Skip to content

[None][feat] Add environment variable to force spec-dec number of accepted tokens #9371

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Nov 26, 2025
Merged

[None][feat] Add environment variable to force spec-dec number of accepted tokens #9371

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d6a4b99
Add environment variable to force EAGLE3 number of accepted tokens
achartier Nov 22, 2025
9514ac8
address review comments
achartier Nov 25, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 14 additions & 0 deletions tensorrt_llm/_torch/speculative/eagle3.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,9 +1,11 @@
import os
from dataclasses import dataclass, field
from typing import TYPE_CHECKING, List, Optional, Set

import torch
from torch import nn

from tensorrt_llm.logger import logger
from tensorrt_llm.mapping import Mapping

from ..attention_backend import AttentionMetadata
Expand Down Expand Up @@ -320,6 +322,15 @@ def __init__(self, spec_config: "EagleDecodingConfig", mapping: Mapping):
self.mapping = mapping
self.guided_decoder: Optional[CapturableGuidedDecoder] = None

env_value = os.environ.get("TLLM_EAGLE3_FORCE_NUM_ACCEPTED_TOKENS", "0")
try:
self.force_num_accepted_tokens = int(env_value)
except ValueError:
logger.warning(
f"TLLM_EAGLE3_FORCE_NUM_ACCEPTED_TOKENS must be a valid integer, "
f"got '{env_value}'. Using default value 0.")
self.force_num_accepted_tokens = 0

# Skip torch.compile for now since current Torch is not compatible with Triton 3.4
# @torch.compile(options={"max-autotune": True})
def forward(self, input_ids, position_ids, hidden_states, logits,
Expand Down Expand Up @@ -481,6 +492,9 @@ def sample_and_accept_draft_tokens(
num_accepted_tokens[num_contexts:] += torch.cumprod(
(draft_tokens == gen_target_tokens[:, :self.max_draft_len]).int(),
dim=-1).sum(1)
# Check for environment variable override
if self.force_num_accepted_tokens != 0:
num_accepted_tokens[num_contexts:] = self.force_num_accepted_tokens
return accepted_tokens, num_accepted_tokens

def draft_decoder(
Expand Down