Changes / Feeds / nvidia/Nemotron-Mini-4B-Instruct
nvidia/Nemotron-Mini-4B-Instruct
2 changes to a tracked key since 2024-09-11, newest first. Head as last read: 791833e92ebddb0bc2c1007f6d2b6764f886a2ae.
| Date | Advisory | What changed, and what it does |
|---|---|---|
| 2024-09-23 | x_NXMC-2024-0263 | the chat template text changed but rendered byte-identically on the four fixed conversations (system+user, multi-turn, a tool call with the result as a string, the same with the result as content parts), all of which end on a user turn; a conversation ending on an assistant turn (training data, prefill) was not rendered (2024-09-23).51918b8bb262 -> 6a417790c444 Open in app |
| 2024-09-11 | x_NXMC-2024-0262 | the chat template text changed; the renderer could not run one of the two versions, so whether rendered prompts changed is not known (2024-09-11).b4d4a52be744 -> 51918b8bb262 Open in app |
Does vLLM send the reference prompt?
This template reads message content as a string, so vLLM hands it strings and sends the reference prompt on the three fixed conversations. Template at 791833e92ebd, read 2026-11-21.
npx neurarch template-check nvidia/Nemotron-Mini-4B-Instruct --engine vllm
Rendered with @huggingface/jinja from the template at that commit, the way vLLM hands it messages with --chat-template-content-format auto (the default): system, user and assistant content as text parts when the template reads parts, tool results as strings. That handling was measured end to end on 2026-11-13 through vLLM 0.33.0's own request parsing (parse_chat_messages) and transformers' apply_chat_template; it agreed with this page on all 18 models it could load; repositories that need custom code (InternVL, sarvam) could not be loaded there and are unconfirmed. SGLang, llama.cpp and other engines are not covered. same holds for a system+user turn, a multi-turn chat and a tool call, not every input.
Pin it, and let a bot move the pin
from_pretrained("nvidia/Nemotron-Mini-4B-Instruct", revision="791833e92ebddb0bc2c1007f6d2b6764f886a2ae")
With the Renovate preset, a pull request moves this pin when the head moves, and links back here.
[](https://www.neurarch.com/changes-feeds/models/nvidia/Nemotron-Mini-4B-Instruct.html)