Menlo/Jan-nano-128k
Text Generation
•
Updated
•
74
AI & ML interests
None defined yet.
Recent Activity
View all activity
jan-hq
updated
a
model
about 11 hours ago
jan-hq
updated
a
model
about 15 hours ago
Curious - Yarn setting different from Qwen3 repo for 128k?
3
#1 opened about 17 hours ago
by
DavidAU
alandao
updated
a
collection
about 17 hours ago
alandao
updated
a
collection
about 19 hours ago
can i use duckduckgo instead of serper
1
#7 opened 3 days ago
by
otiswillie
Bigger models
1
#8 opened 3 days ago
by
MaximusAI
what about coding tasks?
1
#6 opened 7 days ago
by
dfsafdsf
jan-hq
updated
2
models
8 days ago
Post
1115
Don’t give up 🔥
Do you know what I was planning to do this time last week?
I was preparing to write a report declaring that Jan Nano was a failed project because the benchmark results didn’t meet expectations.
But I thought — it can’t be. When loading the model into the app, the performance clearly felt better. So why were the benchmark results worse?
That’s when I reviewed the entire benchmark codebase and realized something fundamental: agentic or workflow-based approaches introduce a huge gap and variation when benchmarking. Jan-nano was trained with an agentic setup — it simply can’t be benchmarked using a rigid workflow-based method.
I made the necessary changes, and the model ended up performing even better than before the issues arose. Turns out the previous benchmarking method conflicted with the way the model was trained.
What if I had given up? That would’ve meant 1.5 months of training and a huge amount of company resources wasted.
But now, this is officially the most successful and biggest release for the whole team — all thanks to Jan-nano.
Menlo/Jan-nano
Do you know what I was planning to do this time last week?
I was preparing to write a report declaring that Jan Nano was a failed project because the benchmark results didn’t meet expectations.
But I thought — it can’t be. When loading the model into the app, the performance clearly felt better. So why were the benchmark results worse?
That’s when I reviewed the entire benchmark codebase and realized something fundamental: agentic or workflow-based approaches introduce a huge gap and variation when benchmarking. Jan-nano was trained with an agentic setup — it simply can’t be benchmarked using a rigid workflow-based method.
I made the necessary changes, and the model ended up performing even better than before the issues arose. Turns out the previous benchmarking method conflicted with the way the model was trained.
What if I had given up? That would’ve meant 1.5 months of training and a huge amount of company resources wasted.
But now, this is officially the most successful and biggest release for the whole team — all thanks to Jan-nano.
Menlo/Jan-nano
Mobile?
1
#5 opened 9 days ago
by
SuperPauly
About MCP Server
2
#4 opened 9 days ago
by
xiaoyaojun
About MCP Server
2
#4 opened 9 days ago
by
xiaoyaojun
alandao
updated
2
models
9 days ago
About using non-jan app
1
#3 opened 9 days ago
by
alandao
About using non-jan app
👍
1
#2 opened 9 days ago
by
alandao
Local Installation Video and Testing - Step by Step
1
#2 opened 10 days ago
by
fahdmirzac
alandao
updated
a
collection
10 days ago