whn09 5a026d3aad Map media-free requests inline instead of through the multimodal worker (DEEPINFER_MULTI_MODAL_INLINE_TEXT=1) ... A text request to the VL model went through the multimodal worker's queue and became schedulable only at a later engine step's fetch(): on Neuron two ~0.36 s idle placeholder steps before every text prefill. vllm-sonnet 512 (prefill bucket 512): TTFT 2.42 -> 1.67 s, tokens unchanged. bench_neuron: BENCH_TRACE_STEPS. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>		2026-10-27 04:56:47 +08:00
..
processor	initial commit	2026-10-20 12:09:42 +08:00
__init__.py	initial commit	2026-10-20 12:09:42 +08:00
multimodal_processor.py	Map media-free requests inline instead of through the multimodal worker (DEEPINFER_MULTI_MODAL_INLINE_TEXT=1)	2026-10-27 04:56:47 +08:00
prepare_tensor_input_vl.py	initial commit	2026-10-20 12:09:42 +08:00
schema.py	initial commit	2026-10-20 12:09:42 +08:00