Repository navigation
Expand file tree
/
Copy pathtolerances.py
More file actions
36 lines (33 loc) · 1.74 KB
/
Copy pathtolerances.py
File metadata and controls
36 lines (33 loc) · 1.74 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import torch
ToleranceSpec = dict[torch.dtype, dict[str, float]]
# Default tolerances for comparing IR op implementations against native.
# These are intentionally conservative (permissive) to avoid false failures
# across different hardware and kernel implementations. Ops that need tighter
# or looser bounds should use override_tolerance.
DEFAULT_TOLERANCES: ToleranceSpec = {
# 52-bit mantissa; machine epsilon ~1.1e-16
torch.float64: {"atol": 1e-8, "rtol": 1e-8},
# 23-bit mantissa; machine epsilon ~1.2e-7.
# Values from PyTorch test_transformers.py reference defaults.
torch.float32: {"atol": 1e-5, "rtol": 1.3e-6},
# 10-bit mantissa; machine epsilon ~9.8e-4.
# Standard tolerance used across vLLM kernel tests.
torch.float16: {"atol": 1e-3, "rtol": 1e-3},
# 7-bit mantissa; machine epsilon ~7.8e-3.
# Wider rtol than float16 to account for the coarser mantissa.
torch.bfloat16: {"atol": 1e-3, "rtol": 1.6e-2},
# 3-bit mantissa; machine epsilon ~6.25e-2.
# Derived from vLLM fp8 kernel tests (merge_attn_states, silu_mul_fp8).
torch.float8_e4m3fn: {"atol": 1e-1, "rtol": 1e-1},
# 2-bit mantissa; machine epsilon ~1.25e-1.
# Wider than e4m3fn due to the smaller mantissa.
torch.float8_e5m2: {"atol": 2e-1, "rtol": 2e-1},
# 1-bit mantissa; machine epsilon ~2.5e-1. Packed pair format (x2).
# Derived from vLLM fp4 tests (test_silu_mul_nvfp4_quant: atol=3e-1).
torch.float4_e2m1fn_x2: {"atol": 3e-1, "rtol": 3e-1},
# Integer quantized; off-by-one from rounding is expected.
# rtol=0 because relative error is meaningless for small integers.
torch.int8: {"atol": 1, "rtol": 0},
}