Skip to content
Merged
Show file tree
Hide file tree
Changes from 1 commit
Commits
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
174 changes: 174 additions & 0 deletions Cargo.lock

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

1 change: 1 addition & 0 deletions Cargo.toml
Original file line number Diff line number Diff line change
Expand Up @@ -46,6 +46,7 @@ dynamo-mocker = { path = "lib/mocker", version = "1.0.0" }
dynamo-kv-router = { path = "lib/kv-router", version = "1.0.0", features = ["metrics"] }
dynamo-async-openai = { path = "lib/async-openai", version = "1.0.0", features = ["byot"] }
dynamo-parsers = { path = "lib/parsers", version = "1.0.0" }
fastokens = { git = "https://github.com/Atero-ai/fastokens", version = "0.1.0" }

# kvbm
kvbm-common = { path = "lib/kvbm-common", version = "0.1.0" }
Expand Down
14 changes: 14 additions & 0 deletions components/src/dynamo/frontend/frontend_args.py
Original file line number Diff line number Diff line change
Expand Up @@ -76,6 +76,7 @@ class FrontendConfig(KvRouterConfigBase):
enable_streaming_tool_dispatch: bool
enable_streaming_reasoning_dispatch: bool
preprocess_workers: int
tokenizer_backend: str

def validate(self) -> None:
if bool(self.tls_cert_path) ^ bool(self.tls_key_path): # ^ is XOR
Expand Down Expand Up @@ -424,3 +425,16 @@ def add_arguments(self, parser) -> None:
),
arg_type=int,
)

add_argument(
g,
flag_name="--dyn-tokenizer-backend",
env_var="DYN_TOKENIZER_BACKEND",
default="default",
dest="tokenizer_backend",
help=(
"Tokenizer backend for BPE models: 'default' (HuggingFace tokenizers library) "
"or 'fasttokens' (fastokens crate for high-performance BPE encoding). "
"Decoding always uses HuggingFace. Has no effect on TikToken models."
),
)
2 changes: 2 additions & 0 deletions components/src/dynamo/frontend/main.py
Original file line number Diff line number Diff line change
Expand Up @@ -165,6 +165,8 @@ async def async_main():
config, vllm_flags, sglang_flags = parse_args()
dump_config(config.dump_config_to, config)
os.environ["DYN_EVENT_PLANE"] = config.event_plane
if config.tokenizer_backend == "fasttokens":
os.environ["DYN_TOKENIZER_BACKEND"] = "fasttokens"
logger.info(
f"Request migration {'enabled' if config.migration_limit > 0 else 'disabled'} "
f"(limit: {config.migration_limit})"
Expand Down
Loading
Loading