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

[Core] Refactor model loading code (#4097)

Browse Source
This commit is contained in:
Antoni Baum
2024-04-16 11:34:39 -07:00
committed by GitHub
parent 05434764cd
commit 69e1d2fb69
67 changed files with 1054 additions and 963 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
tests/samplers/test_sampler.py
View File

@@ -32,7 +32,12 @@ def _prepare_test(
1e-2,
dtype=input_tensor.dtype)
sampler = MockLogitsSampler(fake_logits)
model_runner = ModelRunner(None, None, None, None, None)
model_runner = ModelRunner(model_config=None,
parallel_config=None,
scheduler_config=None,
device_config=None,
load_config=None,
lora_config=None)
return input_tensor, fake_logits, sampler, model_runner
@@ -591,7 +596,12 @@ def test_sampler_top_k_top_p(seed: int, device: str):
device=input_tensor.device,
dtype=input_tensor.dtype)
sampler = MockLogitsSampler(fake_logits)
model_runner = ModelRunner(None, None, None, None, None)
model_runner = ModelRunner(model_config=None,
parallel_config=None,
scheduler_config=None,
device_config=None,
load_config=None,
lora_config=None)
generation_model = GenerationMixin()
generation_config = GenerationConfig(top_k=top_k,