mirror of
https://github.com/zenlm/docs.git
synced 2026-10-30 15:00:08 +00:00
zeekay
0e2ed95f7d
The title, home, overview and sidebar sold Zen4 and a "Zen5 Ultra, 2T+, the largest open-weight model in history" in research preview. None of that had a source. They now say who makes Zen LM, Zoo Labs Foundation, a 501(c)(3) non-profit, the two jobs it is chosen for, that Zen 6 and Zen 6 Flash are available now, and that Zen 7 is a research preview with a link to request access. Zen 5 and Zen 4 are earlier generations. A Zen 6 page carries the numbers from the two model cards, with the loader strings and the License & attribution section marked data-upstream. Upstream base names leave the rest of the read-through, and scripts/check-zen.mjs fails the build when one comes back; the Pages workflow now runs the build script so CI runs the check too.
33 lines
1010 B
Plaintext
33 lines
1010 B
Plaintext
# The upstream names Zen's story is not told with. Read by scripts/check-zen.mjs,
|
|
# which runs at the end of the build and fails it on a hit.
|
|
#
|
|
# One name per line; `#` starts a comment. Case is ignored and a name matches as
|
|
# a whole word — no letter directly before or after it — so `Qwen` also catches
|
|
# Qwen3, Qwen3.8, qwen35, qwen4 and Qwen3_5ForConditionalGeneration.
|
|
#
|
|
# A Zen page says these names in two places only, and marks both `data-upstream`:
|
|
# the Architecture value (the loader string from the weights' own config) and the
|
|
# License & attribution section at the bottom. A post whose subject is another
|
|
# lab's release is marked at its article. Everything else a reader meets — title,
|
|
# description, headings, prose, tables, alt text — says Zen.
|
|
#
|
|
# The list starts from the base models the zenlm cards on Hugging Face declare.
|
|
# A new one goes here, and nowhere else.
|
|
Qwen
|
|
QwenLM
|
|
QwQ
|
|
QVQ
|
|
Flash-Next
|
|
Bonsai
|
|
prism-ml
|
|
DeepSeek
|
|
MiniMax
|
|
Wan
|
|
Granite
|
|
Nemotron
|
|
unsloth
|
|
FLUX
|
|
TRELLIS
|
|
YuE
|
|
SSD-1B
|