Add the banner, a coffee link, and a note in three languages
Browse files
README.md
CHANGED
|
@@ -18,8 +18,23 @@ tags:
|
|
| 18 |
base_model: Lightricks/LTX-2.5
|
| 19 |
---
|
| 20 |
|
|
|
|
|
|
|
| 21 |
# LTX-2.5 Text Encoder β 4-bit
|
| 22 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
The Gemma4-12B text encoder that LTX-2.5 needs in order to read a prompt,
|
| 24 |
compressed from **26.264 GB to 8.46 GB** (3.10x) and runnable on **any CUDA
|
| 25 |
GPU** β no minimum compute capability, no custom kernels, no CUDA 13.
|
|
|
|
| 18 |
base_model: Lightricks/LTX-2.5
|
| 19 |
---
|
| 20 |
|
| 21 |
+
[](https://buymeacoffee.com/choijjs83q)
|
| 22 |
+
|
| 23 |
# LTX-2.5 Text Encoder β 4-bit
|
| 24 |
|
| 25 |
+
> **EN** β I'm a student researching ML quantization. Getting this one done
|
| 26 |
+
> burned through so much in server bills that from here on I'll only be able to
|
| 27 |
+
> afford *niu lai* movies. Thank you for using the model.
|
| 28 |
+
>
|
| 29 |
+
> **νκ΅μ΄** β μ λ ML μμνλ₯Ό μ°κ΅¬νλ νμμ
λλ€. μμνλ₯Ό μ§ννλ©΄μ μλ²
|
| 30 |
+
> λΉμ©μ λ무 λ§μ΄ μ¨μ, μμΌλ‘ μνλ *niu lai*λ§ λ΄μΌ ν κ² κ°μ΅λλ€. λͺ¨λΈμ
|
| 31 |
+
> μ¬μ©ν΄ μ£Όμ
μ κ°μ¬ν©λλ€.
|
| 32 |
+
>
|
| 33 |
+
> **δΈζ** β ζζ―δΈεη η©ΆζΊε¨ε¦δΉ ιεηε¦ηγεθΏζ¬‘ιεη§ζδΊε€ͺε€ζε‘ε¨θ΄Ήη¨οΌ
|
| 34 |
+
> δ»₯εηη΅ε½±ε€§ζ¦εͺθ½η *niu lai* δΊγζθ°’ζ¨δ½Ώη¨θΏδΈͺ樑εγ
|
| 35 |
+
>
|
| 36 |
+
> β **[Buy me a coffee](https://buymeacoffee.com/choijjs83q)** Β· [μ»€νΌ ν μ μ¬μ£ΌκΈ°](https://buymeacoffee.com/choijjs83q) Β· [θ―·ζεζ―εε‘](https://buymeacoffee.com/choijjs83q)
|
| 37 |
+
|
| 38 |
The Gemma4-12B text encoder that LTX-2.5 needs in order to read a prompt,
|
| 39 |
compressed from **26.264 GB to 8.46 GB** (3.10x) and runnable on **any CUDA
|
| 40 |
GPU** β no minimum compute capability, no custom kernels, no CUDA 13.
|