add resource files

LICENSE

@@ -0,0 +1,54 @@
+Tongyi Qianwen LICENSE AGREEMENT
+
+Tongyi Qianwen Release Date: August 3, 2023
+
+By clicking to agree or by using or distributing any portion or element of the Tongyi Qianwen Materials, you will be deemed to have recognized and accepted the content of this Agreement, which is effective immediately.
+
+1. Definitions
+    a. This Tongyi Qianwen LICENSE AGREEMENT (this "Agreement") shall mean the terms and conditions for use, reproduction, distribution and modification of the Materials as defined by this Agreement.
+    b.  "We"(or "Us") shall mean Alibaba Cloud.
+    c.  "You" (or "Your") shall mean a natural person or legal entity exercising the rights granted by this Agreement and/or using the Materials for any purpose and in any field of use.
+    d.  "Third Parties" shall mean individuals or legal entities that are not under common control with Us or You.
+    e. "Tongyi Qianwen" shall mean the large language models (including Qwen-7b model and Qwen-7b-Chat model ), and software and
+algorithms, consisting of trained model weights, parameters (including optimizer states), machine-learning model code, inference-enabling code, training-enabling code, fine-tuning enabling code and other elements of the foregoing distributed by Us.
+    f. "Materials" shall mean, collectively, Alibaba Cloud's proprietary Tongyi Qianwen and Documentation (and any portion thereof) made available under this Agreement.
+    g.  "Source" form shall mean the preferred form for making modifications, including but not limited to model source code, documentation source, and configuration files.
+    h. "Object" form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation,
+ and conversions to other media types.
+
+2. Grant of Rights
+You are granted a non-exclusive, worldwide, non-transferable and royalty-free limited license under Alibaba Cloud's intellectual property or other rights owned by Us embodied in the Materials to use, reproduce, distribute, copy, create derivative works of, and make modifications to the Materials.
+
+3. Redistribution
+You may reproduce and distribute copies of the Materials or derivative works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions:
+    a. You shall give any other recipients of the Materials or derivative works a copy of this Agreement;
+    b. You shall cause any modified files to carry prominent notices stating that You changed the files;
+    c. You shall retain in all copies of the Materials that You distribute the following attribution notices  within a "Notice" text file distributed as a part of such copies:"Tongyi Qianwen is licensed under the Tongyi Qianwen LICENSE AGREEMENT, Copyright (c) Alibaba Cloud. All Rights Reserved."; and
+    d. You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such derivative works as a whole, provided Your use, reproduction, and distribution of the work otherwise complies with the terms and conditions of this Agreement.
+
+4. Restrictions
+If you are commercially using the Materials, and your product or service has more than 100 million monthly active users, You shall request a license from Us. You cannot exercise your rights under this Agreement without our express authorization.
+
+5. Rules of use
+    a. The Materials may be subject to export controls or restrictions in China, the United States or other countries or regions. You shall comply with applicable laws and regulations in your use of the Materials.
+    b. You can not use the Materials or any output therefrom to improve any other large language model (excluding Tongyi Qianwen or derivative works thereof).
+
+6. Intellectual Property
+    a. We retain ownership of all intellectual property rights in and to the Materials and derivatives made by or for Us. Conditioned upon compliance with the terms and conditions of this Agreement, with respect to any derivative works and modifications of the Materials that are made by you, you are and will be the owner of such derivative works and modifications.
+    b. No trademark license is granted to use the trade names, trademarks, service marks, or product names of Us, except as required to fulfill notice requirements under this Agreement or as required for reasonable and customary use in describing and redistributing the Materials.
+    c. If you commence a lawsuit or other proceedings (including a cross-claim or counterclaim in a lawsuit) against Us or any entity alleging that the Materials or any output therefrom, or any part of the foregoing, infringe any intellectual property or other right owned or licensable by you, then all licences granted to you under this Agreement shall terminate as of the date such lawsuit or other proceeding is commenced or brought.
+
+7. Disclaimer of Warranty and Limitation of Liability
+
+    a. We are not obligated to support, update, provide training for, or develop any further version of the Tongyi  Qianwen Model or to grant any license thereto.
+    b. THE MATERIALS ARE PROVIDED "AS IS" WITHOUT ANY EXPRESS OR IMPLIED WARRANTY OF ANY KIND INCLUDING WARRANTIES OF MERCHANTABILITY, NONINFRINGEMENT, OR FITNESS FOR A PARTICULAR PURPOSE. WE MAKE NO WARRANTY AND ASSUME NO RESPONSIBILITY FOR THE SAFETY OR STABILITY OF THE MATERIALS AND ANY OUTPUT THEREFROM.
+    c. IN NO EVENT SHALL WE BE LIABLE TO YOU FOR ANY DAMAGES, INCLUDING, BUT NOT LIMITED TO ANY DIRECT, OR INDIRECT, SPECIAL OR CONSEQUENTIAL DAMAGES ARISING FROM YOUR USE OR INABILITY TO USE THE MATERIALS OR ANY OUTPUT OF IT, NO MATTER HOW IT’S CAUSED.
+    d. You will indemnify and hold armless Us from and against any claim by any third party arising out of or related to your use or distribution of the Materials.
+
+8. Survival and Termination.
+    a. The term of this Agreement shall commence upon your acceptance of this Agreement or access to the Materials and will continue in full force and effect until terminated in accordance with the terms and conditions herein.
+    b. We may terminate this Agreement if you breach any of the terms or conditions of this Agreement. Upon termination of this Agreement, you must delete and cease use of the Materials. Sections 7 and 9 shall survive the termination of this Agreement.
+
+9. Governing Law and Jurisdiction.
+    a. This Agreement and any dispute arising out of or relating to it will be governed by the laws of China, without regard to conflict of law principles, and the UN Convention on Contracts for the International Sale of Goods does not apply to this Agreement.
+    b. The People's Courts in Hangzhou City shall have exclusive jurisdiction over any dispute arising out of this Agreement.

NOTICE

@@ -0,0 +1,27 @@
+------------- LICENSE FOR NVIDIA Megatron-LM code  --------------
+
+Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions
+are met:
+  * Redistributions of source code must retain the above copyright
+    notice, this list of conditions and the following disclaimer.
+  * Redistributions in binary form must reproduce the above copyright
+    notice, this list of conditions and the following disclaimer in the
+    documentation and/or other materials provided with the distribution.
+  * Neither the name of NVIDIA CORPORATION nor the names of its
+    contributors may be used to endorse or promote products derived
+    from this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
+EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
+OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

README.md

@@ -12,27 +12,28 @@ pipeline_tag: text-generation
 <p align="center">
     <img src="assets/logo.jpg" width="400"/>
 <p>
+<br>
 
 <p align="center">
-        ModelScope[Base|Chat]&nbsp ｜ &nbspHuggingface[Base|Chat]&nbsp ｜ &nbspDemo&nbsp ｜ &nbspReport
+        Qwen-7B <a href="https://modelscope.cn/models/qwen/Qwen-7B/summary">🤖 <a> | <a href="https://huggingface.co/Qwen/Qwen-7B">🤗</a>&nbsp ｜ Qwen-7B-Chat <a href="https://modelscope.cn/models/qwen/Qwen-7B-Chat/summary">🤖 <a>| <a href="https://huggingface.co/Qwen/Qwen-7B-Chat">🤗</a>&nbsp ｜ &nbspDemo&nbsp ｜ &nbsp<a href="https://github.com/QwenLM/Qwen-7B/tech_memo.md">Report</a>
 </p>
-<br><br>
+<br>
 
 ## 介绍（Introduction）
 
-通义千问-7B（`Qwen-7B`） 是阿里云研发的通义千问大模型系列的70亿参数规模的模型。`Qwen-7B`是基于Transformer的大语言模型, 在超大规模的预训练数据上进行训练得到。预训练数据类型多样，覆盖广泛，包括大量网络文本、专业书籍、代码等。同时，在`Qwen-7B`的基础上，我们使用对齐机制打造了基于大语言模型的AI助手`Qwen-7B-Chat`。本仓库为`Qwen-7B-Chat`的仓库。
+**通义千问-7B（Qwen-7B）**是阿里云研发的通义千问大模型系列的70亿参数规模的模型。Qwen-7B是基于Transformer的大语言模型, 在超大规模的预训练数据上进行训练得到。预训练数据类型多样，覆盖广泛，包括大量网络文本、专业书籍、代码等。同时，在Qwen-7B的基础上，我们使用对齐机制打造了基于大语言模型的AI助手Qwen-7B-Chat。本仓库为Qwen-7B-Chat的仓库。
 
 如果您想了解更多关于通义千问-7B开源模型的细节，我们建议您参阅Github代码库。
 
-`Qwen-7B` is the 7B-parameter version of the large language model series, Qwen (abbr. of Tongyi Qianwen), proposed by Aibaba Cloud. `Qwen-7B` is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc. Additionally, based on the pretrained `Qwen-7B`, we release `Qwen-7B-Chat`, a large-model-based AI assistant, which is trained with alignment techniques. This repository is the one for `Qwen-7B-Chat`.
+**Qwen-7B** is the 7B-parameter version of the large language model series, Qwen (abbr. Tongyi Qianwen), proposed by Aibaba Cloud. Qwen-7B`is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc. Additionally, based on the pretrained Qwen-7B, we release Qwen-7B-Chat, a large-model-based AI assistant, which is trained with alignment techniques. This repository is the one for Qwen-7B-Chat.
 
 For more details about the open-source model of Qwen-7B, please refer to the Github code repository.
 
 ## 依赖项（Dependency）
 
-运行`Qwen-7B-Chat`，请确保机器环境pytorch版本不低于1.12，再执行以下pip命令安装依赖库
+运行Qwen-7B-Chat，请确保机器环境pytorch版本不低于1.12，再执行以下pip命令安装依赖库
 
-To run `Qwen-7B-Chat`, please make sure that pytorch version is not lower than 1.12, and then execute the following pip commands to install the dependency libraries.
+To run Qwen-7B-Chat, please make sure that pytorch version is not lower than 1.12, and then execute the following pip commands to install the dependent libraries.
 
 ```bash
 pip install transformers==4.31.0 accelerate tiktoken einops
@@ -51,9 +52,9 @@ pip install csrc/rotary
 
 ## 快速使用（Quickstart）
 
-下面我们展示了一个使用`Qwen-7B-Chat`模型，进行多轮对话交互的样例：
+下面我们展示了一个使用Qwen-7B-Chat模型，进行多轮对话交互的样例：
 
-We show an example of multi-turn interaction with `Qwen-7B-Chat` in the following code:
+We show an example of multi-turn interaction with Qwen-7B-Chat in the following code:
 
 ```ipython
 >>> from transformers import AutoModelForCausalLM, AutoTokenizer
@@ -93,9 +94,9 @@ For more information, please refer to our Github repo for more information.
 
 ## 模型细节（Model）
 
-与`Qwen-7B`预训练模型相同，`Qwen-7B-Chat`模型规模基本情况如下所示
+与Qwen-7B预训练模型相同，Qwen-7B-Chat模型规模基本情况如下所示
 
-The details of the model architecture of `Qwen-7B-Chat` are listed as follows
+The details of the model architecture of Qwen-7B-Chat are listed as follows
 
 | Hyperparameter | Value |
 |:--------------:|------:|
@@ -114,17 +115,17 @@ The details of the model architecture of `Qwen-7B-Chat` are listed as follows
 
 For position encoding, FFN activation function, and normalization calculation methods, we adopt the prevalent practices, i.e., RoPE relative position encoding, SwiGLU for activation function, and RMSNorm for normalization (optional installation of flash-attention for acceleration).
 
-For tokenization, compared to the current mainstream open-source models based on Chinese and English vocabularies, `Qwen-7B-Chat` uses a vocabulary of over 150K tokens.
+For tokenization, compared to the current mainstream open-source models based on Chinese and English vocabularies, Qwen-7B-Chat uses a vocabulary of over 150K tokens.
 It first considers efficient encoding of Chinese, English, and code data, and is also more friendly to multilingual languages, enabling users to directly enhance the capability of some languages without expanding the vocabulary.
 It segments numbers by single digit, and calls the [tiktoken](https://github.com/openai/tiktoken) tokenizer library for efficient tokenization.
 
 ## 评测效果（Evaluation）
 
-对于`Qwen-7B-Chat`模型，我们同样评测了常规的中文理解（C-Eval）、英文理解（MMLU）、代码（HumanEval）和数学（GSM8K）等权威任务，同时包含了长序列任务的评测结果。由于Qwen-7B-Chat模型经过对齐后，激发了较强的外部系统调用能力，我们还进行了工具使用能力方面的评测。
+对于Qwen-7B-Chat模型，我们同样评测了常规的中文理解（C-Eval）、英文理解（MMLU）、代码（HumanEval）和数学（GSM8K）等权威任务，同时包含了长序列任务的评测结果。由于Qwen-7B-Chat模型经过对齐后，激发了较强的外部系统调用能力，我们还进行了工具使用能力方面的评测。
 
 提示：由于硬件和框架造成的舍入误差，复现结果如有波动属于正常现象。
 
-For `Qwen-7B-Chat`, we also evaluate the model on C-Eval, MMLU, HumanEval, GSM8K, etc., as well as the benchmark evaluation for long-context understanding, and tool usage.
+For Qwen-7B-Chat, we also evaluate the model on C-Eval, MMLU, HumanEval, GSM8K, etc., as well as the benchmark evaluation for long-context understanding, and tool usage.
 
 Note: Due to rounding errors caused by hardware and framework, differences in reproduced results are possible.
 
@@ -132,9 +133,9 @@ Note: Due to rounding errors caused by hardware and framework, differences in re
 
 #### C-Eval
 
-在[C-Eval](https://arxiv.org/abs/2305.08322)验证集上，我们评价了`Qwen-7B-Chat`模型的zero-shot准确率
+在[C-Eval](https://arxiv.org/abs/2305.08322)验证集上，我们评价了Qwen-7B-Chat模型的zero-shot准确率
 
-We demonstrate the zero-shot accuracy of `Qwen-7B-Chat` on C-Eval validation set
+We demonstrate the zero-shot accuracy of Qwen-7B-Chat on C-Eval validation set
 
 | Model | Avg. Acc. |
 |:--------------:|------:|
@@ -147,9 +148,9 @@ We demonstrate the zero-shot accuracy of `Qwen-7B-Chat` on C-Eval validation set
 | InternLM-7B-Chat | 53.2 |
 | **Qwen-7B-Chat** | **54.2** |
 
-C-Eval测试集上，`Qwen-7B-Chat`模型的zero-shot准确率结果如下：
+C-Eval测试集上，Qwen-7B-Chat模型的zero-shot准确率结果如下：
 
-The zero-shot accuracy of `Qwen-7B-Chat` on C-Eval testing set is provided below:
+The zero-shot accuracy of Qwen-7B-Chat on C-Eval testing set is provided below:
 
 | Model | Avg. | STEM | Social Sciences | Humanities | Others |
 |:--------------:|------:|------:|------:|------:|------:|
@@ -159,18 +160,18 @@ The zero-shot accuracy of `Qwen-7B-Chat` on C-Eval testing set is provided below
 | Baichuan-13B-Chat | 51.5 | 43.7 | 64.6 | 56.2 | 49.2 |
 | **Qwen-7B-Chat** | **54.6** | 47.8 | 67.6 | 59.3 | 50.6 |
 
-在7B规模模型上，经过人类指令对齐的`Qwen-7B-Chat`模型，准确率在同类相近规模模型中仍然处于前列。
+在7B规模模型上，经过人类指令对齐的Qwen-7B-Chat模型，准确率在同类相近规模模型中仍然处于前列。
 
-Compared with other pretrained models with comparable model size, the human-aligned `Qwen-7B-Chat` performs well in C-Eval accuracy.
+Compared with other pretrained models with comparable model size, the human-aligned Qwen-7B-Chat performs well in C-Eval accuracy.
 
 ### 英文评测（English Evaluation）
 
 #### MMLU
 
-[MMLU](https://arxiv.org/abs/2009.03300)评测集上，`Qwen-7B-Chat`模型的zero-shot准确率如下，效果同样在同类对齐模型中同样表现较优。
+[MMLU](https://arxiv.org/abs/2009.03300)评测集上，Qwen-7B-Chat模型的zero-shot准确率如下，效果同样在同类对齐模型中同样表现较优。
 
 The zero-shot accuracy of Qwen-7B-Chat on MMLU is provided below.
-The performance of `Qwen-7B-Chat` still on the top between other human-aligned models with comparable size.
+The performance of Qwen-7B-Chat still on the top between other human-aligned models with comparable size.
 
 | Model | Avg. Acc. |
 |:--------------:|------:|
@@ -185,7 +186,7 @@ The performance of `Qwen-7B-Chat` still on the top between other human-aligned m
 
 Qwen-7B-Chat在[HumanEval](https://github.com/openai/human-eval)的zero-shot Pass@1效果如下
 
-The zero-shot Pass@1 of `Qwen-7B-Chat` on [HumanEval](https://github.com/openai/human-eval) is demonstrated below
+The zero-shot Pass@1 of Qwen-7B-Chat on [HumanEval](https://github.com/openai/human-eval) is demonstrated below
 
 | Model | Pass@1 |
 |:--------------:|------:|
@@ -197,9 +198,9 @@ The zero-shot Pass@1 of `Qwen-7B-Chat` on [HumanEval](https://github.com/openai/
 
 ### 数学评测
 
-在评测数学能力的GSM8K���，`Qwen-7B-Chat`的准确率结果如下
+在评测数学能力的[GSM8K](https://github.com/openai/grade-school-math)上，Qwen-7B-Chat的准确率结果如下
 
-The accuracy of `Qwen-7B-Chat` on GSM8K is shown below
+The accuracy of Qwen-7B-Chat on GSM8K is shown below
 
 | Model | Zero-shot Acc. | 4-shot Acc. |
 |:--------------:|------:|------:|
@@ -213,20 +214,21 @@ The accuracy of `Qwen-7B-Chat` on GSM8K is shown below
 
 ### 长序列评测（Long-Context Understanding）
 
-在长文本摘要数据集[GovReport](https://arxiv.org/abs/2104.02112), [QMSum](https://arxiv.org/abs/2104.05938)和[VCSUM](https://arxiv.org/abs/2305.05280)上，`Qwen-7B-Chat`的Rouge-L结果如下：
+通过NTK插值，LogN注意力缩放可以扩展Qwen-7B-Chat的上下文长度。在长文本摘要数据集[VCSUM](https://arxiv.org/abs/2305.05280)上（文本平均长度在15K左右），Qwen-7B-Chat的Rouge-L结果如下：
 
-The Rouge-L results of `Qwen-7B-Chat` on three long-text summarization datasets ([GovReport](https://arxiv.org/abs/2104.02112), [QMSum](https://arxiv.org/abs/2104.05938), [VCSUM](https://arxiv.org/abs/2305.05280)) are shown below:
+**(若要启用这些技巧，请将config.json里的`use_dynamc_ntk`和`use_logn_attn`设置为true)**
 
-| Model | GovReport | QMSum | VCSUM (zh) |
-|----------------|-------|-------|-------|
-| GPT-3.5-Turbo-16k | 29.5 | 23.4 | 16.0 |
-| LLama2-7B-chat-4k	| 27.3 | 20.6 |	0.2 |
-| LongChat-7B-16k |	28.4 | 23.2 | 14.0 |
-| XGen-7B-8k | 27.8 |	21.7 | 1.5 |
-| InternLM-7B-8k | 9.8 | 16.8 |	13.0 |
-| ChatGLM2-6B	| 23.7 | 22.2 | 14.6 |
-| ChatGLM2-6B-32k | **33.3** | **23.9** | 16.3 |
-| **Qwen-7B-Chat** | 31.1 | 21.5 | **16.6** |
+We introduce NTK-aware interpolation, LogN attention scaling to extend the context length of Qwen-7B-Chat. The Rouge-L results of Qwen-7B-Chat on long-text summarization dataset [VCSUM](https://arxiv.org/abs/2305.05280) (The average length of this dataset is around 15K) are shown below:
+
+**(To use these tricks, please set `use_dynamic_ntk` and `use_long_attn` to true in config.json.)**
+
+| Model | VCSUM (zh) |
+|----------------|-------|
+| GPT-3.5-Turbo-16k | 16.0 |
+| LLama2-7B-Chat	|	0.2 |
+| InternLM-7B-Chat | 13.0 |
+| ChatGLM2-6B-Chat	| 16.3 |
+| **Qwen-7B-Chat** | **16.6** |
 
 ### 工具使用能力的评测（Tool Usage）
 
@@ -234,17 +236,17 @@ The Rouge-L results of `Qwen-7B-Chat` on three long-text summarization datasets
 
 千问支持通过 [ReAct Prompting](https://arxiv.org/abs/2210.03629) 调用插件/工具/API。ReAct 也是 [LangChain](https://python.langchain.com/) 框架采用的主要方式之一。在即将开源的、用于评估工具使用能力的自建评测基准上，千问的表现如下：
 
-`Qwen-7B-Chat` supports calling plugins/tools/APIs through [ReAct Prompting](https://arxiv.org/abs/2210.03629). ReAct is also one of the main approaches used by the [LangChain](https://python.langchain.com/) framework. In the soon-to-be-released evaluation benchmark for assessing tool usage capabilities, `Qwen-7B-Chat`'s performance is as follows:
+Qwen-7B-Chat supports calling plugins/tools/APIs through [ReAct Prompting](https://arxiv.org/abs/2210.03629). ReAct is also one of the main approaches used by the [LangChain](https://python.langchain.com/) framework. In the soon-to-be-released evaluation benchmark for assessing tool usage capabilities, Qwen-7B-Chat's performance is as follows:
 
-| Model | Tool Selection (Acc.↑) | Tool Input (Rouge-L↑) | False Positive Error↓ |
-|-|-|-|-|
-|GPT-4 | 95% | **0.90** | 15%
-|GPT-3.5 | 85% | 0.88 | 75%
-| **Qwen-7B-Chat** | **99%** | 0.89 | **8.5%** |
+| Model            | Tool Selection (Acc.↑) | Tool Input (Rouge-L↑) | False Positive Error↓ |
+|------------------|------------------------|-----------------------|-----------------------|
+| GPT-4            | 95%                    | **0.90**              | 15%                   |
+| GPT-3.5          | 85%                    | 0.88                  | 75%                   |
+| **Qwen-7B-Chat** | **99%**                | 0.89                  | **8.5%**              |
 
 > 评测基准中出现的插件均没有出现在千问的训练集中。该基准评估了模型在多个候选插件中选择正确插件的准确率、传入插件的参数的合理性、以及假阳率。假阳率（False Positive）定义：在处理不该调用插件的请求时，错误地调用了插件。
 
-> The plugins that appear in the evaluation set do not appear in the training set of `Qwen-7B-Chat`. This benchmark evaluates the accuracy of the model in selecting the correct plugin from multiple candidate plugins, the rationality of the parameters passed into the plugin, and the false positive rate. False Positive: Incorrectly invoking a plugin when it should not have been called when responding to a query.
+> The plugins that appear in the evaluation set do not appear in the training set of Qwen-7B-Chat. This benchmark evaluates the accuracy of the model in selecting the correct plugin from multiple candidate plugins, the rationality of the parameters passed into the plugin, and the false positive rate. False Positive: Incorrectly invoking a plugin when it should not have been called when responding to a query.
 
 关于 ReAct Prompting 的 prompt 怎么写、怎么使用，请参考 [ReAct 样例说明](examples/react_prompt.md)。使用工具能使模型更好地完成任务。基于千问的工具使用能力，我们能实现下图所展示的效果：
 
@@ -257,7 +259,7 @@ For how to write and use prompts for ReAct Prompting, please refer to [the ReAct
 
 千问还具备作为 [HuggingFace Agent](https://huggingface.co/docs/transformers/transformers_agents) 的能力。它在 Huggingface 提供的run模式评测基准上的表现如下：
 
-`Qwen-7B-Chat` also has the capability to be used as a [HuggingFace Agent](https://huggingface.co/docs/transformers/transformers_agents). Its performance on the run-mode benchmark provided by HuggingFace is as follows:
+Qwen-7B-Chat also has the capability to be used as a [HuggingFace Agent](https://huggingface.co/docs/transformers/transformers_agents). Its performance on the run-mode benchmark provided by HuggingFace is as follows:
 
 | Model | Tool Selection↑ | Tool Used↑ | Code↑ |
 |-|-|-|-|
@@ -297,7 +299,7 @@ model = AutoModelForCausalLM.from_pretrained(
 
 上述方法可以让我们将模型量化成`NF4`和`Int8`精度的模型进行读取，帮助我们节省显存开销。我们也提供了相关性能数据。我们发现尽管模型在效果上存在损失，但模型的显存开销大幅降低。
 
-With this method, it is available to load `Qwen-7B` in `NF4`and`Int8`, which saves you memory usage. We provide related statistics of model performance below. We find that the quantization downgrades the effectiveness slightly but significantly increases inference efficiency and reduces memory costs.
+With this method, it is available to load Qwen-7B-Chat in `NF4`and `Int8`, which saves you memory usage. We provide related statistics of model performance below. We find that the quantization downgrades the effectiveness slightly but significantly increases inference efficiency and reduces memory costs.
 
 | Precision | MMLU | Memory |
 | :---------: | -------: | -----: |
@@ -305,13 +307,11 @@ With this method, it is available to load `Qwen-7B` in `NF4`and`Int8`, which sav
 |   Int8   |  52.8 |   10.1G |
 |    NF4    |  48.9 |    7.4G |
 
-
-
 ## 使用协议（License Agreement）
 
 我们的代码和模型权重对学术研究完全开放，并支持商用。请查看LICENSE了解具体的开源协议细节。
 
-Our code and checkpoints are open to research purpose, and they are allowed for commercial purposes. Check LICENSE.txt for more details about the license.
+Our code and checkpoints are open to research purpose, and they are allowed for commercial purposes. Check [LICENSE](LICENSE) for more details about the license.
 
 ## 联系我们（Contact Us）
 

config.json

@@ -0,0 +1,45 @@
+{
+  "activation": "swiglu",
+  "apply_residual_connection_post_layernorm": false,
+  "architectures": [
+    "QWenLMHeadModel"
+  ],
+  "auto_map": {
+    "AutoConfig": "configuration_qwen.QWenConfig",
+    "AutoModelForCausalLM": "modeling_qwen.QWenLMHeadModel"
+  },
+  "attn_pdrop": 0.0,
+  "bf16": true,
+  "bias_dropout_fusion": true,
+  "bos_token_id": 151643,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 151643,
+  "ffn_hidden_size": 22016,
+  "fp16": false,
+  "initializer_range": 0.02,
+  "kv_channels": 128,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "qwen",
+  "n_embd": 4096,
+  "n_head": 32,
+  "n_layer": 32,
+  "n_positions": 6144,
+  "no_bias": true,
+  "onnx_safe": null,
+  "padded_vocab_size": 151936,
+  "params_dtype": "torch.bfloat16",
+  "pos_emb": "rotary",
+  "resid_pdrop": 0.1,
+  "rotary_emb_base": 10000,
+  "rotary_pct": 1.0,
+  "scale_attn_weights": true,
+  "seq_length": 2048,
+  "tie_word_embeddings": false,
+  "tokenizer_type": "QWenTokenizer",
+  "transformers_version": "4.31.0",
+  "use_cache": true,
+  "use_flash_attn": false,
+  "vocab_size": 151936,
+  "use_dynamic_ntk": false,
+  "use_logn_attn": false
+}

configuration_qwen.py

@@ -0,0 +1,74 @@
+# Copyright (c) Alibaba Cloud.
+#
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+from transformers import PretrainedConfig
+
+
+class QWenConfig(PretrainedConfig):
+    model_type = "qwen"
+    keys_to_ignore_at_inference = ["past_key_values"]
+    attribute_map = {
+        "hidden_size": "n_embd",
+        "num_attention_heads": "n_head",
+        "max_position_embeddings": "n_positions",
+        "num_hidden_layers": "n_layer",
+    }
+
+    def __init__(
+        self,
+        vocab_size=151851,
+        n_embd=4096,
+        n_layer=32,
+        n_head=32,
+        n_inner=None,
+        embd_pdrop=0.0,
+        attn_pdrop=0.0,
+        layer_norm_epsilon=1e-5,
+        initializer_range=0.02,
+        scale_attn_weights=True,
+        use_cache=True,
+        eos_token_id=151643,
+        apply_residual_connection_post_layernorm=False,
+        bf16=True,
+        kv_channels=128,
+        rotary_pct=1.0,
+        rotary_emb_base=10000,
+        use_dynamic_ntk=False,
+        use_logn_attn=False,
+        use_flash_attn=True,
+        ffn_hidden_size=22016,
+        no_bias=True,
+        tie_word_embeddings=False,
+        **kwargs,
+    ):
+        self.eos_token_id = eos_token_id
+        super().__init__(
+            eos_token_id=eos_token_id, tie_word_embeddings=tie_word_embeddings, **kwargs
+        )
+
+        self.vocab_size = vocab_size
+        self.n_embd = n_embd
+        self.n_layer = n_layer
+        self.n_head = n_head
+        self.n_inner = n_inner
+        self.embd_pdrop = embd_pdrop
+        self.attn_pdrop = attn_pdrop
+        self.layer_norm_epsilon = layer_norm_epsilon
+        self.initializer_range = initializer_range
+        self.scale_attn_weights = scale_attn_weights
+        self.use_cache = use_cache
+        self.apply_residual_connection_post_layernorm = (
+            apply_residual_connection_post_layernorm
+        )
+        self.bf16 = bf16
+        self.kv_channels = kv_channels
+        self.rotary_pct = rotary_pct
+        self.rotary_emb_base = rotary_emb_base
+        self.use_dynamic_ntk = use_dynamic_ntk
+        self.use_logn_attn = use_logn_attn
+        self.use_flash_attn = use_flash_attn
+        self.ffn_hidden_size = ffn_hidden_size
+        self.no_bias = no_bias
+        self.tie_word_embeddings = tie_word_embeddings

generation_config.json

@@ -0,0 +1,16 @@
+{
+  "chat_format": "chatml",
+  "decay_bound": 0.0,
+  "decay_factor": 1.0,
+  "eos_token_id": 151643,
+  "factual_nucleus_sampling": false,
+  "max_context_size": 1024,
+  "max_generate_size": 512,
+  "max_new_tokens": 512,
+  "pad_token_id": 151643,
+  "stop_words_ids": [[151643]],
+  "do_sample": true,
+  "top_k": 0,
+  "top_p": 0.8,
+  "transformers_version": "4.31.0"
+}

modeling_qwen.py

@@ -0,0 +1,1027 @@
+# Copyright (c) Alibaba Cloud.
+#
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+import importlib
+import math
+from typing import TYPE_CHECKING, Optional, Tuple, Union, Callable, List
+
+import torch
+import torch.nn.functional as F
+import torch.utils.checkpoint
+from torch.cuda.amp import autocast
+
+from torch.nn import CrossEntropyLoss
+from transformers import PreTrainedTokenizer, GenerationConfig, StoppingCriteriaList
+from transformers.generation.logits_process import LogitsProcessorList
+if TYPE_CHECKING:
+    from transformers.generation.streamers import BaseStreamer
+from transformers.generation.utils import GenerateOutput
+from transformers.modeling_outputs import (
+    BaseModelOutputWithPast,
+    CausalLMOutputWithPast,
+)
+from transformers.modeling_utils import PreTrainedModel
+from transformers.utils import logging
+
+try:
+    from einops import rearrange
+except ImportError:
+    rearrange = None
+from torch import nn
+
+try:
+    from flash_attn.layers.rotary import apply_rotary_emb_func
+    from einops import rearrange
+
+    use_flash_rotary = True
+    print("use flash_attn rotary")
+except ImportError:
+    use_flash_rotary = False
+    print("import flash_attn rotary fail")
+
+try:
+    from flash_attn.ops.rms_norm import rms_norm
+
+    print("use flash_attn rms_norm")
+except ImportError:
+    rms_norm = None
+    print("import flash_attn rms_norm fail")
+
+from .configuration_qwen import QWenConfig
+from .qwen_generation_utils import (
+    HistoryType,
+    make_context,
+    decode_tokens,
+    get_stop_words_ids,
+    StopWordsLogitsProcessor,
+)
+
+
+logger = logging.get_logger(__name__)
+
+_CHECKPOINT_FOR_DOC = "qwen"
+_CONFIG_FOR_DOC = "QWenConfig"
+
+QWen_PRETRAINED_MODEL_ARCHIVE_LIST = ["qwen-7b"]
+
+try:
+    from flash_attn.flash_attn_interface import flash_attn_unpadded_func
+except ImportError:
+    flash_attn_unpadded_func = None
+
+
+class FlashSelfAttention(torch.nn.Module):
+    def __init__(
+        self,
+        causal=False,
+        softmax_scale=None,
+        attention_dropout=0.0,
+    ):
+        super().__init__()
+        assert flash_attn_unpadded_func is not None, (
+            "Please install FlashAttention first, " "e.g., with pip install flash-attn"
+        )
+        assert (
+            rearrange is not None
+        ), "Please install einops first, e.g., with pip install einops"
+        self.causal = causal
+        self.softmax_scale = softmax_scale
+        self.dropout_p = attention_dropout
+
+    def forward(self, q, k, v):
+        assert all((i.dtype in [torch.float16, torch.bfloat16] for i in (q, k, v)))
+        assert all((i.is_cuda for i in (q, k, v)))
+        batch_size, seqlen_q = q.shape[0], q.shape[1]
+        seqlen_k = k.shape[1]
+        q, k, v = [rearrange(x, "b s ... -> (b s) ...") for x in [q, k, v]]
+        cu_seqlens_q = torch.arange(
+            0,
+            (batch_size + 1) * seqlen_q,
+            step=seqlen_q,
+            dtype=torch.int32,
+            device=q.device,
+        )
+
+        if self.training:
+            assert seqlen_k == seqlen_q
+
+            is_causal = self.causal
+            cu_seqlens_k = cu_seqlens_q
+        else:
+            is_causal = seqlen_q == seqlen_k
+            cu_seqlens_k = torch.arange(
+                0,
+                (batch_size + 1) * seqlen_k,
+                step=seqlen_k,
+                dtype=torch.int32,
+                device=q.device,
+            )
+            self.dropout_p = 0
+        output = flash_attn_unpadded_func(
+            q,
+            k,
+            v,
+            cu_seqlens_q,
+            cu_seqlens_k,
+            seqlen_q,
+            seqlen_k,
+            self.dropout_p,
+            softmax_scale=self.softmax_scale,
+            causal=is_causal,
+        )
+
+        output = rearrange(output, "(b s) ... -> b s ...", b=batch_size)
+        return output
+
+
+class QWenAttention(nn.Module):
+    def __init__(self, config, layer_number=None):
+        super().__init__()
+
+        max_positions = config.max_position_embeddings
+        self.register_buffer(
+            "bias",
+            torch.tril(
+                torch.ones((max_positions, max_positions), dtype=torch.bool)
+            ).view(1, 1, max_positions, max_positions),
+            persistent=False,
+        )
+        self.register_buffer("masked_bias", torch.tensor(-1e4), persistent=False)
+        self.layer_number = max(1, layer_number)
+        self.params_dtype = config.params_dtype
+        self.seq_length = config.seq_length
+
+        self.hidden_size = config.hidden_size
+        self.split_size = config.hidden_size
+        self.num_heads = config.num_attention_heads
+        self.head_dim = self.hidden_size // self.num_heads
+
+        self.use_flash_attn = config.use_flash_attn
+        self.scale_attn_weights = True
+
+        self.layer_idx = None
+
+        self.projection_size = config.kv_channels * config.num_attention_heads
+
+        assert self.projection_size % config.num_attention_heads == 0
+        self.hidden_size_per_attention_head = (
+            self.projection_size // config.num_attention_heads
+        )
+
+        self.c_attn = nn.Linear(config.hidden_size, 3 * self.projection_size)
+
+        self.c_proj = nn.Linear(
+            config.hidden_size, self.projection_size, bias=not config.no_bias
+        )
+
+        if self.use_flash_attn:
+            self.core_attention_flash = FlashSelfAttention(
+                causal=True, attention_dropout=config.attn_pdrop
+            )
+
+        self.bf16 = config.bf16
+
+        if config.rotary_pct == 1.0:
+            self.rotary_ndims = None
+        else:
+            assert config.rotary_pct < 1
+            self.rotary_ndims = int(
+                self.hidden_size_per_attention_head * config.rotary_pct
+            )
+        dim = (
+            self.rotary_ndims
+            if self.rotary_ndims is not None
+            else self.hidden_size_per_attention_head
+        )
+        self.rotary_emb = RotaryEmbedding(
+            dim, base=config.rotary_emb_base
+        )
+
+        self.use_dynamic_ntk = config.use_dynamic_ntk
+        self.use_logn_attn = config.use_logn_attn
+
+        logn_list = [math.log(i, self.seq_length) if i > self.seq_length else 1 for i in range(1, 32768)]
+        self.logn_tensor = torch.Tensor(logn_list)[None, :, None, None]
+        self._ntk_cached = 1.0
+
+        self.attn_dropout = nn.Dropout(config.attn_pdrop)
+
+    def _attn(self, query, key, value, attention_mask=None, head_mask=None):
+        attn_weights = torch.matmul(query, key.transpose(-1, -2))
+
+        if self.scale_attn_weights:
+            attn_weights = attn_weights / torch.full(
+                [],
+                value.size(-1) ** 0.5,
+                dtype=attn_weights.dtype,
+                device=attn_weights.device,
+            )
+
+        query_length, key_length = query.size(-2), key.size(-2)
+        causal_mask = self.bias[
+            :, :, key_length - query_length : key_length, :key_length
+        ]
+        mask_value = torch.finfo(attn_weights.dtype).min
+        mask_value = torch.full([], mask_value, dtype=attn_weights.dtype).to(
+            attn_weights.device
+        )
+        attn_weights = torch.where(
+            causal_mask, attn_weights.to(attn_weights.dtype), mask_value
+        )
+
+        attn_weights = nn.functional.softmax(attn_weights, dim=-1)
+
+        attn_weights = attn_weights.type(value.dtype)
+        attn_weights = self.attn_dropout(attn_weights)
+
+        if head_mask is not None:
+            attn_weights = attn_weights * head_mask
+
+        attn_output = torch.matmul(attn_weights, value)
+        attn_output = attn_output.transpose(1, 2)
+
+        return attn_output, attn_weights
+
+    def _upcast_and_reordered_attn(
+        self, query, key, value, attention_mask=None, head_mask=None
+    ):
+        bsz, num_heads, q_seq_len, dk = query.size()
+        _, _, k_seq_len, _ = key.size()
+
+        attn_weights = torch.empty(
+            bsz * num_heads,
+            q_seq_len,
+            k_seq_len,
+            dtype=torch.float32,
+            device=query.device,
+        )
+
+        scale_factor = 1.0
+        if self.scale_attn_weights:
+            scale_factor /= float(value.size(-1)) ** 0.5
+
+        with autocast(enabled=False):
+            q, k = query.reshape(-1, q_seq_len, dk), key.transpose(-1, -2).reshape(
+                -1, dk, k_seq_len
+            )
+            attn_weights = torch.baddbmm(
+                attn_weights, q.float(), k.float(), beta=0, alpha=scale_factor
+            )
+            attn_weights = attn_weights.reshape(bsz, num_heads, q_seq_len, k_seq_len)
+
+        query_length, key_length = query.size(-2), key.size(-2)
+        causal_mask = self.bias[
+            :, :, key_length - query_length : key_length, :key_length
+        ]
+        mask_value = torch.finfo(attn_weights.dtype).min
+        mask_value = torch.tensor(mask_value, dtype=attn_weights.dtype).to(
+            attn_weights.device
+        )
+        attn_weights = torch.where(causal_mask, attn_weights, mask_value)
+
+        if attention_mask is not None:
+            attn_weights = attn_weights + attention_mask
+
+        attn_weights = nn.functional.softmax(attn_weights, dim=-1)
+
+        if attn_weights.dtype != torch.float32:
+            raise RuntimeError(
+                "Error with upcasting, attn_weights does not have dtype torch.float32"
+            )
+        attn_weights = attn_weights.type(value.dtype)
+        attn_weights = self.attn_dropout(attn_weights)
+
+        if head_mask is not None:
+            attn_weights = attn_weights * head_mask
+
+        attn_output = torch.matmul(attn_weights, value)
+
+        return attn_output, attn_weights
+
+    def _split_heads(self, tensor, num_heads, attn_head_size):
+        new_shape = tensor.size()[:-1] + (num_heads, attn_head_size)
+        tensor = tensor.view(new_shape)
+        return tensor
+
+    def _merge_heads(self, tensor, num_heads, attn_head_size):
+        tensor = tensor.contiguous()
+        new_shape = tensor.size()[:-2] + (num_heads * attn_head_size,)
+        return tensor.view(new_shape)
+
+    def forward(
+        self,
+        hidden_states: Optional[Tuple[torch.FloatTensor]],
+        layer_past: Optional[Tuple[torch.Tensor]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        output_attentions: Optional[bool] = False,
+        use_cache: Optional[bool] = False,
+    ):
+
+        mixed_x_layer = self.c_attn(hidden_states)
+        query, key, value = mixed_x_layer.split(self.split_size, dim=2)
+
+        query = self._split_heads(query, self.num_heads, self.head_dim)
+        key = self._split_heads(key, self.num_heads, self.head_dim)
+        value = self._split_heads(value, self.num_heads, self.head_dim)
+
+        kv_seq_len = hidden_states.size()[1]
+        if layer_past:
+            # layer past[0] shape: bs * seq_len * head_num * dim
+            kv_seq_len += layer_past[0].shape[1]
+        if self.use_dynamic_ntk and kv_seq_len == hidden_states.size()[1]:
+            context_value = math.log(kv_seq_len / self.seq_length, 2) + 1
+            ntk_alpha = 2 ** math.ceil(context_value) - 1
+            ntk_alpha = max(ntk_alpha, 1)
+            self._ntk_cached = ntk_alpha
+        else:
+            ntk_alpha = self._ntk_cached
+        rotary_pos_emb = self.rotary_emb(kv_seq_len, ntk_alpha=ntk_alpha).to(hidden_states.device)
+
+        if rotary_pos_emb is not None:
+            if isinstance(rotary_pos_emb, tuple):
+                rotary_pos_emb = rotary_pos_emb
+            else:
+                rotary_pos_emb = (rotary_pos_emb,) * 2
+
+        if rotary_pos_emb is not None:
+            q_pos_emb, k_pos_emb = rotary_pos_emb
+            # Slice the pos emb for current inference
+            cur_len = query.shape[1]
+            q_pos_emb = q_pos_emb[:, -cur_len:, :, :]
+            k_pos_emb = k_pos_emb[:, -cur_len:, :, :]
+            query = apply_rotary_pos_emb(query, q_pos_emb)
+            key = apply_rotary_pos_emb(key, k_pos_emb)
+
+        if layer_past is not None:
+            past_key, past_value = layer_past[0], layer_past[1]
+            key = torch.cat((past_key, key), dim=1)
+            value = torch.cat((past_value, value), dim=1)
+
+        if use_cache:
+            present = (key, value)
+        else:
+            present = None
+
+        if self.use_logn_attn:
+            if self.logn_tensor.device != query.device:
+                self.logn_tensor = self.logn_tensor.to(query.device).type_as(query)
+            seq_start = key.size(0) - query.size(0)
+            seq_end = key.size(0)
+            logn_tensor = self.logn_tensor[:, seq_start:seq_end, :, :]
+            query = query * logn_tensor.expand_as(query)
+
+        if self.use_flash_attn:
+            q, k, v = query, key, value
+            context_layer = self.core_attention_flash(q, k, v)
+
+            context_layer = rearrange(
+                context_layer, "b s h d -> b s (h d)"
+            ).contiguous()
+        else:
+            query = query.permute(0, 2, 1, 3)
+            key = key.permute(0, 2, 1, 3)
+            value = value.permute(0, 2, 1, 3)
+            attn_output, attn_weight = self._attn(
+                query, key, value, attention_mask, head_mask
+            )
+            context_layer = self._merge_heads(
+                attn_output, self.num_heads, self.head_dim
+            )
+
+        attn_output = self.c_proj(context_layer)
+        outputs = (attn_output, present)
+        if output_attentions:
+            if self.use_flash_attn:
+                raise ValueError("Cannot output attentions while using flash-attn")
+            else:
+                outputs += (attn_weight,)
+
+        return outputs
+
+
+class QWenMLP(nn.Module):
+    def __init__(self, config):
+        super().__init__()
+        self.w1 = nn.Linear(
+            config.hidden_size, config.ffn_hidden_size // 2, bias=not config.no_bias
+        )
+        self.w2 = nn.Linear(
+            config.hidden_size, config.ffn_hidden_size // 2, bias=not config.no_bias
+        )
+        ff_dim_in = config.ffn_hidden_size // 2
+        self.c_proj = nn.Linear(ff_dim_in, config.hidden_size, bias=not config.no_bias)
+
+    def forward(self, hidden_states):
+        a1 = self.w1(hidden_states)
+        a2 = self.w2(hidden_states)
+        intermediate_parallel = a1 * F.silu(a2)
+        output = self.c_proj(intermediate_parallel)
+        return output
+
+
+class QWenBlock(nn.Module):
+    def __init__(self, config, layer_idx=None, num_expert=1):
+        super().__init__()
+        self.num_expert = num_expert
+        self.layer_number = layer_idx
+        self.apply_residual_connection_post_layernorm = (
+            config.apply_residual_connection_post_layernorm
+        )
+        hidden_size = config.hidden_size
+        self.apply_residual_connection_post_layernorm = (
+            config.apply_residual_connection_post_layernorm
+        )
+        self.bf16 = config.bf16
+
+        self.ln_1 = RMSNorm(
+            hidden_size,
+            eps=config.layer_norm_epsilon,
+        )
+        self.attn = QWenAttention(config, layer_number=layer_idx)
+        self.ln_2 = RMSNorm(
+            hidden_size,
+            eps=config.layer_norm_epsilon,
+        )
+
+        self.mlp = QWenMLP(config)
+
+    def forward(
+        self,
+        hidden_states: Optional[Tuple[torch.FloatTensor]],
+        layer_past: Optional[Tuple[torch.Tensor]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        use_cache: Optional[bool] = False,
+        output_attentions: Optional[bool] = False,
+    ):
+        layernorm_output = self.ln_1(hidden_states)
+
+        attn_outputs = self.attn(
+            layernorm_output,
+            layer_past=layer_past,
+            attention_mask=attention_mask,
+            head_mask=head_mask,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+        )
+        attn_output = attn_outputs[0]
+
+        outputs = attn_outputs[1:]
+
+        if self.apply_residual_connection_post_layernorm:
+            residual = layernorm_output
+        else:
+            residual = hidden_states
+        layernorm_input = attn_output + residual
+
+        layernorm_output = self.ln_2(layernorm_input)
+
+        if self.apply_residual_connection_post_layernorm:
+            residual = layernorm_output
+        else:
+            residual = layernorm_input
+
+        mlp_output = self.mlp(layernorm_output)
+        hidden_states = residual + mlp_output
+
+        if use_cache:
+            outputs = (hidden_states,) + outputs
+        else:
+            outputs = (hidden_states,) + outputs[1:]
+
+        return outputs
+
+
+class QWenPreTrainedModel(PreTrainedModel):
+    config_class = QWenConfig
+    base_model_prefix = "transformer"
+    is_parallelizable = False
+    supports_gradient_checkpointing = True
+    _no_split_modules = ["QWenBlock"]
+
+    def __init__(self, *inputs, **kwargs):
+        super().__init__(*inputs, **kwargs)
+
+    def _init_weights(self, module):
+        """Initialize the weights."""
+        if isinstance(module, nn.Linear):
+            module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
+            if module.bias is not None:
+                module.bias.data.zero_()
+        elif isinstance(module, nn.Embedding):
+            module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
+            if module.padding_idx is not None:
+                module.weight.data[module.padding_idx].zero_()
+        elif isinstance(module, RMSNorm):
+            module.weight.data.fill_(1.0)
+
+        for name, p in module.named_parameters():
+            if name == "c_proj.weight":
+                p.data.normal_(
+                    mean=0.0,
+                    std=(
+                        self.config.initializer_range
+                        / math.sqrt(2 * self.config.n_layer)
+                    ),
+                )
+
+    def _set_gradient_checkpointing(self, module, value=False):
+        if isinstance(module, QWenModel):
+            module.gradient_checkpointing = value
+
+
+class QWenModel(QWenPreTrainedModel):
+    _keys_to_ignore_on_load_missing = ["attn.masked_bias"]
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.vocab_size = config.padded_vocab_size
+        self.num_hidden_layers = config.num_hidden_layers
+        self.embed_dim = config.hidden_size
+
+        max_sequence_length = config.max_position_embeddings
+        self.position_embedding_type = config.pos_emb
+        self.gradient_checkpointing = False
+
+        if self.position_embedding_type == "learned":
+            self.wpe = nn.Embedding(max_sequence_length, self.embed_dim)
+            self.init_method(self.position_embeddings.weight)
+            self._position_embeddings_key = "position_embeddings"
+            self.init_method(self.position_embeddings.weight)
+        else:
+            self.wpe = None
+            self._position_embeddings_key = ""
+
+        self.wte = nn.Embedding(self.vocab_size, self.embed_dim)
+
+        self.drop = nn.Dropout(config.embd_pdrop)
+        self.h = nn.ModuleList(
+            [
+                QWenBlock(
+                    config,
+                    layer_idx=i,
+                )
+                for i in range(config.num_hidden_layers)
+            ]
+        )
+        self.ln_f = RMSNorm(
+            self.embed_dim,
+            eps=config.layer_norm_epsilon,
+        )
+
+        self.post_init()
+
+    def get_input_embeddings(self):
+        return self.wte
+
+    def set_input_embeddings(self, new_embeddings):
+        self.wte = new_embeddings
+
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[Tuple[Tuple[torch.Tensor]]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        token_type_ids: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ):
+        output_attentions = (
+            output_attentions
+            if output_attentions is not None
+            else self.config.output_attentions
+        )
+        output_hidden_states = (
+            output_hidden_states
+            if output_hidden_states is not None
+            else self.config.output_hidden_states
+        )
+        use_cache = use_cache if use_cache is not None else self.config.use_cache
+        return_dict = (
+            return_dict if return_dict is not None else self.config.use_return_dict
+        )
+
+        if input_ids is not None and inputs_embeds is not None:
+            raise ValueError(
+                "You cannot specify both input_ids and inputs_embeds at the same time"
+            )
+        elif input_ids is not None:
+            input_shape = input_ids.size()
+            input_ids = input_ids.view(-1, input_shape[-1])
+            batch_size = input_ids.shape[0]
+        elif inputs_embeds is not None:
+            input_shape = inputs_embeds.size()[:-1]
+            batch_size = inputs_embeds.shape[0]
+        else:
+            raise ValueError("You have to specify either input_ids or inputs_embeds")
+
+        device = input_ids.device if input_ids is not None else inputs_embeds.device
+
+        if token_type_ids is not None:
+            token_type_ids = token_type_ids.view(-1, input_shape[-1])
+        if position_ids is not None:
+            position_ids = position_ids.view(-1, input_shape[-1])
+
+        if past_key_values is None:
+            past_length = 0
+            past_key_values = tuple([None] * len(self.h))
+        else:
+            past_length = past_key_values[0][0].size(-2)
+
+        if position_ids is None:
+            position_ids = torch.arange(
+                past_length,
+                input_shape[-1] + past_length,
+                dtype=torch.long,
+                device=device,
+            )
+            position_ids = position_ids.unsqueeze(0).view(-1, input_shape[-1])
+
+        if attention_mask is not None:
+            if batch_size <= 0:
+                raise ValueError("batch_size has to be defined and > 0")
+            attention_mask = attention_mask.view(batch_size, -1)
+            attention_mask = attention_mask[:, None, None, :]
+            attention_mask = attention_mask.to(dtype=self.dtype)
+            attention_mask = (1.0 - attention_mask) * torch.finfo(self.dtype).min
+
+        encoder_attention_mask = None
+        head_mask = self.get_head_mask(head_mask, self.config.n_layer)
+
+        if inputs_embeds is None:
+            inputs_embeds = self.wte(input_ids)
+        hidden_states = inputs_embeds
+        if self.wpe is not None:
+            position_embeds = self.wpe(position_ids)
+            hidden_states = hidden_states + position_embeds
+
+        hidden_states = self.drop(hidden_states)
+        output_shape = input_shape + (hidden_states.size(-1),)
+
+        if self.gradient_checkpointing and self.training:
+            if use_cache:
+                logger.warning_once(
+                    "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..."
+                )
+                use_cache = False
+
+        presents = () if use_cache else None
+        all_self_attentions = () if output_attentions else None
+        all_hidden_states = () if output_hidden_states else None
+        for i, (block, layer_past) in enumerate(zip(self.h, past_key_values)):
+
+            if output_hidden_states:
+                all_hidden_states = all_hidden_states + (hidden_states,)
+
+            if self.gradient_checkpointing and self.training:
+
+                def create_custom_forward(module):
+                    def custom_forward(*inputs):
+                        # None for past_key_value
+                        return module(*inputs, use_cache, output_attentions)
+
+                    return custom_forward
+
+                outputs = torch.utils.checkpoint.checkpoint(
+                    create_custom_forward(block),
+                    hidden_states,
+                    None,
+                    attention_mask,
+                    head_mask[i],
+                    encoder_hidden_states,
+                    encoder_attention_mask,
+                )
+            else:
+                outputs = block(
+                    hidden_states,
+                    layer_past=layer_past,
+                    attention_mask=attention_mask,
+                    head_mask=head_mask[i],
+                    encoder_hidden_states=encoder_hidden_states,
+                    encoder_attention_mask=encoder_attention_mask,
+                    use_cache=use_cache,
+                    output_attentions=output_attentions,
+                )
+
+            hidden_states = outputs[0]
+            if use_cache is True:
+                presents = presents + (outputs[2 if output_attentions else 1],)
+
+            if output_attentions:
+                all_self_attentions = all_self_attentions + (outputs[1],)
+
+        hidden_states = self.ln_f(hidden_states)
+        hidden_states = hidden_states.view(output_shape)
+
+        if not return_dict:
+            return tuple(
+                v for v in [hidden_states, presents, all_hidden_states] if v is not None
+            )
+
+        return BaseModelOutputWithPast(
+            last_hidden_state=hidden_states,
+            past_key_values=presents,
+            hidden_states=all_hidden_states,
+            attentions=all_self_attentions,
+        )
+
+
+class QWenLMHeadModel(QWenPreTrainedModel):
+    _keys_to_ignore_on_load_missing = [r"h\.\d+\.attn\.rotary_emb\.inv_freq"]
+    _keys_to_ignore_on_load_unexpected = [r"h\.\d+\.attn\.masked_bias"]
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.transformer = QWenModel(config)
+        self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
+        self.post_init()
+
+    def get_output_embeddings(self):
+        return self.lm_head
+
+    def set_output_embeddings(self, new_embeddings):
+        self.lm_head = new_embeddings
+
+    def prepare_inputs_for_generation(
+        self, input_ids, past_key_values=None, inputs_embeds=None, **kwargs
+    ):
+        token_type_ids = kwargs.get("token_type_ids", None)
+        if past_key_values:
+            input_ids = input_ids[:, -1].unsqueeze(-1)
+            if token_type_ids is not None:
+                token_type_ids = token_type_ids[:, -1].unsqueeze(-1)
+
+        attention_mask = kwargs.get("attention_mask", None)
+        position_ids = kwargs.get("position_ids", None)
+
+        if attention_mask is not None and position_ids is None:
+            position_ids = attention_mask.long().cumsum(-1) - 1
+            position_ids.masked_fill_(attention_mask == 0, 1)
+            if past_key_values:
+                position_ids = position_ids[:, -1].unsqueeze(-1)
+        else:
+            position_ids = None
+
+        if inputs_embeds is not None and past_key_values is None:
+            model_inputs = {"inputs_embeds": inputs_embeds}
+        else:
+            model_inputs = {"input_ids": input_ids}
+
+        model_inputs.update(
+            {
+                "past_key_values": past_key_values,
+                "use_cache": kwargs.get("use_cache"),
+                "position_ids": position_ids,
+                "attention_mask": attention_mask,
+                "token_type_ids": token_type_ids,
+            }
+        )
+        return model_inputs
+
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[Tuple[Tuple[torch.Tensor]]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        token_type_ids: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: Optional[torch.FloatTensor] = None,
+        labels: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, CausalLMOutputWithPast]:
+
+        return_dict = (
+            return_dict if return_dict is not None else self.config.use_return_dict
+        )
+
+        transformer_outputs = self.transformer(
+            input_ids,
+            past_key_values=past_key_values,
+            attention_mask=attention_mask,
+            token_type_ids=token_type_ids,
+            position_ids=position_ids,
+            head_mask=head_mask,
+            inputs_embeds=inputs_embeds,
+            encoder_hidden_states=encoder_hidden_states,
+            encoder_attention_mask=encoder_attention_mask,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+        hidden_states = transformer_outputs[0]
+
+        lm_logits = self.lm_head(hidden_states)
+
+        loss = None
+        if labels is not None:
+            labels = labels.to(lm_logits.device)
+            shift_logits = lm_logits[..., :-1, :].contiguous()
+            shift_labels = labels[..., 1:].contiguous()
+            loss_fct = CrossEntropyLoss()
+            loss = loss_fct(
+                shift_logits.view(-1, shift_logits.size(-1)), shift_labels.view(-1)
+            )
+
+        if not return_dict:
+            output = (lm_logits,) + transformer_outputs[1:]
+            return ((loss,) + output) if loss is not None else output
+
+        return CausalLMOutputWithPast(
+            loss=loss,
+            logits=lm_logits,
+            past_key_values=transformer_outputs.past_key_values,
+            hidden_states=transformer_outputs.hidden_states,
+            attentions=transformer_outputs.attentions,
+        )
+
+    @staticmethod
+    def _reorder_cache(
+        past_key_values: Tuple[Tuple[torch.Tensor]], beam_idx: torch.Tensor
+    ) -> Tuple[Tuple[torch.Tensor]]:
+
+        return tuple(
+            tuple(
+                past_state.index_select(0, beam_idx.to(past_state.device))
+                for past_state in layer_past
+            )
+            for layer_past in past_key_values
+        )
+
+    def chat(
+        self,
+        tokenizer: PreTrainedTokenizer,
+        query: str,
+        history: Optional[HistoryType],
+        system: str = "You are a helpful assistant.",
+        append_history: bool = True,
+    ) -> Tuple[str, HistoryType]:
+
+        if history is None:
+            history = []
+
+        raw_text, context_tokens = make_context(
+            tokenizer,
+            query,
+            history=history,
+            system=system,
+            max_window_size=6144,
+            chat_format=self.generation_config.chat_format,
+        )
+
+        stop_words_ids = get_stop_words_ids(
+            self.generation_config.chat_format, tokenizer
+        )
+        input_ids = torch.tensor([context_tokens]).to(self.device)
+
+        outputs = self.generate(
+            input_ids,
+            stop_words_ids=stop_words_ids,
+            return_dict_in_generate=False,
+        )
+
+        response = decode_tokens(
+            outputs[0],
+            tokenizer,
+            raw_text_len=len(raw_text),
+            context_length=len(context_tokens),
+            chat_format=self.generation_config.chat_format,
+            verbose=False,
+        )
+
+        if append_history:
+            history.append((query, response))
+
+        return response, history
+
+    def generate(
+        self,
+        inputs: Optional[torch.Tensor] = None,
+        generation_config: Optional[GenerationConfig] = None,
+        logits_processor: Optional[LogitsProcessorList] = None,
+        stopping_criteria: Optional[StoppingCriteriaList] = None,
+        prefix_allowed_tokens_fn: Optional[Callable[[int, torch.Tensor], List[int]]] = None,
+        synced_gpus: Optional[bool] = None,
+        streamer: Optional["BaseStreamer"] = None,
+        **kwargs,
+    ) -> Union[GenerateOutput, torch.LongTensor]:
+        # Process stop_words_ids.
+        stop_words_ids = kwargs.pop('stop_words_ids', None)
+        if stop_words_ids is None and generation_config is not None:
+            stop_words_ids = getattr(generation_config, 'stop_words_ids', None)
+        if stop_words_ids is None:
+            stop_words_ids = getattr(self.generation_config, 'stop_words_ids', None)
+
+        if stop_words_ids is not None:
+            stop_words_logits_processor = StopWordsLogitsProcessor(
+                stop_words_ids=stop_words_ids, eos_token_id=self.generation_config.eos_token_id)
+            if logits_processor is None:
+                logits_processor = LogitsProcessorList([stop_words_logits_processor])
+            else:
+                logits_processor.append(stop_words_logits_processor)
+
+        return super().generate(
+            inputs,
+            generation_config,
+            logits_processor,
+            stopping_criteria,
+            prefix_allowed_tokens_fn,
+            synced_gpus,
+            streamer,
+            **kwargs,
+        )
+
+
+class RotaryEmbedding(torch.nn.Module):
+    def __init__(self, dim, base=10000):
+        super().__init__()
+        self.dim = dim
+        self.base = base
+        inv_freq = 1.0 / (base ** (torch.arange(0, dim, 2).float() / dim))
+        self.register_buffer("inv_freq", inv_freq)
+        if importlib.util.find_spec("einops") is None:
+            raise RuntimeError("einops is required for Rotary Embedding")
+
+        self._rotary_pos_emb_cache = None
+        self._seq_len_cached = 0
+        self._ntk_alpha_cached = 1.0
+
+    def update_rotary_pos_emb_cache(self, max_seq_len, offset=0, ntk_alpha=1.0):
+        seqlen = max_seq_len + offset
+        if seqlen > self._seq_len_cached or ntk_alpha != self._ntk_alpha_cached:
+            base = self.base * ntk_alpha ** (self.dim / (self.dim - 2))
+            self.inv_freq = 1.0 / (base ** (torch.arange(0, self.dim, 2, device=self.inv_freq.device).float() / self.dim))
+            self._seq_len_cached = seqlen
+            self._ntk_alpha_cached = ntk_alpha
+            seq = torch.arange(seqlen, device=self.inv_freq.device)
+            freqs = torch.outer(seq.type_as(self.inv_freq), self.inv_freq)
+            emb = torch.cat((freqs, freqs), dim=-1)
+            from einops import rearrange
+
+            self._rotary_pos_emb_cache = rearrange(emb, "n d -> 1 n 1 d")
+
+    def forward(self, max_seq_len, offset=0, ntk_alpha=1.0):
+        self.update_rotary_pos_emb_cache(max_seq_len, offset, ntk_alpha)
+        return self._rotary_pos_emb_cache[:, offset : offset + max_seq_len]
+
+
+def _rotate_half(x):
+    from einops import rearrange
+
+    x = rearrange(x, "... (j d) -> ... j d", j=2)
+    x1, x2 = x.unbind(dim=-2)
+    return torch.cat((-x2, x1), dim=-1)
+
+
+def apply_rotary_pos_emb(t, freqs, use_flash_rotary=False):
+    if use_flash_rotary:
+        t_ = t.float()
+        freqs = freqs.squeeze(0).squeeze(1)
+        cos = freqs[:, : freqs.shape[-1] // 2].cos()
+        sin = freqs[:, : freqs.shape[-1] // 2].sin()
+        output = apply_rotary_emb_func(t_, cos, sin).type_as(t)
+        return output
+    else:
+        rot_dim = freqs.shape[-1]
+        t_, t_pass_ = t[..., :rot_dim], t[..., rot_dim:]
+        t_ = t_.float()
+        t_pass_ = t_pass_.float()
+        t_ = (t_ * freqs.cos()) + (_rotate_half(t_) * freqs.sin())
+        return torch.cat((t_, t_pass_), dim=-1).type_as(t)
+
+
+class RMSNorm(torch.nn.Module):
+    def __init__(self, dim: int, eps: float = 1e-6):
+        super().__init__()
+        self.eps = eps
+        self.weight = nn.Parameter(torch.ones(dim))
+
+    def _norm(self, x):
+        return x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + self.eps)
+
+    def forward(self, x):
+        if rms_norm is not None:
+            return rms_norm(x, self.weight, self.eps)
+        else:
+            output = self._norm(x.float()).type_as(x)
+            return output * self.weight

qwen_generation_utils.py

@@ -0,0 +1,411 @@
+# Copyright (c) Alibaba Cloud.
+#
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+"""Generation support."""
+
+from typing import Tuple, List, Union, Iterable
+
+import numpy as np
+import torch
+import torch.nn.functional as F
+from transformers import PreTrainedTokenizer
+from transformers import logging
+from transformers.generation import LogitsProcessor
+
+logger = logging.get_logger(__name__)
+
+# Types.
+HistoryType = List[Tuple[str, str]]
+TokensType = List[int]
+BatchTokensType = List[List[int]]
+
+
+def pad_batch(batch: BatchTokensType, pad_id: int, seq_length: int) -> BatchTokensType:
+    for tokens in batch:
+        context_length = len(tokens)
+        if context_length < seq_length:
+            tokens.extend([pad_id] * (seq_length - context_length))
+    return batch
+
+
+def get_ltor_masks_and_position_ids(
+    data,
+    eod_token,
+    reset_position_ids,
+    reset_attention_mask,
+    eod_mask_loss,
+):
+    """Build masks and position id for left to right model."""
+
+    # Extract batch size and sequence length.
+    micro_batch_size, seq_length = data.size()
+
+    # Attention mask (lower triangular).
+    if reset_attention_mask:
+        att_mask_batch = micro_batch_size
+    else:
+        att_mask_batch = 1
+    attention_mask = torch.tril(
+        torch.ones((att_mask_batch, seq_length, seq_length), device=data.device)
+    ).view(att_mask_batch, 1, seq_length, seq_length)
+
+    # Loss mask.
+    loss_mask = torch.ones(data.size(), dtype=torch.float, device=data.device)
+    if eod_mask_loss:
+        loss_mask[data == eod_token] = 0.0
+
+    # Position ids.
+    position_ids = torch.arange(seq_length, dtype=torch.long, device=data.device)
+    position_ids = position_ids.unsqueeze(0).expand_as(data)
+    # We need to clone as the ids will be modifed based on batch index.
+    if reset_position_ids:
+        position_ids = position_ids.clone()
+
+    if reset_position_ids or reset_attention_mask:
+        # Loop through the batches:
+        for b in range(micro_batch_size):
+
+            # Find indecies where EOD token is.
+            eod_index = position_ids[b, data[b] == eod_token]
+            # Detach indecies from positions if going to modify positions.
+            if reset_position_ids:
+                eod_index = eod_index.clone()
+
+            # Loop through EOD indecies:
+            prev_index = 0
+            for j in range(eod_index.size()[0]):
+                i = eod_index[j]
+                # Mask attention loss.
+                if reset_attention_mask:
+                    attention_mask[b, 0, (i + 1) :, : (i + 1)] = 0
+                # Reset positions.
+                if reset_position_ids:
+                    position_ids[b, (i + 1) :] -= i + 1 - prev_index
+                    prev_index = i + 1
+
+    # Convert attention mask to binary:
+    attention_mask = attention_mask < 0.5
+
+    return attention_mask, loss_mask, position_ids
+
+
+def get_batch(context_tokens: torch.LongTensor, eod_id: int):
+    """Generate batch from context tokens."""
+    # Move to GPU.
+    tokens = context_tokens.contiguous().to(context_tokens.device)
+    # Get the attention mask and postition ids.
+    attention_mask, _, position_ids = get_ltor_masks_and_position_ids(
+        tokens,
+        eod_id,
+        reset_position_ids=False,
+        reset_attention_mask=False,
+        eod_mask_loss=False,
+    )
+    return tokens, attention_mask, position_ids
+
+
+def get_stop_words_ids(chat_format, tokenizer):
+    if chat_format == "raw":
+        stop_words_ids = [tokenizer.encode("Human:"), [tokenizer.eod_id]]
+    elif chat_format == "chatml":
+        stop_words_ids = [[tokenizer.im_end_id], [tokenizer.im_start_id]]
+    else:
+        raise NotImplementedError(f"Unknown chat format {chat_format!r}")
+    return stop_words_ids
+
+
+def make_context(
+    tokenizer: PreTrainedTokenizer,
+    query: str,
+    history: List[Tuple[str, str]] = None,
+    system: str = "",
+    max_window_size: int = 6144,
+    chat_format: str = "chatml",
+):
+    if history is None:
+        history = []
+
+    if chat_format == "chatml":
+        im_start, im_end = "<|im_start|>", "<|im_end|>"
+        im_start_tokens = [tokenizer.im_start_id]
+        im_end_tokens = [tokenizer.im_end_id]
+        nl_tokens = tokenizer.encode("\n")
+
+        def _tokenize_str(role, content):
+            return f"{role}\n{content}", tokenizer.encode(
+                role
+            ) + nl_tokens + tokenizer.encode(content)
+
+        system_text, system_tokens_part = _tokenize_str("system", system)
+        system_tokens = im_start_tokens + system_tokens_part + im_end_tokens
+
+        raw_text = ""
+        context_tokens = []
+
+        for turn_query, turn_response in reversed(history):
+            query_text, query_tokens_part = _tokenize_str("user", turn_query)
+            query_tokens = im_start_tokens + query_tokens_part + im_end_tokens
+            response_text, response_tokens_part = _tokenize_str(
+                "assistant", turn_response
+            )
+            response_tokens = im_start_tokens + response_tokens_part + im_end_tokens
+
+            next_context_tokens = nl_tokens + query_tokens + nl_tokens + response_tokens
+            prev_chat = (
+                f"\n{im_start}{query_text}{im_end}\n{im_start}{response_text}{im_end}"
+            )
+
+            current_context_size = (
+                len(system_tokens) + len(next_context_tokens) + len(context_tokens)
+            )
+            if current_context_size < max_window_size:
+                context_tokens = next_context_tokens + context_tokens
+                raw_text = prev_chat + raw_text
+            else:
+                break
+
+        context_tokens = system_tokens + context_tokens
+        raw_text = f"{im_start}{system_text}{im_end}" + raw_text
+        context_tokens += (
+            nl_tokens
+            + im_start_tokens
+            + _tokenize_str("user", query)[1]
+            + im_end_tokens
+            + nl_tokens
+            + im_start_tokens
+            + tokenizer.encode("assistant")
+            + nl_tokens
+        )
+        raw_text += f"\n{im_start}user\n{query}{im_end}\n{im_start}assistant\n"
+
+    elif chat_format == "raw":
+        raw_text = query
+        context_tokens = tokenizer.encode(raw_text)
+    else:
+        raise NotImplementedError(f"Unknown chat format {chat_format!r}")
+
+    return raw_text, context_tokens
+
+
+def _decode_default(
+    tokens: List[int],
+    *,
+    stop_words: List[str],
+    eod_words: List[str],
+    tokenizer: PreTrainedTokenizer,
+    raw_text_len: int,
+    verbose: bool = False,
+    return_end_reason: bool = False,
+):
+    trim_decode_tokens = tokenizer.decode(tokens)[raw_text_len:]
+    if verbose:
+        print("\nRaw Generate: ", trim_decode_tokens)
+
+    end_reason = f"Gen length {len(tokens)}"
+    for stop_word in stop_words:
+        trim_decode_tokens = trim_decode_tokens.replace(stop_word, "").strip()
+    for eod_word in eod_words:
+        if eod_word in trim_decode_tokens:
+            end_reason = f"Gen {eod_word!r}"
+        trim_decode_tokens = trim_decode_tokens.split(eod_word)[0]
+    trim_decode_tokens = trim_decode_tokens.strip()
+    if verbose:
+        print("\nEnd Reason:", end_reason)
+        print("\nGenerate: ", trim_decode_tokens)
+
+    if return_end_reason:
+        return trim_decode_tokens, end_reason
+    else:
+        return trim_decode_tokens
+
+
+def _decode_chatml(
+    tokens: List[int],
+    *,
+    stop_words: List[str],
+    eod_token_ids: List[int],
+    tokenizer: PreTrainedTokenizer,
+    raw_text_len: int,
+    context_length: int,
+    verbose: bool = False,
+    return_end_reason: bool = False,
+):
+    end_reason = f"Gen length {len(tokens)}"
+    eod_token_idx = context_length
+    for eod_token_idx in range(context_length, len(tokens)):
+        if tokens[eod_token_idx] in eod_token_ids:
+            end_reason = f"Gen {tokenizer.decode([tokens[eod_token_idx]])!r}"
+            break
+
+    trim_decode_tokens = tokenizer.decode(tokens[:eod_token_idx])[raw_text_len:]
+    if verbose:
+        print("\nRaw Generate w/o EOD:", tokenizer.decode(tokens)[raw_text_len:])
+        print("\nRaw Generate:", trim_decode_tokens)
+        print("\nEnd Reason:", end_reason)
+    for stop_word in stop_words:
+        trim_decode_tokens = trim_decode_tokens.replace(stop_word, "").strip()
+    trim_decode_tokens = trim_decode_tokens.strip()
+    if verbose:
+        print("\nGenerate:", trim_decode_tokens)
+
+    if return_end_reason:
+        return trim_decode_tokens, end_reason
+    else:
+        return trim_decode_tokens
+
+
+def decode_tokens(
+    tokens: Union[torch.LongTensor, TokensType],
+    tokenizer: PreTrainedTokenizer,
+    raw_text_len: int,
+    context_length: int,
+    chat_format: str,
+    verbose: bool = False,
+    return_end_reason: bool = False,
+) -> str:
+    if torch.is_tensor(tokens):
+        tokens = tokens.cpu().numpy().tolist()
+
+    if chat_format == "chatml":
+        return _decode_chatml(
+            tokens,
+            stop_words=[],
+            eod_token_ids=[tokenizer.im_start_id, tokenizer.im_end_id],
+            tokenizer=tokenizer,
+            raw_text_len=raw_text_len,
+            context_length=context_length,
+            verbose=verbose,
+            return_end_reason=return_end_reason,
+        )
+    elif chat_format == "raw":
+        return _decode_default(
+            tokens,
+            stop_words=["<|endoftext|>"],
+            eod_words=["<|endoftext|>"],
+            tokenizer=tokenizer,
+            raw_text_len=raw_text_len,
+            verbose=verbose,
+            return_end_reason=return_end_reason,
+        )
+    else:
+        raise NotImplementedError(f"Unknown chat format {chat_format!r}")
+
+
+class StopWordsLogitsProcessor(LogitsProcessor):
+    """
+    :class:`transformers.LogitsProcessor` that enforces that when specified sequences appear, stop geration.
+
+    Args:
+        stop_words_ids (:obj:`List[List[int]]`):
+            List of list of token ids of stop ids. In order to get the tokens of the words
+            that should not appear in the generated text, use :obj:`tokenizer(bad_word,
+            add_prefix_space=True).input_ids`.
+        eos_token_id (:obj:`int`):
+            The id of the `end-of-sequence` token.
+    """
+
+    def __init__(self, stop_words_ids: Iterable[Iterable[int]], eos_token_id: int):
+
+        if not isinstance(stop_words_ids, List) or len(stop_words_ids) == 0:
+            raise ValueError(
+                f"`stop_words_ids` has to be a non-emtpy list, but is {stop_words_ids}."
+            )
+        if any(not isinstance(bad_word_ids, list) for bad_word_ids in stop_words_ids):
+            raise ValueError(
+                f"`stop_words_ids` has to be a list of lists, but is {stop_words_ids}."
+            )
+        if any(
+            any(
+                (not isinstance(token_id, (int, np.integer)) or token_id < 0)
+                for token_id in stop_word_ids
+            )
+            for stop_word_ids in stop_words_ids
+        ):
+            raise ValueError(
+                f"Each list in `stop_words_ids` has to be a list of positive integers, but is {stop_words_ids}."
+            )
+
+        self.stop_words_ids = list(
+            filter(
+                lambda bad_token_seq: bad_token_seq != [eos_token_id], stop_words_ids
+            )
+        )
+        self.eos_token_id = eos_token_id
+        for stop_token_seq in self.stop_words_ids:
+            assert (
+                len(stop_token_seq) > 0
+            ), "Stop words token sequences {} cannot have an empty list".format(
+                stop_words_ids
+            )
+
+    def __call__(
+        self, input_ids: torch.LongTensor, scores: torch.FloatTensor
+    ) -> torch.FloatTensor:
+        stopped_samples = self._calc_stopped_samples(input_ids)
+        for i, should_stop in enumerate(stopped_samples):
+            if should_stop:
+                scores[i, self.eos_token_id] = float(2**30)
+        return scores
+
+    def _tokens_match(self, prev_tokens: torch.LongTensor, tokens: List[int]) -> bool:
+        if len(tokens) == 0:
+            # if bad word tokens is just one token always ban it
+            return True
+        elif len(tokens) > len(prev_tokens):
+            # if bad word tokens are longer then prev input_ids they can't be equal
+            return False
+        elif prev_tokens[-len(tokens) :].tolist() == tokens:
+            # if tokens match
+            return True
+        else:
+            return False
+
+    def _calc_stopped_samples(self, prev_input_ids: Iterable[int]) -> Iterable[int]:
+        stopped_samples = []
+        for prev_input_ids_slice in prev_input_ids:
+            match = False
+            for stop_token_seq in self.stop_words_ids:
+                if self._tokens_match(prev_input_ids_slice, stop_token_seq):
+                    # if tokens do not match continue
+                    match = True
+                    break
+            stopped_samples.append(match)
+
+        return stopped_samples
+
+
+def top_k_logits(logits, top_k=0, top_p=0.0, filter_value=-float("Inf")):
+    """This function has been mostly taken from huggingface conversational
+    ai code at
+        https://medium.com/huggingface/how-to-build-a-state-of-the-art-
+             conversational-ai-with-transfer-learning-2d818ac26313"""
+
+    if top_k > 0:
+        # Remove all tokens with a probability less than the
+        # last token of the top-k
+        indices_to_remove = logits < torch.topk(logits, top_k)[0][..., -1, None]
+        logits[indices_to_remove] = filter_value
+
+    if top_p > 0.0:
+        # Cconvert to 1D
+        sorted_logits, sorted_indices = torch.sort(logits, descending=True, dim=-1)
+        cumulative_probs = torch.cumsum(F.softmax(sorted_logits, dim=-1), dim=-1)
+
+        # Remove tokens with cumulative probability above the threshold
+        sorted_indices_to_remove = cumulative_probs > top_p
+        # Shift the indices to the right to keep also the first token
+        # above the threshold
+        sorted_indices_to_remove[..., 1:] = sorted_indices_to_remove[..., :-1].clone()
+        sorted_indices_to_remove[..., 0] = 0
+        for i in range(sorted_indices.size(0)):
+            indices_to_remove = sorted_indices[i][sorted_indices_to_remove[i]]
+            logits[i][indices_to_remove] = filter_value
+
+    return logits
+
+
+def switch(val1, val2, boolean):
+    boolean = boolean.type_as(val1)
+    return (1 - boolean) * val1 + boolean * val2

tokenization_qwen.py

@@ -0,0 +1,243 @@
+# Copyright (c) Alibaba Cloud.
+#
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+"""Tokenization classes for QWen."""
+
+from __future__ import absolute_import, division, print_function, unicode_literals
+
+import json
+import logging
+import os
+import unicodedata
+from io import open
+import base64
+import tiktoken
+from typing import List, Optional, Tuple, Union
+
+from transformers import PreTrainedTokenizer, AddedToken
+
+logger = logging.getLogger(__name__)
+
+TIKTOKEN_NAME = "qwen.tiktoken"
+
+
+class QWenTokenizer(PreTrainedTokenizer):
+    """QWen tokenizer."""
+
+    """NOTE: This tokenizer will not handle special tokens to avoid injection attacks"""
+
+    @classmethod
+    def from_pretrained(
+        cls, pretrained_model_name_or_path, cache_dir=None, *inputs, **kwargs
+    ):
+        merges_file = os.path.join(pretrained_model_name_or_path, TIKTOKEN_NAME)
+        tokenizer = cls(merges_file, *inputs, **kwargs)
+        return tokenizer
+
+    def __init__(
+        self,
+        merges_file,
+        errors="replace",
+        max_len=None,
+        unk_token="<|endoftext|>",
+        bos_token="<|endoftext|>",
+        eos_token="<|endoftext|>",
+        pad_token=None,
+        add_prefix_space=False,
+        add_bos_token=False,
+        add_more_sp_tokens=True,
+        **kwargs,
+    ):
+        bos_token = (
+            AddedToken(bos_token, lstrip=False, rstrip=False)
+            if isinstance(bos_token, str)
+            else bos_token
+        )
+        eos_token = (
+            AddedToken(eos_token, lstrip=False, rstrip=False)
+            if isinstance(eos_token, str)
+            else eos_token
+        )
+        unk_token = (
+            AddedToken(unk_token, lstrip=False, rstrip=False)
+            if isinstance(unk_token, str)
+            else unk_token
+        )
+        pad_token = (
+            AddedToken(pad_token, lstrip=False, rstrip=False)
+            if isinstance(pad_token, str)
+            else pad_token
+        )
+        super().__init__(
+            errors=errors,
+            unk_token=unk_token,
+            bos_token=bos_token,
+            eos_token=eos_token,
+            pad_token=pad_token,
+            add_prefix_space=add_prefix_space,
+            add_bos_token=add_bos_token,
+        )
+        self.add_bos_token = add_bos_token
+        self.max_len = max_len if max_len is not None else int(1e12)
+
+        self.errors = errors  # how to handle errors in decoding
+
+        name = "QWen"
+        ENDOFTEXT = "<|endoftext|>"
+        IMSTART = "<|im_start|>"
+        IMEND = "<|im_end|>"
+        if add_more_sp_tokens:
+            special_tokens = (
+                ENDOFTEXT,
+                IMSTART,
+                IMEND,
+                "<R>",
+                "<S>",
+                "<X>",
+                "<mask>",
+                "<sep>",
+            ) + tuple([f"<extra_{i}>" for i in range(200)])
+        else:
+            special_tokens = (ENDOFTEXT, IMSTART, IMEND)
+
+        PAT_STR = r"""(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\r\n\p{L}\p{N}]?\p{L}+|\p{N}| ?[^\s\p{L}\p{N}]+[\r\n]*|\s*[\r\n]+|\s+(?!\S)|\s+"""
+
+        def load_tiktoken_bpe(tiktoken_bpe_file: str) -> "dict[bytes, int]":
+            contents = open(tiktoken_bpe_file, "rb").read()
+            return {
+                base64.b64decode(token): int(rank)
+                for token, rank in (
+                    line.split() for line in contents.splitlines() if line
+                )
+            }
+
+        mergeable_ranks = load_tiktoken_bpe(merges_file)
+        special_tokens = {
+            token: index
+            for index, token in enumerate(special_tokens, start=len(mergeable_ranks))
+        }
+        self.special_tokens = special_tokens
+        enc = tiktoken.Encoding(
+            name,
+            pat_str=PAT_STR,
+            mergeable_ranks=mergeable_ranks,
+            special_tokens=special_tokens,
+        )
+        assert (
+            len(mergeable_ranks) + len(special_tokens) == enc.n_vocab
+        ), f"{len(mergeable_ranks) + len(special_tokens)} != {enc.n_vocab} in encoding"
+
+        self.mergeable_ranks = mergeable_ranks
+        self.encoder = self.mergeable_ranks
+        self.decoder = {v: k for k, v in self.encoder.items()}
+        self.tokenizer = enc  # type: tiktoken.Encoding
+        self.eod_id = self.tokenizer.eot_token
+        self.im_start_id = special_tokens[IMSTART]
+        self.im_end_id = special_tokens[IMEND]
+
+    def __len__(self):
+        return self.tokenizer.n_vocab
+
+    def get_vocab(self):
+        return self.mergeable_ranks
+
+    def convert_tokens_to_ids(self, tokens):
+        ids = []
+        # Remove support for py2
+        if isinstance(tokens, str):
+            if tokens in self.special_tokens:
+                return self.special_tokens[tokens]
+            else:
+                return self.encoder.get(tokens)
+        for token in tokens:
+            if token in self.special_tokens:
+                ids.append(self.special_tokens[token])
+            else:
+                ids.append(self.encoder.get(token))
+        if len(ids) > self.max_len:
+            logger.warning(
+                "Token indices sequence length is longer than the specified maximum "
+                " sequence length for this OpenAI GPT model ({} > {}). Running this"
+                " sequence through the model will result in indexing errors".format(
+                    len(ids), self.max_len
+                )
+            )
+        return ids
+
+    def save_vocabulary(self, save_directory: str) -> Tuple[str]:
+        """
+        Save only the vocabulary of the tokenizer (vocabulary + added tokens).
+
+        Returns:
+            `Tuple(str)`: Paths to the files saved.
+        """
+        file_path = os.path.join(save_directory, "qwen.tiktoken")
+        with open(file_path, "w", encoding="utf8") as w:
+            for k, v in self.mergeable_ranks.items():
+                line = base64.b64encode(k).decode("utf8") + " " + str(v) + "\n"
+                w.write(line)
+        return (file_path,)
+
+    def tokenize(self, text: str, **kwargs) -> List[str]:
+        """
+        Converts a string in a sequence of tokens, replacing unknown tokens with the `unk_token`.
+
+        Args:
+            text (`str`):
+                The sequence to be encoded.
+            pair (`str`, *optional*):
+                A second sequence to be encoded with the first.
+            add_special_tokens (`bool`, *optional*, defaults to `False`):
+                Whether or not to add the special tokens associated with the corresponding model.
+            kwargs (additional keyword arguments, *optional*):
+                Will be passed to the underlying model specific encode method. See details in
+                [`~PreTrainedTokenizerBase.__call__`]
+
+        Returns:
+            `List[str]`: The list of tokens.
+        """
+        tokens = []
+        text = unicodedata.normalize("NFC", text)
+        for t in self.tokenizer.encode_ordinary(text):
+            tokens.append(self.decoder[t])
+        return tokens
+
+    def convert_tokens_to_string(self, tokens: List[str]) -> str:
+        """
+        Converts a sequence of tokens in a single string. The most simple way to do it is `" ".join(tokens)` but we
+        often want to remove sub-word tokenization artifacts at the same time.
+        """
+        text = "".join(tokens)
+        text = bytearray([self.byte_decoder[c] for c in text]).decode(
+            "utf-8", errors=self.errors
+        )
+        return text
+
+    @property
+    def vocab_size(self):
+        return self.tokenizer.n_vocab
+
+    def _convert_id_to_token(self, index: int) -> str:
+        raise NotImplementedError
+
+    def _tokenize(self, text, **kwargs):
+        """
+        Converts a string in a sequence of tokens (string), using the tokenizer. Split in words for word-based
+        vocabulary or sub-words for sub-word-based vocabularies (BPE/SentencePieces/WordPieces).
+
+        Do NOT take care of added tokens.
+        """
+        raise NotImplementedError
+
+    def _decode(
+        self,
+        token_ids: Union[int, List[int]],
+        skip_special_tokens: bool = False,
+        clean_up_tokenization_spaces: bool = None,
+        **kwargs,
+    ) -> str:
+        if isinstance(token_ids, int):
+            token_ids = [token_ids]
+        return self.tokenizer.decode(token_ids)

tokenizer_config.json

@@ -0,0 +1,11 @@
+{
+  "remove_space": false,
+  "do_lower_case": false,
+  "tokenizer_class": "QWenTokenizer",
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_qwen.QWenTokenizer",
+      null
+      ]
+  }
+}