5,6 +5,7 @@ import json
from time import sleep from time import sleep
def print_dict(data): def print_dict(data):
if isinstance(data, dict):
for k, v in data.items(): for k, v in data.items():
if isinstance(v, dict): if isinstance(v, dict):
print_dict(v) print_dict(v)
@ -24,7 +25,7 @@ def make_empty_bar(num_requests):
bar.append("\u2589") bar.append("\u2589")
bar = ' '.join(bar) bar = ' '.join(bar)
bar = bar.replace(' ','') bar = bar.replace(' ','')
print(f"Bar is now {bar}.") print(f"Bar is now {bar}.\n")
return bar return bar
def make_progress_bar(bar, count, num_requests): def make_progress_bar(bar, count, num_requests):
@ -32,9 +33,9 @@ def make_progress_bar(bar, count, num_requests):
stride2 = len("\u23F1") stride2 = len("\u23F1")
for i in range(num_requests): for i in range(num_requests):
if i == count: if i == count:
print(f"Bar position {i} is {bar[i]}") # print(f"Bar position {i} is {bar[i]}\n")
bar = bar[:i*stride1] + "\u23F1" + bar[i*stride1 + stride2:] bar = bar[:i*stride1] + "\u23F1" + bar[i*stride1 + stride2:]
print(f"Bar is now {bar}") print(f"Bar is now {bar}\n")
return bar return bar
def send_request(q, question, event, count, num_requests): def send_request(q, question, event, count, num_requests):
@ -49,9 +50,16 @@ def send_request(q, question, event, count, num_requests):
response = requests.post(url, headers=headers, json=data) response = requests.post(url, headers=headers, json=data)
if response.status_code in [200,300]: if response.status_code in [200,300]:
print(f"Current Queue Size: {q.qsize()}; processing request {count} / {num_requests}\n") print(f"Current Queue Size: {q.qsize()}; processing request {count} / {num_requests}\n")
print(f"Status Code for {question}: {response.status_code}") print(f"Status Code for {question}: {response.status_code}\n")
print(f"Response to {question}:\n") print(f"Response to {question}:\n")
print_dict(json.loads(response.text)) if isinstance(response.text, str):
data = json.loads(response.text)
if isinstance(data, dict):
print_dict(data)
elif isinstance(data, str):
print(data)
else:
print("\nServer returned data of wrong type.\n")
# put the response text in the queue # put the response text in the queue
q.put(response.text) q.put(response.text)
if not q.empty(): if not q.empty():

1
examples/CMakeLists.txt
View file

@ -17,7 +17,6 @@ else()
add_subdirectory(batched-bench) add_subdirectory(batched-bench)
add_subdirectory(beam-search) add_subdirectory(beam-search)
add_subdirectory(benchmark) add_subdirectory(benchmark)
add_subdirectory(cmap-example)
add_subdirectory(convert-llama2c-to-ggml) add_subdirectory(convert-llama2c-to-ggml)
add_subdirectory(embedding) add_subdirectory(embedding)
add_subdirectory(finetune) add_subdirectory(finetune)

1394
examples/server/server.cpp
View file

File diff suppressed because it is too large Load diff