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.
40 lines
1.4 KiB
Python
40 lines
1.4 KiB
Python
from dataclasses import dataclass
|
|
from typing import Dict, List, Optional, TypedDict, Union
|
|
|
|
from pydantic import BaseModel
|
|
|
|
|
|
# These classes are deprecated, see SamplingParams
|
|
class LLMGuidedOptions(TypedDict, total=False):
|
|
guided_json: Union[Dict, BaseModel, str]
|
|
guided_regex: str
|
|
guided_choice: List[str]
|
|
guided_grammar: str
|
|
guided_decoding_backend: str
|
|
guided_whitespace_pattern: str
|
|
guided_json_object: bool
|
|
|
|
|
|
@dataclass
|
|
class GuidedDecodingRequest:
|
|
"""One of the fields will be used to retrieve the logit processor."""
|
|
guided_json: Optional[Union[Dict, BaseModel, str]] = None
|
|
guided_regex: Optional[str] = None
|
|
guided_choice: Optional[List[str]] = None
|
|
guided_grammar: Optional[str] = None
|
|
guided_decoding_backend: Optional[str] = None
|
|
guided_whitespace_pattern: Optional[str] = None
|
|
guided_json_object: Optional[bool] = None
|
|
|
|
def __post_init__(self):
|
|
"""Validate that some fields are mutually exclusive."""
|
|
guide_count = sum([
|
|
self.guided_json is not None, self.guided_regex is not None,
|
|
self.guided_choice is not None, self.guided_grammar is not None,
|
|
self.guided_json_object is not None
|
|
])
|
|
if guide_count > 1:
|
|
raise ValueError(
|
|
"You can only use one kind of guided decoding but multiple are "
|
|
f"specified: {self.__dict__}")
|