ef6abf3dc7
Adds ALL files needed for Dockerfile build:
- qwen3_6_scripts/ (baseline patches + our optimizations)
- vllm/ (full vllm package)
- paged_attention_v2_pytorch.py (V2 with single-bmm optimization)
- Dockerfile + computility-run.yaml
Our optimizations vs baseline:
1. paged_attn.py: pre-gathered context KV (eliminates 194 gather calls),
Triton try/fallback, V2 heuristic, threshold 32K→64K
2. paged_attention_v2_pytorch.py: fills NotImplementedError,
single-bmm Phase 1 (195 launches → 3)
3. patch_enable_triton.py: HAS_TRITON=True with safety fallback
4. patch_triton_tuning.py: BLOCK=64, NUM_WARPS=4 for BI-V100
5. computility-run.yaml: gpu-memory-utilization 0.9→0.95,
max-num-batched-tokens 8192→16384
This repo can now be submitted to dev.modelhub.org.cn as-is.
52 lines
1.8 KiB
Python
52 lines
1.8 KiB
Python
'''
|
|
Worker-related helper functions.
|
|
'''
|
|
|
|
from vllm.utils import STR_NOT_IMPL_ENC_DEC_ERR_STRS
|
|
from vllm.worker.model_runner import GPUModelRunnerBase
|
|
|
|
|
|
def assert_enc_dec_mr_supported_scenario(
|
|
enc_dec_mr: GPUModelRunnerBase) -> None:
|
|
'''
|
|
Asserted that the provided encoder/decoder model runner instance reflects
|
|
a supported scenario.
|
|
'''
|
|
|
|
# Reminder: Please update docs/source/serving/compatibility_matrix.rst
|
|
# If the feature combo become valid
|
|
|
|
if enc_dec_mr.cache_config.enable_prefix_caching:
|
|
raise NotImplementedError(
|
|
STR_NOT_IMPL_ENC_DEC_ERR_STRS['STR_NOT_IMPL_ENC_DEC_PREFIX_CACHE'])
|
|
|
|
if enc_dec_mr.sliding_window is not None:
|
|
raise NotImplementedError(
|
|
STR_NOT_IMPL_ENC_DEC_ERR_STRS['STR_NOT_IMPL_ENC_DEC_SWA'])
|
|
|
|
if enc_dec_mr.scheduler_config.chunked_prefill_enabled:
|
|
raise NotImplementedError(STR_NOT_IMPL_ENC_DEC_ERR_STRS[
|
|
'STR_NOT_IMPL_ENC_DEC_CHUNKED_PREFILL'])
|
|
|
|
if getattr(enc_dec_mr.model_config.hf_config, 'attn_logit_softcapping',
|
|
None) is not None:
|
|
raise NotImplementedError(
|
|
STR_NOT_IMPL_ENC_DEC_ERR_STRS['STR_NOT_IMPL_ENC_DEC_LOGIT_SOFTCAP']
|
|
)
|
|
|
|
if enc_dec_mr.lora_config is not None:
|
|
raise NotImplementedError(
|
|
STR_NOT_IMPL_ENC_DEC_ERR_STRS['STR_NOT_IMPL_ENC_DEC_LORA'])
|
|
|
|
if enc_dec_mr.parallel_config.pipeline_parallel_size > 1:
|
|
raise NotImplementedError(
|
|
STR_NOT_IMPL_ENC_DEC_ERR_STRS['STR_NOT_IMPL_ENC_DEC_PP'])
|
|
|
|
if enc_dec_mr.scheduler_config.num_lookahead_slots > 0:
|
|
raise NotImplementedError(
|
|
STR_NOT_IMPL_ENC_DEC_ERR_STRS['STR_NOT_IMPL_ENC_DEC_SPEC_DEC'])
|
|
|
|
if enc_dec_mr.prompt_adapter_config is not None:
|
|
raise NotImplementedError(STR_NOT_IMPL_ENC_DEC_ERR_STRS[
|
|
'STR_NOT_IMPL_ENC_DEC_PROMPT_ADAPTER'])
|