[Minor] Various small code cleanups/simplifications (#31508)
Signed-off-by: njhill <nickhill123@gmail.com>
This commit is contained in:
@@ -829,7 +829,7 @@ class InputBatch:
|
||||
presence_penalties=self.presence_penalties[:num_reqs],
|
||||
repetition_penalties=self.repetition_penalties[:num_reqs],
|
||||
output_token_ids=output_token_ids,
|
||||
spec_token_ids=cast(list[list[int]], self.spec_token_ids),
|
||||
spec_token_ids=self.spec_token_ids,
|
||||
no_penalties=self.no_penalties,
|
||||
allowed_token_ids_mask=allowed_token_ids_mask,
|
||||
bad_words_token_ids=self.bad_words_token_ids,
|
||||
|
||||
@@ -1026,7 +1026,7 @@ class GPUModelRunner(
|
||||
each sequence, and a shifting is done during the next iteration
|
||||
based on the number of accepted tokens.
|
||||
"""
|
||||
if not self.model_config.is_hybrid or not self.speculative_config:
|
||||
if not self.speculative_config or not self.model_config.is_hybrid:
|
||||
return
|
||||
|
||||
# Find the number of accepted tokens for each sequence.
|
||||
|
||||
Reference in New Issue
Block a user