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.
16 lines
486 B
Python
16 lines
486 B
Python
import logging
|
|
|
|
|
|
class NewLineFormatter(logging.Formatter):
|
|
"""Adds logging prefix to newlines to align multi-line messages."""
|
|
|
|
def __init__(self, fmt, datefmt=None, style="%"):
|
|
logging.Formatter.__init__(self, fmt, datefmt, style)
|
|
|
|
def format(self, record):
|
|
msg = logging.Formatter.format(self, record)
|
|
if record.message != "":
|
|
parts = msg.split(record.message)
|
|
msg = msg.replace("\n", "\r\n" + parts[0])
|
|
return msg
|