AI & ML interests

The last paradigm of LLMs

Recent Activity

Banaxi-TechΒ  updated a Space 9 days ago
Project-LTM/README
Banaxi-TechΒ  published a Space 9 days ago
Project-LTM/README
View all activity

Banaxi-TechΒ 
posted an update 2 days ago
view post
Post
88
We're announcing Klein Ternary 1.7B.

Its our latest, ternary-weight model based on LFM 3 1.7B. Our target is 80-85% top 1 alignment, putting it close to 4 bit.

Klein Ternary is aimed at coding, with a footprint under 1 GB. It’s best suited for Raspberry Pis, phones and other low-power devices where even 4-bit models are too heavy.

The training pipeline is, normal self distillation, with a small seperate stage to bring out more coding performance after, and then a bit of post-training on Qwen distilled data to make it better.

We plan to release it in about 10 days.

Banaxi-TechΒ 
posted an update 6 days ago
view post
Post
6,636
We're introducing KV1 1.2B.
Its our first KV1 model using a 2 bit KV cache, based on LFM 2.5 1.2B.

KV1 1.2B retains 75.8% of original intelligence.

More details: klein-ml/blog

Model klein-ml/KV1-1.2B
  • 5 replies
Β·
Banaxi-TechΒ 
posted an update 7 days ago
view post
Post
6,360
BananaMind 3 is coming.

Biggest context in all SLMs with NSA.
10M, 50M, 150M, 400M.

And maybe, 1B.
  • 14 replies
Β·
Banaxi-TechΒ 
posted an update 8 days ago
view post
Post
6,709
We've released the BananaMind/BETU-SLM-Leaderboard

Its so simple to use, it ranks models, and by one command add your own model to it.

It scores by Overall Index which is a weighted mix of Hellaswawg, ARC EASY, PIQA, Arithmark 3, Tiny Theory of Mind, and BananaMind Base Bench.
  • 1 reply
Β·
Banaxi-TechΒ 
posted an update 9 days ago
view post
Post
106
We ARE Announcing BananaMind 3 Ultra.

400M parameters, designed to beat LFM 2.5 1.2B.

This model is a usable small agent. Not a demo model like our 3M models.


FP8. KV1. Optimization. FactBase DCLM. Fable 5.2 for maximum speed kernels. TST.

This model will M.
What do you think? 50+ INT Index on the Open SLM leaderboard? BananaMind 3 is coming before christmas.

  • 11 replies
Β·
Banaxi-TechΒ 
updated a Space 9 days ago
Banaxi-TechΒ 
published a Space 9 days ago
Banaxi-TechΒ 
posted an update 12 days ago
Banaxi-TechΒ 
posted an update 13 days ago
view post
Post
71
Hi guys.

We're announcing Klein SL Pro, our new model at Klein ML.
It's doing post-training of Qwen3.8 Flash Next to make IT fit on 16GB VRAM.

It works by converting its MoE to SLMoE, SLMoE is an architecture, that chooses experts per response, not per token anymore.

This makes it so you can offload everything else to the disk and not get slow speed.


We're also seeking more compute to do this same thing on Deepseek V4.1 Flash, allowing it to run on 16GB VRAM tightly and definitely on 24GB. If you have hardware, more than 200B of VRAM in total, create a discussion klein-ml/README.
  • 2 replies
Β·
Banaxi-TechΒ 
posted an update 17 days ago
Banaxi-TechΒ 
posted an update 19 days ago
view post
Post
136
SupraTTS is going to be cooked lol

Klein TTS X1.5 is releasing in under a WEEK


(supratts scoring 743 on tts index: 😭)


we are just doing our post training now before releasing and are so excited with the results x2 will be insane!
  • 19 replies
Β·
Banaxi-TechΒ 
posted an update 21 days ago
view post
Post
115
Happy halloween everyone!


We've did some updates to Banaxi-Tech/TTS-Arena and now, everyone who uses the TTS Arena alot will get featured in our Klein TTS Release. Using it REALLY helps us alot so please use it.

Updates on the sizes: Klein TTS Flash Lite with 0.4M parameters, Klein TTS Flash with 1.7M parameters, Klein TTS Pro with 13M parameters and Ultra with 21M.
Banaxi-TechΒ 
posted an update 23 days ago
view post
Post
92
Hi everyone, we're releasing the TTS Arena.

You can test models and see rankings, its like the SLM Arena but for TTS.

Right now there are 2 stealth models on the arena, "cirxinx" and "tiger-alpha", try them out to preview a new model πŸ₯·


We're also training Klein TTS X1.5 right now, they will release soon. We're also adding a 210K parameter size.


If your excited on them vote here: Banaxi-Tech/HuggingPolls

If you'r wondering what that is, its HuggingPolls, a space we released for Polls on huggingface.

Because huggingface doesent have any polls, we made it.

You can also make your own polls!
  • 1 reply
Β·
Banaxi-TechΒ 
posted an update 24 days ago
view post
Post
109
Klein TTS X1.5!!!


It will feature 1.7M,4M 13M and a 21M size

klein-ml


Early access opening soon
Its training right now!

We are also working on the X2 architecture featuring a feature that allows the model to iterate on the audio it produced.
Banaxi-TechΒ 
posted an update 26 days ago
view post
Post
5,618
BananaMind 3 is going to be an comeback. Sorry for the BGA things ( i genuinely didnt know about nsa existing and yes I used aiiii)
Were going to use our BananaMind 2.1 Unified architecture with some changes + NSA and some other stuff.
  • 48 replies
Β·
Banaxi-TechΒ 
posted an update 27 days ago
view post
Post
4,912
BananaMind is in the google situation before argon right now, right?
  • 29 replies
Β·
Banaxi-TechΒ 
posted an update about 1 month ago
view post
Post
206
So... im on vacation for 5 more days and I just realized... my hotel doesent have a grounding port 😟 so my pc shocked me 3 times.

So no progress for 5 days... only me and this phone
  • 2 replies
Β·
Banaxi-TechΒ 
posted an update about 1 month ago
view post
Post
4,253
Hi everyone
We at
saicr
have found a very interesting "thing" we call ADN via interpretabillity. It appears on EVERY prompt sent to the model. More on that later.
Also I made a X account follow it for more updates:
@banaxitech

And KleinML:
@kleinmlai

Here's my website for Klein ML: https://kleinml.com/?ref=huggingface.co
  • 4 replies
Β·
Banaxi-TechΒ 
posted an update about 1 month ago
view post
Post
209
bci is the future.


and i dont like that
  • 4 replies
Β·
Banaxi-TechΒ 
posted an update about 1 month ago
view post
Post
7,311
hi everyone
we at
klein-ml
are announcing KleinMem

These models will have Native memory, without any system-prompt or prompt, it still remembers exactly.

Get early access at https://tinyurl.com/klein-eap limited to 15 users.


Architecture details are coming soon toon.

If you know alot about architecture and want to help build it, join
klein-ml
.