Update README.md
Browse files
README.md
CHANGED
@@ -62,7 +62,7 @@ The new methods available are:
|
|
62 |
## Provided files
|
63 |
|
64 |
| Name | Quant method | Bits | Use case |
|
65 |
-
| ---- | ---- | ---- | ---- |
|
66 |
| Merak-7B-v3-Mini-Orca-Indo.gguf | Q2_K | 2 | smallest, significant quality loss - not recommended for most purposes |
|
67 |
| Merak-7B-v3-Mini-Orca-Indo.q3_K_S.gguf | Q3_K_S | 3 | very small, high quality loss |
|
68 |
| Merak-7B-v3-Mini-Orca-Indo.q3_K_M.gguf | Q3_K_M | 3 | very small, high quality loss |
|
@@ -78,23 +78,17 @@ The new methods available are:
|
|
78 |
|
79 |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
|
80 |
|
81 |
-
readme adapted from [TheBloke](https://huggingface.co/TheBloke)
|
82 |
-
|
83 |
## How to run in `text-generation-webui`
|
84 |
|
85 |
Further instructions here: [text-generation-webui/docs/llama.cpp.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp.md).
|
86 |
|
|
|
87 |
# Original model card: 🦚Merak-7B-v3-Mini-Orca🐳
|
88 |
|
89 |
**Merak-7B-v3-Mini-Orca** is Ichsan2895's [Merak-7B-v3](https://huggingface.co/Ichsan2895/Merak-7B-v3) fine-tuned
|
90 |
on Bahasa Indonesia translated psmathur's [orca_mini_v1_dataset](https://huggingface.co/datasets/psmathur/orca_mini_v1_dataset).
|
91 |
|
92 |
|
93 |
-
## Usage
|
94 |
-
This model fit on 16GB VRAM GPU (Google Collab T4 wil do), by using BitsandBytes it can run on 6GB VRAM GPU.
|
95 |
-
|
96 |
-
[](https://colab.research.google.com/drive/11xmPcRNirGwZcpgmNPNpUioJUG4PQBuh)
|
97 |
-
|
98 |
**Quantized** versions is available:
|
99 |
|
100 |
GPTQ: https://huggingface.co/asyafiqe/Merak-7B-v3-Mini-Orca-Indo-GPTQ
|
@@ -188,4 +182,7 @@ Please see the Responsible Use Guide available at https://ai.meta.com/llama/resp
|
|
188 |
journal={CoRR},
|
189 |
year={2021}
|
190 |
}
|
191 |
-
```
|
|
|
|
|
|
|
|
62 |
## Provided files
|
63 |
|
64 |
| Name | Quant method | Bits | Use case |
|
65 |
+
| ---- | ---- | ---- | ---- |
|
66 |
| Merak-7B-v3-Mini-Orca-Indo.gguf | Q2_K | 2 | smallest, significant quality loss - not recommended for most purposes |
|
67 |
| Merak-7B-v3-Mini-Orca-Indo.q3_K_S.gguf | Q3_K_S | 3 | very small, high quality loss |
|
68 |
| Merak-7B-v3-Mini-Orca-Indo.q3_K_M.gguf | Q3_K_M | 3 | very small, high quality loss |
|
|
|
78 |
|
79 |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
|
80 |
|
|
|
|
|
81 |
## How to run in `text-generation-webui`
|
82 |
|
83 |
Further instructions here: [text-generation-webui/docs/llama.cpp.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp.md).
|
84 |
|
85 |
+
|
86 |
# Original model card: 🦚Merak-7B-v3-Mini-Orca🐳
|
87 |
|
88 |
**Merak-7B-v3-Mini-Orca** is Ichsan2895's [Merak-7B-v3](https://huggingface.co/Ichsan2895/Merak-7B-v3) fine-tuned
|
89 |
on Bahasa Indonesia translated psmathur's [orca_mini_v1_dataset](https://huggingface.co/datasets/psmathur/orca_mini_v1_dataset).
|
90 |
|
91 |
|
|
|
|
|
|
|
|
|
|
|
92 |
**Quantized** versions is available:
|
93 |
|
94 |
GPTQ: https://huggingface.co/asyafiqe/Merak-7B-v3-Mini-Orca-Indo-GPTQ
|
|
|
182 |
journal={CoRR},
|
183 |
year={2021}
|
184 |
}
|
185 |
+
```
|
186 |
+
|
187 |
+
|
188 |
+
readme adapted from [TheBloke](https://huggingface.co/TheBloke)
|