biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[ci/lint] Add back default arg for pre-commit (#12279)

Browse Source 
Signed-off-by: kevin <kevin@anyscale.com>
This commit is contained in:
Kevin H. Luu
2025-01-21 17:15:27 -08:00
committed by GitHub
parent df76e5af26
commit 64ea24d0b3
6 changed files with 26 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
tests/models/decoder_only/language/test_gguf.py
View File

@@ -74,11 +74,7 @@ DOLPHIN_CONFIG = GGUFTestConfig(
)
MODELS = [
LLAMA_CONFIG,
QWEN2_CONFIG,
PHI3_CONFIG,
GPT2_CONFIG,
STABLELM_CONFIG,
LLAMA_CONFIG, QWEN2_CONFIG, PHI3_CONFIG, GPT2_CONFIG, STABLELM_CONFIG,
DOLPHIN_CONFIG
# STARCODER_CONFIG, # broken
]
@@ -114,11 +110,12 @@ def test_models(
messages, tokenize=False, add_generation_prompt=True)
# Run unquantized model.
with vllm_runner(model_name=model.original_model,
enforce_eager=True, # faster tests
dtype=dtype,
max_model_len=MAX_MODEL_LEN,
tensor_parallel_size=tp_size) as original_model:
with vllm_runner(
model_name=model.original_model,
enforce_eager=True, # faster tests
dtype=dtype,
max_model_len=MAX_MODEL_LEN,
tensor_parallel_size=tp_size) as original_model:
original_outputs = original_model.generate_greedy_logprobs(
example_prompts[:-1], max_tokens, num_logprobs)