dynamo-renderer 5.1.0

Standalone OpenAI chat-template / prompt formatting (HF chat_template via minijinja).
Documentation 
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

# SPDX-FileCopyrightText: Copyright (c) 2024-2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: Apache-2.0

[package]
name = "dynamo-renderer"
readme = "README.md"
version = "5.1.0"
edition.workspace = true
description = "Standalone OpenAI chat-template / prompt formatting (HF chat_template via minijinja)."
authors.workspace = true
license = "Apache-2.0"
homepage.workspace = true
repository.workspace = true
keywords = ["llm", "chat-template", "prompt", "jinja", "inference"]

[dependencies]
# The chat-template engine bridges OpenAI request types (dynamo-protocols) with
# the HF chat_template jinja engine (minijinja). It does NOT depend on tokenizer
# internals; `dynamo-tokenizers` is re-exported only as a one-import convenience
# for consumers that want tokenize + templating together.
dynamo-protocols = { workspace = true }
dynamo-tokenizers = { workspace = true }

anyhow = { workspace = true }
chrono = { workspace = true }
either = { workspace = true }
serde = { workspace = true }
serde_json = { workspace = true, features = ["preserve_order", "raw_value"] }
tracing = { workspace = true }

minijinja = { workspace = true, features = ["loader", "loop_controls"] }
minijinja-contrib = { workspace = true, features = ["pycompat"] }

[dev-dependencies]
tempfile = { workspace = true }