Skip to content

GitLab

Unverified Commit 42488dae authored by Roger Wang's avatar Roger Wang Committed by GitHub
Browse files

[Bugfix] Fix dummy video number of frames calculation (#25553) 


Signed-off-by: default avatarRoger Wang <hey@rogerw.io>
parent b67dece2
Hide whitespace changes
Inline Side-by-side
Showing with 2 additions and 2 deletions
vllm/v1/worker/gpu_model_runner.py
View file @ 42488dae
...@@ -2807,7 +2807,7 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin): ...@@ -2807,7 +2807,7 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
dummy_decoder_data = self.mm_registry.get_decoder_dummy_data( dummy_decoder_data = self.mm_registry.get_decoder_dummy_data(
model_config=self.model_config, model_config=self.model_config,
seq_len=self.max_num_tokens, seq_len=self.max_model_len,
mm_counts={modality: 1}, mm_counts={modality: 1},
cache=self.mm_budget.cache, cache=self.mm_budget.cache,
) )
......
vllm/v1/worker/tpu_model_runner.py
View file @ 42488dae
...@@ -1795,7 +1795,7 @@ class TPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin): ...@@ -1795,7 +1795,7 @@ class TPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
dummy_decoder_data = self.mm_registry.get_decoder_dummy_data( dummy_decoder_data = self.mm_registry.get_decoder_dummy_data(
model_config=self.model_config, model_config=self.model_config,
seq_len=self.max_num_tokens, seq_len=self.max_model_len,
mm_counts={modality: 1}, mm_counts={modality: 1},
cache=self.mm_budget.cache, cache=self.mm_budget.cache,
) )
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment