vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-10 09:04:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

server : disable context shift by default (#15416)

Browse source 
* server : disable context shift by default

ggml-ci

* server : make scopr of test parameters local
This commit is contained in:
Georgi Gerganov 2025-08-19 16:46:37 +03:00 • committed by GitHub
parent a6d3cfe7fa
commit d2fcd91cf9
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
16 changed files with 27 additions and 20 deletions
Download patch file Download diff file Expand all files Collapse all files

6
tools/server/tests/unit/test_ctx_shift.py
View file

@ -11,7 +11,7 @@ Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu
Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
""".strip()
@pytest.fixture(scope="module", autouse=True)
@pytest.fixture(autouse=True)
def create_server():
global server
server = ServerPreset.tinyllama2()
@ -25,6 +25,7 @@ def test_ctx_shift_enabled():
# the prompt is truncated to keep the last 109 tokens
# 64 tokens are generated thanks to shifting the context when it gets full
global server
server.enable_ctx_shift = True
server.start()
res = server.make_request("POST", "/completion", data={
"n_predict": 64,
@ -42,7 +43,6 @@ def test_ctx_shift_enabled():
])
def test_ctx_shift_disabled_short_prompt(n_predict: int, n_token_output: int, truncated: bool):
global server
server.disable_ctx_shift = True
server.n_predict = -1
server.start()
res = server.make_request("POST", "/completion", data={
@ -56,7 +56,6 @@ def test_ctx_shift_disabled_short_prompt(n_predict: int, n_token_output: int, tr
def test_ctx_shift_disabled_long_prompt():
global server
server.disable_ctx_shift = True
server.start()
res = server.make_request("POST", "/completion", data={
"n_predict": 64,
@ -68,7 +67,6 @@ def test_ctx_shift_disabled_long_prompt():
def test_ctx_shift_disabled_stream():
global server
server.disable_ctx_shift = True
server.start()
res = server.make_stream_request("POST", "/v1/completions", data={
"n_predict": 256,