Update README.md
Browse files
README.md
CHANGED
|
@@ -2,6 +2,9 @@ Requantizations of a Q5_K_M quant of a trending 70b model without better quant/f
|
|
| 2 |
|
| 3 |
Q3_K_M, Q3_K_S, Q3_K_XS, Q2_K_S, IQ3_XXS, IQ2_XS available. Miqudev provided Q5_K_M, Q4_K_M, and Q2_K from his probable FP16.
|
| 4 |
|
|
|
|
|
|
|
|
|
|
| 5 |
Miku 70b has a theta of 1,000,000, like CodeLlama, and not 10,000, like Llama 2 models usually have.
|
| 6 |
That feature singularizes it to my knowledge to ALL Llama 2 models, beside Codellamas which also have a theta of 1,000,000..
|
| 7 |
|
|
|
|
| 2 |
|
| 3 |
Q3_K_M, Q3_K_S, Q3_K_XS, Q2_K_S, IQ3_XXS, IQ2_XS available. Miqudev provided Q5_K_M, Q4_K_M, and Q2_K from his probable FP16.
|
| 4 |
|
| 5 |
+
Bonus : a Kobold.CPP Frankenstein which reads IQ3_XXS models and is not affected by the Kobold.CPP 1.56/1.57 slowdown at the cost of an absent Mixtral fix.
|
| 6 |
+
https://github.com/Nexesenex/kobold.cpp/releases/tag/v1.57_b2030
|
| 7 |
+
|
| 8 |
Miku 70b has a theta of 1,000,000, like CodeLlama, and not 10,000, like Llama 2 models usually have.
|
| 9 |
That feature singularizes it to my knowledge to ALL Llama 2 models, beside Codellamas which also have a theta of 1,000,000..
|
| 10 |
|