fixed docs

Signed-off-by: ftgreat <[email protected]>

README.md

@@ -155,14 +155,16 @@ git clone https://github.com/OpenBMB/BMTrain
 cd BMTrain
 python setup.py install
 ```
-- [Tips] For single-node docker environments, we need to set up ports for your ssh. e.g., [email protected] with port 7110
+- [Optional] For BMInf low-resource inference, install [BMInf](https://github.com/OpenBMB/BMInf)
 ```
->>> vim ~/.ssh/config
-Host 127.0.0.1
-    Hostname 127.0.0.1
-    Port 7110
-    User root
+pip install bminf
+
+```
+- [Optional] For Flash Attention, install [Flash-attention](https://github.com/HazyResearch/flash-attention)
 ```
+pip install flash-attn
+```
+
 - [Tips] For multi-node docker environments, generate ssh keys and copy the public key to all nodes (in `~/.ssh/`)
 ```
 >>> ssh-keygen -t rsa -C "[email protected]"

README_zh.md

@@ -144,6 +144,17 @@ git clone https://github.com/OpenBMB/BMTrain
 cd BMTrain
 python setup.py install 
 ```
+
+- [可选] 开启BMInf低资源推理, 需要安装[BMInf](https://github.com/OpenBMB/BMInf)
+```
+pip install bminf
+
+```
+- [可选] 对于FlashAttention, 需要安装[Flash-attention](https://github.com/HazyResearch/flash-attention)
+```
+pip install flash-attn
+```
+
 - [可选] 镜像构建，请参照 [Dockerfile](https://github.com/FlagAI-Open/FlagAI/blob/master/Dockerfile)
 - [提示] 单节点docker环境下，运行多卡数据并行需要设置host。 例如，docker节点 [email protected]，其端口 7110。
 ```

examples/Aquila/Aquila-chat/README.md

@@ -45,6 +45,9 @@ The tokenizer used in the Aquila model was trained from scratch by us and suppor
 | LLaMA | 32000 | sp(bpe)|1805| 1257|1970 |
 | Aquila | 100000 | bpe|1575 | 477|1679 |
 
+Aquila系列模型均可在24G显卡上运行。
+
+The Aquila series models can all run on a 24GB graphics card.
 
 ## 训练数据集/Training data 
 
@@ -61,6 +64,8 @@ We have added support for two additional multimodal instructions: text-to-image
 
 ### 1. 推理/Inference
 
+运行成功后模型会自动下载在`./checkpoints_in`里，
+
 ```python
 import os
 import torch
@@ -71,7 +76,7 @@ from flagai.data.tokenizer import Tokenizer
 import bminf
 
 state_dict = "./checkpoints_in"
-model_name = 'aquilachat-33b'
+model_name = 'aquilachat-7b'
 
 loader = AutoLoader(
     "lm",
@@ -174,10 +179,10 @@ with torch.no_grad():
 Create a new directory named `aquilachat-7b` inside `./checkpoints_in`. Place the fine-tuned checkpoint and all other files from the original `aquilachat-7b` model, including `config.json`, `mergex.txt`, `vocab.json`, and `special_tokens_map.json`, into this directory.
 
 #### Step 2: 修改参数/ Modify Parameters
-* `cd /examples/Aquila/Aquila-sft`
+* `cd /examples/Aquila/Aquila-chat`
 * 配置`hostfile`文件, 参考[这里](../../../doc_zh/TUTORIAL_8_ENVIRONMENT_SETUP.md#a配置hostfilehostfile-中的v100-1-与sshconfig-对应) ; Configure the `hostfile` file, refer to [here](../../../docs/TUTORIAL_8_ENVIRONMENT_SETUP.md)
-* 配置`bmtrain_mgpu.sh`文件, 将`SCRIPT_FILE`改成`aquila_sft.py`; configure the `bmtrain_mgpu.sh` file, change `SCRIPT_FILE` to `aquila_sft.py`
-* (可选) 在`Aquila-sft.yaml`文件里更改参数 ; (optional) change parameters in `Aquila-sft.yaml`
+* 配置`bmtrain_mgpu.sh`文件, 将`SCRIPT_FILE`改成`aquila_chat.py`; configure the `bmtrain_mgpu.sh` file, change `SCRIPT_FILE` to `aquila_chat.py`
+* (可选) 在`Aquila-chat.yaml`文件里更改参数 ; (optional) change parameters in `Aquila-chat.yaml`
 
 | 参数名 Parameter             | 类型 Type | 描述 Description                                        |
 |--------------------------------|------------|-------------------------------------------------------|
@@ -195,7 +200,7 @@ Create a new directory named `aquilachat-7b` inside `./checkpoints_in`. Place th
 
 #### Step 3: 启动可监督微调/Start SFT
 ```
-bash dist_trigger_docker.sh hostfile Aquila-sft.yaml aquilachat-7b [实验名]
+bash dist_trigger_docker.sh hostfile Aquila-chat.yaml aquilachat-7b [实验名]
 ```
 接下来会输出下列信息，注意`NODES_NUM`应该与节点数相等，`LOGFILE`是模型运行的日志文件；The following information will be output. Note that `NODES_NUM` should be equal to the number of nodes, and `LOGFILE` is the log file for the model run.
 

examples/Aquila/Aquila-code/README.md

@@ -43,6 +43,9 @@ We used different tokenizers to extract ten thousand data samples from English,
 | LLaMA | 32000 | sp(bpe)|1805| 1257|1970 |
 | Aquila | 100000 | bpe|1575 | 477|1679 |
 
+Aquila系列模型均可在24G显卡上运行。
+
+The Aquila series models can all run on a 24GB graphics card.
 
 ## 训练数据集/Training data 
 `AquilaCode-7B-NV`和`AquilaCode-7B-TS`模型训练使用了[starcoderdata](https://huggingface.co/datasets/bigcode/starcoderdata)中的shell, sql，C, C++, Java, Javascript, Python, git-commits, github-issues, jupyter-scripts, jupyter-structured-text数据

examples/Aquila/Aquila-code/generate_code.py

@@ -4,6 +4,7 @@
 import torch
 import os
 from flagai import mpu
+import sys;sys.path.append("/data2/yzd/workspace/FlagAI")
 from flagai.auto_model.auto_loader import AutoLoader
 import random
 import numpy as np

examples/Aquila/Aquila-pretrain/README.md

@@ -37,14 +37,17 @@ The tokenizer used in the Aquila model was trained from scratch by us and suppor
 | LLaMA | 32000 | sp(bpe)|1805| 1257|1970 |
 | Aquila | 100000 | bpe|1575 | 477|1679 |
 
+Aquila系列模型均可在24G显卡上运行。
+
+The Aquila series models can all run on a 24GB graphics card.
 
 
 ## 训练数据集/Training data 
 Aquila预训练使用了Pile，[RedPajama-Data-1T](https://huggingface.co/datasets/togethercomputer/RedPajama-Data-1T), [Wikipedia](https://huggingface.co/datasets/wikipedia), [C4](https://huggingface.co/datasets/c4), 悟道中文数据集、电子书、专利、百科、论坛, github数据等, 详情可见下图。
 
 The Aquila-7B model was pretrained on Pile，[RedPajama-Data-1T](https://huggingface.co/datasets/togethercomputer/RedPajama-Data-1T), [Wikipedia](https://huggingface.co/datasets/wikipedia), [C4](https://huggingface.co/datasets/c4), Wudao Corpus、e-book、Patent, encyclopedia, forum, github etc. Details are given in the figure below.
 
-![Screenshot](../img/data_dist.png)
+<!-- ![Screenshot](../img/data_dist.png) -->
 
 
 
@@ -97,15 +100,15 @@ bash dist_trigger_docker.sh hostfile Aquila-pretrain-33B.yaml aquila-33b [实验
 
 #### Step 2: 启动可监督微调/Start SFT
 ```
-cd ../Aquila-sft/
+cd ../Aquila-chat/
 ```
 对于Aquila-7B模型：
 ```
-bash dist_trigger_docker.sh hostfile Aquila-sft.yaml aquila-7b [实验名 experiment name]
+bash dist_trigger_docker.sh hostfile Aquila-chat.yaml aquila-7b [实验名 experiment name]
 ```
 对于Aquila-33B模型:
 ```
-bash dist_trigger_docker.sh hostfile Aquila-sft.yaml aquila-33b [实验名 experiment name]
+bash dist_trigger_docker.sh hostfile Aquila-chat.yaml aquila-33b [实验名 experiment name]
 ```
 接下来会输出下列信息，注意`NODES_NUM`应该与节点数相等，`LOGFILE`是模型运行的日志文件；The following information will be output. Note that `NODES_NUM` should be equal to the number of nodes, and `LOGFILE` is the log file for the model run.
 

examples/Aquila/README.md

@@ -36,14 +36,16 @@ The tokenizer used in the Aquila model was trained from scratch by us and suppor
 | LLaMA | 32000 | sp(bpe)|1805| 1257|1970 |
 | Aquila | 100000 | bpe|1575 | 477|1679 |
 
+Aquila系列模型均可在24G显卡上运行。
 
+The Aquila series models can all run on a 24GB graphics card.
 
 ## 训练数据集/Training data 
 Aquila预训练使用了Pile，[RedPajama-Data-1T](https://huggingface.co/datasets/togethercomputer/RedPajama-Data-1T), [Wikipedia](https://huggingface.co/datasets/wikipedia), [C4](https://huggingface.co/datasets/c4), 悟道中文数据集、电子书、专利、百科、论坛, github数据等, 详情可见下图。
 
 The Aquila-7B model was pretrained on Pile，[RedPajama-Data-1T](https://huggingface.co/datasets/togethercomputer/RedPajama-Data-1T), [Wikipedia](https://huggingface.co/datasets/wikipedia), [C4](https://huggingface.co/datasets/c4), Wudao Corpus、e-book、Patent, encyclopedia, forum, github etc. Details are given in the figure below.
 
-![Screenshot](./img/data_dist.png)
+<!-- ![Screenshot](./img/data_dist.png) -->
 
 
 
@@ -96,15 +98,15 @@ bash dist_trigger_docker.sh hostfile Aquila-pretrain-33B.yaml aquila-33b [实验
 
 #### Step 2: 启动可监督微调/Start SFT
 ```
-cd ../Aquila-sft/
+cd ../Aquila-chat/
 ```
 对于Aquila-7B模型：
 ```
-bash dist_trigger_docker.sh hostfile Aquila-sft.yaml aquila-7b [实验名 experiment name]
+bash dist_trigger_docker.sh hostfile Aquila-chat.yaml aquila-7b [实验名 experiment name]
 ```
 对于Aquila-33B模型:
 ```
-bash dist_trigger_docker.sh hostfile Aquila-sft.yaml aquila-33b [实验名 experiment name]
+bash dist_trigger_docker.sh hostfile Aquila-chat.yaml aquila-33b [实验名 experiment name]
 ```
 接下来会输出下列信息，注意`NODES_NUM`应该与节点数相等，`LOGFILE`是模型运行的日志文件；The following information will be output. Note that `NODES_NUM` should be equal to the number of nodes, and `LOGFILE` is the log file for the model run.