Skip to content

[CI/Build][LoRA] Temporarily fix long context failure issue #9579

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Oct 22, 2024
Merged

[CI/Build][LoRA] Temporarily fix long context failure issue #9579

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
07c415d
Temporarily fix the test failure issue
jeejeelee Oct 22, 2024
e59c8e4
Add fixme
jeejeelee Oct 22, 2024
d42aac8
Fix format
jeejeelee Oct 22, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
31 changes: 20 additions & 11 deletions tests/lora/test_long_context.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,9 +28,15 @@
def _create_lora_request(lora_id, long_context_infos):
context_len = long_context_infos[lora_id]["context_length"]
scaling_factor = context_len_to_scaling_factor[context_len]
return LoRARequest(context_len, lora_id,
long_context_infos[lora_id]["lora"], None,
4096 * scaling_factor)
return LoRARequest(
# There are 2 LoRAs for 16K, we need to add lora_id to indicate
# they are different LoRAs.
context_len + str(lora_id),
lora_id,
long_context_infos[lora_id]["lora"],
None,
4096 * scaling_factor,
)


def evaluate_json_response(model_response, golden_response):
Expand Down Expand Up @@ -108,14 +114,17 @@ def lora_llm(long_context_infos):
for info in long_context_infos.values()
]

llm = vllm.LLM("meta-llama/Llama-2-13b-chat-hf",
enable_lora=True,
max_num_seqs=16,
max_loras=2,
long_lora_scaling_factors=tuple(scaling_factors),
max_num_batched_tokens=4096 * 8,
tensor_parallel_size=4,
distributed_executor_backend="mp")
llm = vllm.LLM(
"meta-llama/Llama-2-13b-chat-hf",
enable_lora=True,
max_num_seqs=16,
max_loras=2,
long_lora_scaling_factors=tuple(scaling_factors),
max_num_batched_tokens=4096 * 8,
tensor_parallel_size=4,
# FIXME enable async output processor
disable_async_output_proc=True,
distributed_executor_backend="mp")
yield llm
del llm

Expand Down