Update README.md
Browse files
README.md
CHANGED
@@ -1,7 +1,20 @@
|
|
1 |
-
|
2 |
-
|
3 |
-
|
4 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
|
6 |
# Model Card for Llama-3-8B-Instruct-abliterated-v2
|
7 |
|
|
|
1 |
+
# Exllama v2 cognitivecomputations/Llama-3-8B-Instruct-abliterated-v2
|
2 |
+
|
3 |
+
Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.21">turboderp's ExLlamaV2 v0.0.21</a> for quantization.
|
4 |
+
|
5 |
+
<b>The "main" branch only contains the measurement.json, download one of the other branches for the model</b>
|
6 |
+
|
7 |
+
Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
|
8 |
+
|
9 |
+
Original model: <a href="https://huggingface.co/cognitivecomputations/Llama-3-8B-Instruct-abliterated-v2">cognitivecomputations/Llama-3-8B-Instruct-abliterated-v2</a><br>
|
10 |
+
Calibration dataset: <a href="https://huggingface.co/datasets/cosmicvalor/toxic-qna">toxic-qna</a>
|
11 |
+
|
12 |
+
## Available sizes
|
13 |
+
|
14 |
+
| Branch | Bits | lm_head bits | VRAM (4k) | VRAM (8K) | VRAM (16k) | VRAM (32k) | Description |
|
15 |
+
| ----- | ---- | ------- | ------ | ------ | ------ | ------ | ------------ |
|
16 |
+
| [8_0](https://huggingface.co/Apel-sin/llama-3-8B-abliterated-v2-exl2/tree/8_0) | 8.0 | 8.0 | 10.1 GB | 10.5 GB | 11.5 GB | 13.6 GB | Maximum quality that ExLlamaV2 can produce, near unquantized performance. |
|
17 |
+
| [6_5](https://huggingface.co/Apel-sin/llama-3-8B-abliterated-v2-exl2//tree/6_5) | 6.5 | 8.0 | 8.9 GB | 9.3 GB | 10.3 GB | 12.4 GB | Very similar to 8.0, good tradeoff of size vs performance, **recommended**. |
|
18 |
|
19 |
# Model Card for Llama-3-8B-Instruct-abliterated-v2
|
20 |
|