[Core][2/N] Model runner refactoring part 2. Combine prepare prefill / decode to a single API (#4681)

This PR combines prepare_prompt and prepare_decode into a single API. This PR also coelsce the attn metadata for prefill/decode to a single class and allow to slice them when running attn backend.

It also refactors subquery_start_loc which was not refactored in the previous PR
This commit is contained in:
SangBin Cho
2024-05-15 14:00:10 +09:00
committed by GitHub
parent 8a7cc254a0
commit 65bf2ac165
18 changed files with 781 additions and 734 deletions
+1
View File
@@ -114,6 +114,7 @@ class MultiStepWorker(Worker):
token_logprob = seq_output.logprobs[token_id]
seq.append_token_id(token_id, token_logprob.logprob)
seq.update_num_computed_tokens(1)
def _shallow_copy_inputs(
self, seq_group_metadata_list: List[SequenceGroupMetadata]