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

server : fix format_infill (#10724)

Browse source 
* server : fix format_infill

* fix

* rename

* update test

* use another model

* update test

* update test

* test_invalid_input_extra_req
This commit is contained in:
Xuan Son Nguyen 2024-12-08 23:04:29 +01:00 • committed by GitHub
parent e52522b869
commit ce8784bdb1
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 53 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

36
examples/server/tests/unit/test_infill.py
View file

@ -13,28 +13,28 @@ def test_infill_without_input_extra():
global server
server.start()
res = server.make_request("POST", "/infill", data={
"prompt": "Complete this",
"input_prefix": "#include <cstdio>\n#include \"llama.h\"\n\nint main() {\n int n_threads = llama_",
"input_prefix": "#include <cstdio>\n#include \"llama.h\"\n\nint main() {\n",
"prompt": " int n_threads = llama_",
"input_suffix": "}\n",
})
assert res.status_code == 200
assert match_regex("(One|day|she|saw|big|scary|bird)+", res.body["content"])
assert match_regex("(Ann|small|shiny)+", res.body["content"])
def test_infill_with_input_extra():
global server
server.start()
res = server.make_request("POST", "/infill", data={
"prompt": "Complete this",
"input_extra": [{
"filename": "llama.h",
"text": "LLAMA_API int32_t llama_n_threads();\n"
}],
"input_prefix": "#include <cstdio>\n#include \"llama.h\"\n\nint main() {\n int n_threads = llama_",
"input_prefix": "#include <cstdio>\n#include \"llama.h\"\n\nint main() {\n",
"prompt": " int n_threads = llama_",
"input_suffix": "}\n",
})
assert res.status_code == 200
assert match_regex("(cuts|Jimmy|mom|came|into|the|room)+", res.body["content"])
assert match_regex("(Dad|excited|park)+", res.body["content"])
@pytest.mark.parametrize("input_extra", [
@ -48,10 +48,30 @@ def test_invalid_input_extra_req(input_extra):
global server
server.start()
res = server.make_request("POST", "/infill", data={
"prompt": "Complete this",
"input_extra": [input_extra],
"input_prefix": "#include <cstdio>\n#include \"llama.h\"\n\nint main() {\n int n_threads = llama_",
"input_prefix": "#include <cstdio>\n#include \"llama.h\"\n\nint main() {\n",
"prompt": " int n_threads = llama_",
"input_suffix": "}\n",
})
assert res.status_code == 400
assert "error" in res.body
@pytest.mark.skipif(not is_slow_test_allowed(), reason="skipping slow test")
def test_with_qwen_model():
global server
server.model_file = None
server.model_hf_repo = "ggml-org/Qwen2.5-Coder-1.5B-IQ3_XXS-GGUF"
server.model_hf_file = "qwen2.5-coder-1.5b-iq3_xxs-imat.gguf"
server.start(timeout_seconds=600)
res = server.make_request("POST", "/infill", data={
"input_extra": [{
"filename": "llama.h",
"text": "LLAMA_API int32_t llama_n_threads();\n"
}],
"input_prefix": "#include <cstdio>\n#include \"llama.h\"\n\nint main() {\n",
"prompt": " int n_threads = llama_",
"input_suffix": "}\n",
})
assert res.status_code == 200
assert res.body["content"] == "n_threads();\n printf(\"Number of threads: %d\\n\", n_threads);\n return 0;\n"

3
examples/server/tests/utils.py
View file

@ -371,3 +371,6 @@ def match_regex(regex: str, text: str) -> bool:
).search(text)
is not None
)
def is_slow_test_allowed():
return os.environ.get("SLOW_TESTS") == "1" or os.environ.get("SLOW_TESTS") == "ON"