xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-07-11 05:17:09 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix set forward context error

Browse Source 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
yewentao256 2025-08-11 14:07:29 +00:00
parent 28e7c30b01
commit 44ead56ad5
1 changed files with 6 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
vllm/v1/worker/gpu_model_runner.py
View File

@ -2031,7 +2031,12 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
# Run the model. # Run the model.
# Use persistent buffers for CUDA graphs. # Use persistent buffers for CUDA graphs.
self.maybe_setup_kv_connector(scheduler_output) with set_forward_context(attn_metadata,
vllm_config=self.vllm_config,
num_tokens=num_input_tokens or 1,
num_tokens_across_dp=num_tokens_after_padding,
skip_cuda_graphs=skip_cuda_graphs):
self.maybe_setup_kv_connector(scheduler_output)
model_output = self._run_model( model_output = self._run_model(
attn_metadata=attn_metadata, attn_metadata=attn_metadata,
num_scheduled_tokens=num_input_tokens, num_scheduled_tokens=num_input_tokens,