Skip to content
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

Jump to bottom

[Bugfix] Fix InternVL2 vision embeddings process with pipeline parallel #8299

Merged
merged 6 commits into from
Sep 11, 2024
Merged

[Bugfix] Fix InternVL2 vision embeddings process with pipeline parallel #8299

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
b456676
fix internvl pipeline parallel
Isotr0py Sep 9, 2024
ce5dfe5
extend internvl distributed test
Isotr0py Sep 9, 2024
30dc7e5
disable CHUNKED_PREFILL for InternVL2-4B
Isotr0py Sep 9, 2024
a42c99e
Merge branch 'vllm-project:main' into fix-internvl-pp
Isotr0py Sep 10, 2024
53a42c3
limit max model len
Isotr0py Sep 10, 2024
9cb6bf2
align dtype
Isotr0py Sep 10, 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
10 changes: 8 additions & 2 deletions tests/distributed/test_pipeline_parallel.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,7 +32,9 @@
(1, 4, 1, 0, 0, "meta-llama/Meta-Llama-3-8B", "ray"),
(2, 2, 1, 0, 0, "meta-llama/Meta-Llama-3-8B", "ray"),
(2, 2, 0, 1, 0, "meta-llama/Meta-Llama-3-8B", "ray"),
(2, 2, 1, 1, 1, "internlm/internlm2_5-7b-chat", "ray"),
(1, 2, 1, 1, 1, "OpenGVLab/InternVL2-1B", "ray"),
(1, 2, 1, 1, 1, "OpenGVLab/InternVL2-2B", "ray"),
(1, 2, 1, 0, 1, "OpenGVLab/InternVL2-4B", "ray"),
],
)
@fork_new_process_for_each_test
Expand All @@ -46,6 +48,8 @@ def test_compare_tp(TP_SIZE, PP_SIZE, EAGER_MODE, CHUNKED_PREFILL,
# use half precision for speed and memory savings in CI environment
"--dtype",
"float16",
"--max-model-len",
"8192",
"--pipeline-parallel-size",
str(PP_SIZE),
"--tensor-parallel-size",
Expand All @@ -62,7 +66,9 @@ def test_compare_tp(TP_SIZE, PP_SIZE, EAGER_MODE, CHUNKED_PREFILL,
tp_args = [
# use half precision for speed and memory savings in CI environment
"--dtype",
"bfloat16",
"float16",
"--max-model-len",
"8192",
"--tensor-parallel-size",
str(max(TP_SIZE, 2)), # We only use 2 GPUs in the CI.
"--distributed-executor-backend",
Expand Down
3 changes: 2 additions & 1 deletion vllm/model_executor/models/internvl.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,7 @@

from vllm.attention import AttentionMetadata
from vllm.config import CacheConfig, MultiModalConfig
from vllm.distributed import get_pp_group
from vllm.inputs import INPUT_REGISTRY, InputContext, LLMInputs
from vllm.model_executor.layers.quantization import QuantizationConfig
from vllm.model_executor.layers.sampler import SamplerOutput
Expand Down Expand Up @@ -480,7 +481,7 @@ def forward(
**kwargs: object,
) -> SamplerOutput:
image_input = self._parse_and_validate_image_input(**kwargs)
if image_input is not None:
if image_input is not None and get_pp_group().is_first_rank:
inputs_embeds = self.language_model.model.get_input_embeddings(
input_ids)
vision_embeddings = self._process_image_input(image_input)
Expand Down
Loading