Commit
•
f62b218
1
Parent(s):
501746a
Update README.md
Browse files
README.md
CHANGED
|
@@ -9,7 +9,7 @@ inference: false
|
|
| 9 |
|
| 10 |
## 更新信息
|
| 11 |
|
| 12 |
-
- **[2024/03/25]** 发布XVERSE-65B-Chat-GGUF模型，支持llama.cpp在MacOS/Linux/Windows系统上推理
|
| 13 |
- **[2023/12/08]** 发布 **XVERSE-65B-2** 底座模型，该模型在前一版本的基础上进行了 **Continual Pre-Training**，训练总 token 量达到 **3.2** 万亿；模型各方面的能力均得到提升，尤其是数学和代码能力，在 GSM8K 上提升 **20**%，HumanEval 上提升 **41**%。
|
| 14 |
- **[2023/11/29]** 更新模型架构及更多底座数据的相关信息。
|
| 15 |
- **[2023/11/24]** 更新预训练数据的相关信息。
|
|
@@ -17,7 +17,7 @@ inference: false
|
|
| 17 |
|
| 18 |
## Update Information
|
| 19 |
|
| 20 |
-
- **[2024/03/25]** Released the XVERSE-65B-Chat GGUF models, supporting llama.cpp to
|
| 21 |
- **[2023/12/08]** Released the **XVERSE-65B-2** base model. This model builds upon its predecessor through **Continual Pre-Training**, reaching a total training volume of **3.2** trillion tokens. It exhibits enhancements in all capabilities, particularly in mathematics and coding skills, with a **20%** improvement on the GSM8K benchmark and a **41%** increase on HumanEval.
|
| 22 |
- **[2023/11/29]** Update model architecture and additional pre-training data information.
|
| 23 |
- **[2023/11/24]** Update the related information of the pre-training data.
|
|
@@ -75,7 +75,7 @@ huggingface-cli download xverse/XVERSE-65B-Chat-GGUF xverse-65b-chat-q2_k.gguf -
|
|
| 75 |
cat xverse-65b-chat-q5_0.gguf.* > xverse-65b-chat-q5_0.gguf
|
| 76 |
```
|
| 77 |
|
| 78 |
-
我们演示了如何使用 `llama.cpp` 来运行xverse-65b:
|
| 79 |
|
| 80 |
```bash
|
| 81 |
./main -m xverse-65b-chat-q5_0.gguf -n 512 --color -i --temp 0.85 --top_k 30 --top_p 0.85 --repeat_penalty 1.1 -ins # -ngl 99 for GPU
|
|
@@ -95,7 +95,7 @@ For the `q5_k_m` model, due to maximum file size for uploading, we split the GGU
|
|
| 95 |
cat xverse-65b-chat-q5_0.gguf.* > xverse-65b-chat-q5_0.gguf
|
| 96 |
```
|
| 97 |
|
| 98 |
-
We demonstrate how to use `llama.cpp` to run xverse-65b:
|
| 99 |
|
| 100 |
```shell
|
| 101 |
./main -m xverse-65b-chat-q5_0.gguf -n 512 --color -i --temp 0.85 --top_k 30 --top_p 0.85 --repeat_penalty 1.1 -ins # -ngl 99 for GPU
|
|
|
|
| 9 |
|
| 10 |
## 更新信息
|
| 11 |
|
| 12 |
+
- **[2024/03/25]** 发布XVERSE-65B-Chat-GGUF模型，支持llama.cpp在MacOS/Linux/Windows系统上推理XVERSE-65B-Chat模型。
|
| 13 |
- **[2023/12/08]** 发布 **XVERSE-65B-2** 底座模型，该模型在前一版本的基础上进行了 **Continual Pre-Training**，训练总 token 量达到 **3.2** 万亿；模型各方面的能力均得到提升，尤其是数学和代码能力，在 GSM8K 上提升 **20**%，HumanEval 上提升 **41**%。
|
| 14 |
- **[2023/11/29]** 更新模型架构及更多底座数据的相关信息。
|
| 15 |
- **[2023/11/24]** 更新预训练数据的相关信息。
|
|
|
|
| 17 |
|
| 18 |
## Update Information
|
| 19 |
|
| 20 |
+
- **[2024/03/25]** Released the XVERSE-65B-Chat GGUF models, supporting llama.cpp to inference the XVERSE-65B-Chat model on MacOS/Linux/Windows systems.
|
| 21 |
- **[2023/12/08]** Released the **XVERSE-65B-2** base model. This model builds upon its predecessor through **Continual Pre-Training**, reaching a total training volume of **3.2** trillion tokens. It exhibits enhancements in all capabilities, particularly in mathematics and coding skills, with a **20%** improvement on the GSM8K benchmark and a **41%** increase on HumanEval.
|
| 22 |
- **[2023/11/29]** Update model architecture and additional pre-training data information.
|
| 23 |
- **[2023/11/24]** Update the related information of the pre-training data.
|
|
|
|
| 75 |
cat xverse-65b-chat-q5_0.gguf.* > xverse-65b-chat-q5_0.gguf
|
| 76 |
```
|
| 77 |
|
| 78 |
+
我们演示了如何使用 `llama.cpp` 来运行xverse-65b-chat-q5_0.gguf:
|
| 79 |
|
| 80 |
```bash
|
| 81 |
./main -m xverse-65b-chat-q5_0.gguf -n 512 --color -i --temp 0.85 --top_k 30 --top_p 0.85 --repeat_penalty 1.1 -ins # -ngl 99 for GPU
|
|
|
|
| 95 |
cat xverse-65b-chat-q5_0.gguf.* > xverse-65b-chat-q5_0.gguf
|
| 96 |
```
|
| 97 |
|
| 98 |
+
We demonstrate how to use `llama.cpp` to run xverse-65b-chat-q5_0.gguf:
|
| 99 |
|
| 100 |
```shell
|
| 101 |
./main -m xverse-65b-chat-q5_0.gguf -n 512 --color -i --temp 0.85 --top_k 30 --top_p 0.85 --repeat_penalty 1.1 -ins # -ngl 99 for GPU
|