Menlo/R-Horizon-vision-train-ready
Viewer
•
Updated
•
2.28k
•
15
AI & ML interests
None defined yet.
Recent Activity
View all activity
Scaling
🔥
1
2
#11 opened about 15 hours ago
by
ziryeg
bachvudinh
updated
a
dataset
6 days ago
bachvudinh
published
a
dataset
6 days ago
bachvudinh
updated
a
dataset
19 days ago
bachvudinh
published
a
dataset
19 days ago
bachvudinh
updated
a
dataset
19 days ago
alandao
authored
a
paper
4 months ago
Post
1337
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
alandao
authored
2
papers
7 months ago
alandao
authored
a
paper
8 months ago
alandao
authored
a
paper
about 1 year ago
jan-hq
updated
a
model
about 1 year ago
alandao
updated
a
model
about 1 year ago
jan-hq
updated
4
models
about 1 year ago
Menlo/mini-Ichigo-llama3.2-3B-s-base
Audio-Text-to-Text
•
3B
•
Updated
•
6
•
1
Menlo/Ichigo-llama3.1-s-instruct-v0.3-phase-2
Audio-Text-to-Text
•
8B
•
Updated
•
11
•
5
Menlo/Ichigo-llama3.1-s-base-v0.3
Audio-Text-to-Text
•
8B
•
Updated
•
1
•
4
Menlo/Ichigo-llama3.2-3B-s-instruct-cp-4000
Text Generation
•
3B
•
Updated
jan-hq
updated
2
datasets
about 1 year ago