rejoanul karim rafi

Owner of Mini Institute

[rejoanul karim rafi @RedoneRafi]

@dummerspast39 I think it should be 47-48. it worked better than muse spark 1.3 in some cases. and it found some opus 5.5 mistake in orchestrated workflow

[rejoanul karim rafi @RedoneRafi]

@elshayib_ xiaomi mimo 2.6 was released few weeks ago.

[rejoanul karim rafi @RedoneRafi]

@hairbyriike In real life they are Hamas in movie they are hero.

[rejoanul karim rafi @RedoneRafi]

@Dheemanthredy wow! when it will publish?

[rejoanul karim rafi @RedoneRafi]

@VikParuchuri What would be the comparison for indic languages in quality if we compare with Chandra ocr 2?

[rejoanul karim rafi @RedoneRafi]

@QwenDevs small

[rejoanul karim rafi @RedoneRafi]

@fal sorry fal, we don't need your model anymore.

[rejoanul karim rafi @RedoneRafi]

@HaochengXiUCB Is any minimax h3 variant works?

[rejoanul karim rafi @RedoneRafi]

@PrunaAI @DesignArena open-sourcing it anytime?

[rejoanul karim rafi @RedoneRafi]

@Shirink_13 Is it correct to consider the gov of the UAE as part of the ‘Ummah’?

[rejoanul karim rafi @RedoneRafi]

@ivanfioravanti Please make some panel image, where sequence of scene in one image for multi shot. in 2x2 panel or 3x3 panel.

[rejoanul karim rafi @RedoneRafi]

@aihacs @agnt_gg did you used any lora? or other technique for lip-sync?

[rejoanul karim rafi @RedoneRafi]

@PinakiTweetsBD ভাইয়া, আপনার দিকনির্দেশনা চেয়ে ইনবক্সে একটি মেসেজ দিয়েছি। ব্যস্ততার মধ্যেও কষ্ট করে একটু নজর দিলে চিরকৃতজ্ঞ থাকব।

[rejoanul karim rafi @RedoneRafi]

@thsottiaux 5 hour limit & it's super slow, it's too slower than Muse spark 1.3.
Faster token speeds save users valuable time, making high productivity well & worth exhausting limits sooner. creating a direct win-win for user satisfaction and company revenue.

[rejoanul karim rafi @RedoneRafi]

@Leacvikas @fal 😂 Bro!!! use some brain or follow reddit.

[rejoanul karim rafi @RedoneRafi]

@TeksEdge in diffusion model it sucks. such as LTX, WAN etc..

[rejoanul karim rafi @RedoneRafi]

I Canceled My Subscription, Join the League!!!

[rejoanul karim rafi @RedoneRafi]

@TheDibsterX @RabbiJeffx 🫡

[rejoanul karim rafi @RedoneRafi]

@louszbd The problem with GLM subscription :
- GLM-5.3-flash caps you at ~150M–300M tok/w
- Two prompts burned ~100Mtok which will barely use 20–30% of ChatGPT 5-hour limit.
In Price for performance ChatGPT 8-10x better than GLM for 20$ sub. It's useful token-gap.