Skip to content

GitLab

Commit 4b752578 authored by jujl1's avatar jujl1
Browse files

fix: 修复deepseek在PP>1开chunkprefill精度问题

parent 03ca39bd
Hide whitespace changes
Inline Side-by-side
Showing with 6 additions and 0 deletions
vllm/v1/worker/gpu_model_runner.py
View file @ 4b752578
...@@ -1638,6 +1638,9 @@ class GPUModelRunnerBase(LoRAModelRunnerMixin): ...@@ -1638,6 +1638,9 @@ class GPUModelRunnerBase(LoRAModelRunnerMixin):
# Mask out the sampled tokens that should not be sampled. # Mask out the sampled tokens that should not be sampled.
for i in discard_sampled_tokens_req_indices: for i in discard_sampled_tokens_req_indices:
valid_sampled_token_ids[i].clear() valid_sampled_token_ids[i].clear()
if spec_token_ids is not None:
for i in discard_sampled_tokens_req_indices:
spec_token_ids[i].clear()
# Cache the sampled tokens in the model runner, so that the scheduler # Cache the sampled tokens in the model runner, so that the scheduler
# doesn't need to send them back. # doesn't need to send them back.
...@@ -3519,6 +3522,9 @@ class GPUModelRunnerMTP(GPUModelRunnerBase): ...@@ -3519,6 +3522,9 @@ class GPUModelRunnerMTP(GPUModelRunnerBase):
# Mask out the sampled tokens that should not be sampled. # Mask out the sampled tokens that should not be sampled.
for i in discard_sampled_tokens_req_indices: for i in discard_sampled_tokens_req_indices:
valid_sampled_token_ids[i].clear() valid_sampled_token_ids[i].clear()
if spec_token_ids is not None:
for i in discard_sampled_tokens_req_indices:
spec_token_ids[i].clear()
# Cache the sampled tokens in the model runner, so that the scheduler # Cache the sampled tokens in the model runner, so that the scheduler
# doesn't need to send them back. # doesn't need to send them back.
......
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