A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 17, 00:30 UTC
[misc] remove comments that were supposed to be removed (#4977)
SangBin Cho committed
May 22, 2024 at 22:02 UTC
c74c913bfbefc5d7a1302557eb35cdcbecd91f67
2 files changed
-3
tests/lora/conftest.py
-1
@@ -185,7 +185,6 @@ def long_context_lora_files_32k():
185
return snapshot_download(repo_id="SangBinCho/long_context_32k_testing")
186
187
188
-# SANG-TODO Download long lora files.
188
@pytest.fixture(scope="session")
189
def long_context_infos(long_context_lora_files_16k_1,
190
long_context_lora_files_16k_2,
vllm/lora/models.py
-2
@@ -105,8 +105,6 @@ def convert_mapping(
105
lora_offset: int = long_lora_context.offsets_by_lora_id.get(
106
index_mapping_indices[i], 0)
107
long_lora_offsets[i] = lora_offset
108
- # SANG-TODO
109
- # index_mapping_indices[i] = i
108
109
indices_list: List[Union[List[int], torch.Tensor]] = [
110
index_mapping_indices, lora_indices, embedding_indices