[DEPLOY] Complete submission: baseline + all optimizations

Browse Source

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.

...

This commit is contained in:

dylanyunlon

2026-07-30 16:06:20 +00:00

parent 3722503dee

commit ef6abf3dc7

920 changed files with 149563 additions and 0 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

BIN

vllm/model_executor/models/__pycache__/__init__.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/arctic.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/baichuan.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/bart.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/blip.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/blip2.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/bloom.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/chameleon.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/chatglm.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/clip.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/commandr.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/dbrx.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/decilm.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/deepseek.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/deepseek_v2.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/eagle.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/exaone.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/falcon.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/fuyu.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/gemma.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/gemma2.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/gemma2_embedding.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/glm4.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/glm4_vision_encoder.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/gpt2.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/gpt_bigcode.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/gpt_j.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/gpt_neox.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/granite.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/granitemoe.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/idefics2_vision_model.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/interfaces.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/interfaces_base.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/intern_vit.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/internlm2.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/internvl.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/jais.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/jamba.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/llama.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/llama_embedding.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/llava.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/llava_next.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/llava_next_video.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/llava_onevision.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/mamba.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/mamba_cache.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/medusa.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/minicpm.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/minicpm3.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/minicpmv.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/mixtral.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/mixtral_quant.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/mllama.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/mlp_speculator.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/module_mapping.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/molmo.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/mpt.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/nemotron.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/nvlm_d.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/olmo.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/olmoe.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/opt.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/orion.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/paligemma.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/persimmon.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/phi.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/phi3.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/phi3_small.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/phi3v.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/phimoe.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/pixtral.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/qwen.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/qwen2.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/qwen2_5_vl.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/qwen2_moe.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/qwen2_rm.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/qwen2_vl.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/qwen3.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/qwen3_moe.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/registry.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/siglip.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/solar.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/stablelm.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/starcoder2.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/ultravox.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/utils.cpython-310.pyc Normal file

Binary file not shown.

BIN

vllm/model_executor/models/__pycache__/xverse.cpython-310.pyc Normal file

Binary file not shown.