Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
title: "ESM2 Quantized Models"
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
## ESM2 Quantized
|
| 6 |
+
|
| 7 |
+
ESM2 Quantized is an adapted version of the ESM2 architectures. It uses local attention instead of global attention, allowing for models with longer input sizes. ESM2 Quantized models have a context size of 2,050, double that of the standard ESM2 model. This kind of model was trained with int4 quantization. Several ESM2 Quantized models are available:
|
| 8 |
+
|
| 9 |
+
| Model | Num layers |
|
| 10 |
+
|------------------------------|----|
|
| 11 |
+
| [gabrielbianchin/esm2_t36_long_int4](https://huggingface.co/gabrielbianchin/esm2_t36_long_int4) | 36 |
|
| 12 |
+
| [gabrielbianchin/esm2_t33_long_int4](https://huggingface.co/gabrielbianchin/esm2_t33_long_int4) | 33 |
|
| 13 |
+
| [gabrielbianchin/esm2_t30_long_int4](https://huggingface.co/gabrielbianchin/esm2_t30_long_int4) | 30 |
|
| 14 |
+
| [gabrielbianchin/esm2_t12_long_int4](https://huggingface.co/gabrielbianchin/esm2_t12_long_int4) | 12 |
|
| 15 |
+
| [gabrielbianchin/esm2_t6_long_int4](https://huggingface.co/gabrielbianchin/esm2_t6_long_int4) | 6 |
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
For detailed information, please refer to the paper.
|