Upload 18 files

README.md

@@ -0,0 +1,252 @@
+## *Recreation of the Original Model Card As It Was: [08/28/23](https://web.archive.org/web/20230828190052/https://huggingface.co/Qwen/Qwen-7B)*
+
+[](#qwen-7b)Qwen-7B
+===================
+
+![](https://web.archive.org/web/20230828190052im_/https://qianwen-res.oss-cn-beijing.aliyuncs.com/logo.jpg)
+
+  
+
+Qwen-7B [🤖](https://modelscope.cn/models/qwen/Qwen-7B/summary) | [🤗](https://huggingface.co/Qwen/Qwen-7B)  ｜ Qwen-7B-Chat [🤖](https://modelscope.cn/models/qwen/Qwen-7B-Chat/summary) | [🤗](https://huggingface.co/Qwen/Qwen-7B-Chat)  | Qwen-7B-Chat-Int4 [🤗](https://huggingface.co/Qwen/Qwen-7B-Chat-Int4)  
+[WeChat](https://github.com/QwenLM/Qwen-7B/blob/main/assets/wechat.png)   |   [Discord](https://discord.gg/z3GAxXZ9Ce)   |   [Demo](https://modelscope.cn/studios/qwen/Qwen-7B-Chat-Demo/summary)  ｜  [Report](https://github.com/QwenLM/Qwen-7B/blob/main/tech_memo.md)
+
+  
+
+[](#介绍-introduction)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的仓库。
+
+通义千问-7B（Qwen-7B）主要有以下特点：
+
+1.  **大规模高质量训练语料**：使用超过2.2万亿tokens的数据进行预训练，包含高质量中、英、多语言、代码、数学等数据，涵盖通用及专业领域的训练语料。通过大量对比实验对预训练语料分布进行了优化。
+2.  **强大的性能**：Qwen-7B在多个中英文下游评测任务上（涵盖常识推理、代码、数学、翻译等），效果显著超越现有的相近规模开源模型，甚至在部分指标上相比更大尺寸模型也有较强竞争力。具体评测结果请详见下文。
+3.  **覆盖更全面的词表**：相比目前以中英词表为主的开源模型，Qwen-7B使用了约15万大小的词表。该词表对多语言更加友好，方便用户在不扩展词表的情况下对部分语种进行能力增强和扩展。
+
+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](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B)。
+
+**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.
+
+The features of Qwen-7B include:
+
+1.  **Large-scale high-quality training corpora**: It is pretrained on over 2.2 trillion tokens, including Chinese, English, multilingual texts, code, and mathematics, covering general and professional fields. The distribution of the pre-training corpus has been optimized through a large number of ablation experiments.
+2.  **Competitive performance**: It significantly surpasses existing open-source models of similar scale on multiple Chinese and English downstream evaluation tasks (including commonsense, reasoning, code, mathematics, etc.), and even surpasses some larger-scale models in several benchmarks. See below for specific evaluation results.
+3.  **More comprehensive vocabulary coverage**: Compared with other open-source models based on Chinese and English vocabularies, Qwen-7B uses a vocabulary of over 150K tokens. This vocabulary is more friendly to multiple languages, enabling users to directly further enhance the capability for certain languages without expanding the vocabulary.
+
+For more details about the open-source model of Qwen-7B, please refer to the [Github](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B) code repository.
+
+[](#要求（requirements）)要求（Requirements）
+-------------------------------------
+
+*   python 3.8及以上版本
+*   pytorch 1.12及以上版本，推荐2.0及以上版本
+*   建议使用CUDA 11.4及以上（GPU用户、flash-attention用户等需考虑此选项）
+*   python 3.8 and above
+*   pytorch 1.12 and above, 2.0 and above are recommended
+*   CUDA 11.4 and above are recommended (this is for GPU users, flash-attention users, etc.)
+
+[](#依赖项-dependency)依赖项 (Dependency)
+-----------------------------------
+
+运行Qwen-7B，请确保满足上述要求，再执行以下pip命令安装依赖库
+
+To run Qwen-7B, please make sure you meet the above requirements, and then execute the following pip commands to install the dependent libraries.
+
+    pip install transformers==4.31.0 accelerate tiktoken einops
+    
+
+另��，推荐安装`flash-attention`库，以实现更高的效率和更低的显存占用。
+
+In addition, it is recommended to install the `flash-attention` library for higher efficiency and lower memory usage.
+
+    git clone -b v1.0.8 https://github.com/Dao-AILab/flash-attention
+    cd flash-attention && pip install .
+    # 下方安装可选，安装可能比较缓慢。
+    # Below are optional. Installing them might be slow.
+    # pip install csrc/layer_norm
+    # pip install csrc/rotary
+    
+
+[](#快速使用（quickstart）)快速使用（Quickstart）
+-------------------------------------
+
+您可以通过以下代码轻松调用：
+
+You can easily call the model with the following code:
+
+    from transformers import AutoModelForCausalLM, AutoTokenizer
+    from transformers.generation import GenerationConfig
+    
+    # Note: The default behavior now has injection attack prevention off.
+    tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen-7B", trust_remote_code=True)
+    
+    # use bf16
+    # model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen-7B", device_map="auto", trust_remote_code=True, bf16=True).eval()
+    # use fp16
+    # model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen-7B", device_map="auto", trust_remote_code=True, fp16=True).eval()
+    # use cpu only
+    # model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen-7B", device_map="cpu", trust_remote_code=True).eval()
+    # use auto mode, automatically select precision based on the device.
+    model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen-7B", device_map="auto", trust_remote_code=True).eval()
+    
+    # Specify hyperparameters for generation
+    model.generation_config = GenerationConfig.from_pretrained("Qwen/Qwen-7B", trust_remote_code=True)
+    
+    inputs = tokenizer('蒙古国的首都是乌兰巴托（Ulaanbaatar）\n冰岛的首都是雷克雅未克（Reykjavik）\n埃塞俄比亚的首都是', return_tensors='pt')
+    inputs = inputs.to(model.device)
+    pred = model.generate(**inputs)
+    print(tokenizer.decode(pred.cpu()[0], skip_special_tokens=True))
+    # 蒙古国的首都是乌兰巴托（Ulaanbaatar）\n冰岛的首都是雷克雅未克（Reykjavik）\n埃塞俄比亚的首都是亚的斯亚贝巴（Addis Ababa）...
+    
+
+关于更多的使用说明，请参考我们的[Github repo](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B)获取更多信息。
+
+For more information, please refer to our [Github repo](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B) for more information.
+
+[](#tokenizer)Tokenizer
+-----------------------
+
+> 注：作为术语的“tokenization”在中文中尚无共识的概念对应，本文档采用英文表达以利说明。
+
+基于tiktoken的分词器有别于其他分词器，比如sentencepiece分词器。尤其在微调阶段，需要特别注意特殊token的使用。关于tokenizer的更多信息，以及微调时涉及的相关使用，请参阅[文档](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B/blob/main/tokenization_note_zh.md)。
+
+Our tokenizer based on tiktoken is different from other tokenizers, e.g., sentencepiece tokenizer. You need to pay attention to special tokens, especially in finetuning. For more detailed information on the tokenizer and related use in fine-tuning, please refer to the [documentation](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B/blob/main/tokenization_note.md).
+
+[](#模型细节-model)模型细节 (Model)
+---------------------------
+
+Qwen-7B模型规模基本情况如下所示：
+
+The details of the model architecture of Qwen-7B are listed as follows:
+
+![](https://files.catbox.moe/cgwcel.png)
+
+在位置编码、FFN激活函数和normalization的实现方式上，我们也采用了目前最流行的做法， 即RoPE相对位置编码、SwiGLU激活函数、RMSNorm（可选安装flash-attention加速）。
+
+在分词器方面，相比目前主流开源模型以中英词表为主，Qwen-7B使用了超过15万token大小的词表。 该词表在GPT-4使用的BPE词表`cl100k_base`基础上，对中文、多语言进行了优化，在对中、英、代码数据的高效编解码的基础上，对部分多语言更加友好，方便用户在不扩展词表的情况下对部分语种进行能力增强。 词表对数字按单个数字位切分。调用较为高效的[tiktoken分词库](https://web.archive.org/web/20230828190052/https://github.com/openai/tiktoken)进行分词。
+
+我们从部分语种各随机抽取100万个文档语料，以对比不同模型的编码压缩率（以支持100语种的XLM-R为基准值1，越低越好），具体性能见图。
+
+可以看到Qwen-7B在保持中英代码高效解码的前提下，对部分使用人群较多的语种（泰语th、希伯来语he、阿拉伯语ar、韩语ko、越南语vi、日语ja、土耳其语tr、印尼语id、波兰语pl、俄语ru、荷兰语nl、葡萄牙语pt、意大利语it、德语de、西班牙语es、法语fr等）上也实现了较高的压缩率，使得模型在这些语种上也具备较强的可扩展性和较高的训练和推理效率。
+
+在预训练数据方面，Qwen-7B模型一方面利用了部分开源通用语料， 另一方面也积累了海量全网语料以及高质量文本内容，去重及过滤后的语料超过2.2T tokens。 囊括全网文本、百科、书籍、代码、数学及各个领域垂类。
+
+![](/web/20230828190052im_/https://huggingface.co/Qwen/assets/tokenizer.png)
+
+For position encoding, FFN activation function, and normalization 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 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://web.archive.org/web/20230828190052/https://github.com/openai/tiktoken) tokenizer library for efficient tokenization.
+
+We randomly selected 1 million document corpus of each language to test and compare the encoding compression rates of different models (with XLM-R, which supports 100 languages, as the base value 1). The specific performance is shown in the figure above.
+
+As can be seen, while ensuring the efficient decoding of Chinese, English, and code, Qwen-7B also achieves a high compression rate for many other languages (such as th, he, ar, ko, vi, ja, tr, id, pl, ru, nl, pt, it, de, es, fr etc.), equipping the model with strong scalability as well as high training and inference efficiency in these languages.
+
+For pre-training data, on the one hand, Qwen-7B uses part of the open-source generic corpus. On the other hand, it uses a massive amount of accumulated web corpus and high-quality text content. The scale of corpus reaches over 2.2T tokens after deduplication and filtration, encompassing web text, encyclopedias, books, code, mathematics, and various domain.
+
+[](#评测效果（evaluation）)评测效果（Evaluation）
+-------------------------------------
+
+### [](#中文评测（chinese-evaluation）)中文评测（Chinese Evaluation）
+
+#### [](#c-eval)C-Eval
+
+[C-Eval](https://web.archive.org/web/20230828190052/https://arxiv.org/abs/2305.08322)是评测预训练模型中文常识能力的常用测评框架，覆盖人文、社科、理工、其他专业四个大方向共52个学科。 我们按照标准做法，以开发集样本作为few-shot来源，评价Qwen-7B预训练模型的5-shot验证集与测试集准确率。
+
+[C-Eval](https://web.archive.org/web/20230828190052/https://arxiv.org/abs/2305.08322) is a common evaluation benchmark for testing the common sense capability of pre-trained models in Chinese. It covers 52 subjects in four major directions: humanities, social sciences, STEM, and other specialties. According to the standard practice, we use the development set samples as the source of few-shot, to evaluate the 5-shot validation set and test set accuracy of the Qwen-7B pre-trained model.
+
+在C-Eval验证集上，Qwen-7B模型和其他模型的准确率对比如下：
+
+The accuracy comparison of Qwen-7B and the other models on the C-Eval validation set is shown as follows:
+
+![](https://files.catbox.moe/deyqre.png)
+
+在C-Eval测试集上，Qwen-7B预训练模型与其他模型的效果对比如下表所示：
+
+The performance comparison of Qwen-7B and other models on the C-Eval test set is shown in the following table:
+
+![](https://files.catbox.moe/gudsnj.png)
+
+可以看到，Qwen-7B在同等规模现有模型中取得了最高的分数，甚至相比更大规模模型也具有较强竞争力。
+
+As can be seen, Qwen-7B achieves the best performance out of all existing models with similar scale and even surpasses larger-scale models.
+
+### [](#英文评测（english-evaluation）)英文评测（English Evaluation）
+
+#### [](#mmlu)MMLU
+
+[MMLU](https://web.archive.org/web/20230828190052/https://arxiv.org/abs/2009.03300)是目前评测英文综合能力最权威的基准评测之一，同样覆盖了不同学科领域、不同难度层级的57个子任务。
+
+Qwen-7B在MMLU 5-shot准确率表现如下表：
+
+[MMLU](https://web.archive.org/web/20230828190052/https://arxiv.org/abs/2009.03300) is currently one of the most recognized benchmarks for evaluating English comprehension abilities, covering 57 subtasks across different academic fields and difficulty levels. The MMLU 5-shot accuracy performance of Qwen-7B is shown in the following table:
+
+![](https://files.catbox.moe/v62y13.png)
+
+在英文方面，Qwen-7B的效果同样超过了目前国内外其他同类开源预训练模型，同样对比更大规模版本的模型也具有较��竞争力。
+
+In terms of English, Qwen-7B also surpasses other similar open-source pre-trained models, and is competitive when compared to larger versions of other models.
+
+### [](#代码评测（coding-evaluation）)代码评测（Coding Evaluation）
+
+我们在[HumanEval](https://web.archive.org/web/20230828190052/https://github.com/openai/human-eval)（0-shot）上对比预训练模型的代码能力，结果如下：
+
+We compared the code capabilities of pre-trained models on [HumanEval](https://web.archive.org/web/20230828190052/https://github.com/openai/human-eval), and the results are as follows:
+
+![](https://files.catbox.moe/z1u5dl.png)
+
+### [](#数学评测（mathematics-evaluation）)数学评测（Mathematics Evaluation）
+
+数学能力使用常用的[GSM8K](https://web.archive.org/web/20230828190052/https://github.com/openai/grade-school-math)数据集（8-shot）评价：
+
+We compared the math capabilities of pre-trained models on [GSM8K](https://web.archive.org/web/20230828190052/https://github.com/openai/grade-school-math) (8-shot), and the results are as follows:
+
+![](https://files.catbox.moe/ff7wt3.png)
+
+### [](#翻译评测（translation-evaluation）)翻译评测（Translation Evaluation）
+
+我们使用[WMT22](https://web.archive.org/web/20230828190052/https://www.statmt.org/wmt22/translation-task.html)中-英（zh-en）和英-中（en-zh）数据集（5-shot BLEU）评测：
+
+We compared the translation capabilities of pre-trained models on [WMT22](https://web.archive.org/web/20230828190052/https://www.statmt.org/wmt22/translation-task.html) zh-en and en-zh (5-shot BLEU), and the results are as follows:
+
+![](https://files.catbox.moe/keskkd.png)
+
+### [](#长序列评测（long-context-evaluation）)长序列评测（Long-Context Evaluation）
+
+我们引入NTK插值，LogN注意力缩放，窗口注意力等技巧，将模型的上下文长度扩展到8K以上。在arXiv数据上使用PPL指标测试Qwen-7B在不同长度下的表现，结果如下：
+
+**(若要启用NTK和LogN注意力缩放，请将config.json里的`use_dynamic_ntk`和`use_logn_attn`设置为true)**
+
+We introduce NTK-aware interpolation, LogN attention scaling, Window attention, etc. to extend the context length to over 8K tokens. We conduct language modeling experiments on the arXiv dataset with the PPL evaluation. Results are demonstrated below:
+
+**(To use NTK interpolation and LogN scaling, please set `use_dynamic_ntk` and `use_long_attn` to true in config.json.)**
+
+![](https://files.catbox.moe/d3ntwj.png)
+
+[](#评测复现（reproduction）)评测复现（Reproduction）
+-----------------------------------------
+
+我们提供了评测脚本，方便大家复现模型效果，详见[链接](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B/tree/main/eval)。提示：由于硬件和框架造成的舍入误差，复现结果如有小幅波动属于正常现象。
+
+We have provided evaluation scripts to reproduce the performance of our model, details as [link](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B/tree/main/eval).
+
+[](#faq)FAQ
+-----------
+
+如遇到问题，敬请查阅[FAQ](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B/blob/main/FAQ_zh.md)以及issue区，如仍无法解决再提交issue。
+
+If you meet problems, please refer to [FAQ](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B/blob/main/FAQ.md) and the issues first to search a solution before you launch a new issue.
+
+[](#使用协议（license-agreement）)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](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B/blob/main/LICENSE)了解具体的开源协议细节。如需商用，请填写[问卷](https://web.archive.org/web/20230828190052/https://dashscope.console.aliyun.com/openModelApply/qianwen)申请。
+
+Our code and checkpoints are open to research purpose, and they are allowed for commercial purposes. Check [LICENSE](https://web.archive.org/web/20230828190052/https://github.com/QwenLM/Qwen-7B/blob/main/LICENSE) for more details about the license. If you have requirements for commercial use, please fill out the [form](https://web.archive.org/web/20230828190052/https://dashscope.console.aliyun.com/openModelApply/qianwen) to apply.
+
+[](#联系我们（contact-us）)联系我们（Contact Us）
+-------------------------------------
+
+如果你想给我们的研发团队和产品团队留言，请通过邮件（[qianwen\_opensource@alibabacloud.com](https://web.archive.org/web/20230828190052/mailto:qianwen_opensource@alibabacloud.com)）联系我们。
+
+If you are interested to leave a message to either our research team or product team, feel free to send an email to [qianwen\_opensource@alibabacloud.com](https://web.archive.org/web/20230828190052/mailto:qianwen_opensource@alibabacloud.com).
+---------------------------

config.json

@@ -0,0 +1,37 @@
+{
+  "architectures": [
+    "QWenLMHeadModel"
+  ],
+  "auto_map": {
+    "AutoConfig": "configuration_qwen.QWenConfig",
+    "AutoModelForCausalLM": "modeling_qwen.QWenLMHeadModel"
+  },
+  "attn_dropout_prob": 0.0,
+  "bf16": false,
+  "fp16": false,
+  "fp32": false,
+  "emb_dropout_prob": 0.0,
+  "intermediate_size": 22016,
+  "initializer_range": 0.02,
+  "kv_channels": 128,
+  "layer_norm_epsilon": 1e-06,
+  "model_type": "qwen",
+  "hidden_size": 4096,
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "max_position_embeddings": 8192,
+  "no_bias": true,
+  "onnx_safe": null,
+  "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": "auto",
+  "vocab_size": 151936,
+  "use_dynamic_ntk": true,
+  "use_logn_attn": true
+}

configuration_qwen.py

@@ -0,0 +1,65 @@
+# 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"]
+
+    def __init__(
+        self,
+        vocab_size=151936,
+        hidden_size=4096,
+        num_hidden_layers=32,
+        num_attention_heads=32,
+        emb_dropout_prob=0.0,
+        attn_dropout_prob=0.0,
+        layer_norm_epsilon=1e-6,
+        initializer_range=0.02,
+        max_position_embeddings=8192,
+        scale_attn_weights=True,
+        use_cache=True,
+        bf16=False,
+        fp16=False,
+        fp32=False,
+        kv_channels=128,
+        rotary_pct=1.0,
+        rotary_emb_base=10000,
+        use_dynamic_ntk=True,
+        use_logn_attn=True,
+        use_flash_attn="auto",
+        intermediate_size=22016,
+        no_bias=True,
+        tie_word_embeddings=False,
+        **kwargs,
+    ):
+        self.vocab_size = vocab_size
+        self.hidden_size = hidden_size
+        self.intermediate_size = intermediate_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.emb_dropout_prob = emb_dropout_prob
+        self.attn_dropout_prob = attn_dropout_prob
+        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.max_position_embeddings = max_position_embeddings
+        self.bf16 = bf16
+        self.fp16 = fp16
+        self.fp32 = fp32
+        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.no_bias = no_bias
+        super().__init__(
+            tie_word_embeddings=tie_word_embeddings,
+            **kwargs
+        )

generation_config.json

@@ -0,0 +1,11 @@
+{
+  "chat_format": "raw",
+  "eos_token_id": 151643,
+  "pad_token_id": 151643,
+  "stop_words_ids": [[151643]],
+  "max_new_tokens": 512,
+  "do_sample": true,
+  "top_k": 0,
+  "top_p": 0.8,
+  "transformers_version": "4.31.0"
+}

modeling_qwen.py

@@ -0,0 +1,1207 @@
+# 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, Any, Generator
+
+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
+
+SUPPORT_CUDA = torch.cuda.is_available()
+SUPPORT_BF16 = SUPPORT_CUDA and torch.cuda.is_bf16_supported()
+SUPPORT_FP16 = SUPPORT_CUDA and torch.cuda.get_device_capability(0)[0] >= 7
+
+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"]
+
+_ERROR_BAD_CHAT_FORMAT = """\
+We detect you are probably using the pretrained model (rather than chat model) for chatting, since the chat_format in generation_config is not "chatml".
+If you are directly using the model downloaded from Huggingface, please make sure you are using our "Qwen/Qwen-7B-Chat" Huggingface model (rather than "Qwen/Qwen-7B") when you call model.chat().
+我们检测到您可能在使用预训练模型（而非chat模型）进行多轮chat，因为您当前在generation_config指定的chat_format，并未设置为我们在对话中所支持的"chatml"格式。
+如果您在直接使用我们从Huggingface提供的模型，请确保您在调用model.chat()时，使用的是"Qwen/Qwen-7B-Chat"模型（而非"Qwen/Qwen-7B"预训练模型）。
+"""
+
+_SENTINEL = object()
+_ERROR_STREAM_IN_CHAT = """\
+Pass argument `stream` to model.chat() is buggy, deprecated, and marked for removal. Please use model.chat_stream(...) instead of model.chat(..., stream=True).
+向model.chat()传入参数stream的用法可能存在Bug，该用法已被废弃，将在未来被移除。请使用model.chat_stream(...)代替model.chat(..., stream=True)。
+"""
+
+apply_rotary_emb_func = None
+rms_norm = None
+flash_attn_unpadded_func = None
+
+
+def _import_flash_attn():
+    global apply_rotary_emb_func, rms_norm, flash_attn_unpadded_func
+    try:
+        from flash_attn.layers.rotary import apply_rotary_emb_func as __apply_rotary_emb_func
+        apply_rotary_emb_func = __apply_rotary_emb_func
+    except ImportError:
+        logger.warn(
+            "Warning: import flash_attn rotary fail, please install FlashAttention rotary to get higher efficiency "
+            "https://github.com/Dao-AILab/flash-attention/tree/main/csrc/rotary"
+        )
+
+    try:
+        from flash_attn.ops.rms_norm import rms_norm as __rms_norm
+        rms_norm = __rms_norm
+    except ImportError:
+        logger.warn(
+            "Warning: import flash_attn rms_norm fail, please install FlashAttention layer_norm to get higher efficiency "
+            "https://github.com/Dao-AILab/flash-attention/tree/main/csrc/layer_norm"
+        )
+
+    try:
+        import flash_attn
+        if not hasattr(flash_attn, '__version__'):
+            from flash_attn.flash_attn_interface import flash_attn_unpadded_func as __flash_attn_unpadded_func
+        else:
+            if int(flash_attn.__version__.split(".")[0]) >= 2:
+                from flash_attn.flash_attn_interface import flash_attn_varlen_func as __flash_attn_unpadded_func
+            else:
+                from flash_attn.flash_attn_interface import flash_attn_unpadded_func as __flash_attn_unpadded_func
+        flash_attn_unpadded_func = __flash_attn_unpadded_func
+    except ImportError:
+        logger.warn(
+            "Warning: import flash_attn fail, please install FlashAttention to get higher efficiency "
+            "https://github.com/Dao-AILab/flash-attention"
+        )
+
+
+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,
+        )
+
+        new_shape = (batch_size, output.shape[0] // batch_size) + output.shape[1:]
+        output = output.view(new_shape)
+        return output
+
+
+class QWenAttention(nn.Module):
+    def __init__(self, config):
+        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.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.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
+        )
+
+        self.is_fp32 = not (config.bf16 or config.fp16)
+        if (
+            self.use_flash_attn
+            and flash_attn_unpadded_func is not None
+            and not self.is_fp32
+        ):
+            self.core_attention_flash = FlashSelfAttention(
+                causal=True, attention_dropout=config.attn_dropout_prob
+            )
+
+        self.bf16 = config.bf16
+
+
+        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.attn_dropout = nn.Dropout(config.attn_dropout_prob)
+
+    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]],
+        rotary_pos_emb: Optional[List[torch.Tensor]] = None,
+        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)
+
+        if rotary_pos_emb is not None:
+            cur_len = query.shape[1]
+            rotary_pos_emb = [i[:, -cur_len:, :, :] for i in rotary_pos_emb]
+            rotary_pos_emb = (rotary_pos_emb,) * 2
+            q_pos_emb, k_pos_emb = rotary_pos_emb
+            # Slice the pos emb for current inference
+            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 and not self.training:
+            if self.logn_tensor.device != query.device or self.logn_tensor.dtype != query.dtype:
+                self.logn_tensor = self.logn_tensor.to(query.device).type_as(query)
+            seq_start = key.size(1) - query.size(1)
+            seq_end = key.size(1)
+            logn_tensor = self.logn_tensor[:, seq_start:seq_end, :, :]
+            query = query * logn_tensor.expand_as(query)
+
+        if (
+            self.use_flash_attn
+            and flash_attn_unpadded_func is not None
+            and not self.is_fp32
+            and query.is_cuda
+        ):
+            q, k, v = query, key, value
+            context_layer = self.core_attention_flash(q, k, v)
+
+            # b s h d -> b s (h d)
+            context_layer = context_layer.flatten(2,3).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
+                and flash_attn_unpadded_func is not None
+                and not self.is_fp32
+            ):
+                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.intermediate_size // 2, bias=not config.no_bias
+        )
+        self.w2 = nn.Linear(
+            config.hidden_size, config.intermediate_size // 2, bias=not config.no_bias
+        )
+        ff_dim_in = config.intermediate_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):
+        super().__init__()
+        hidden_size = config.hidden_size
+        self.bf16 = config.bf16
+
+        self.ln_1 = RMSNorm(
+            hidden_size,
+            eps=config.layer_norm_epsilon,
+        )
+        self.attn = QWenAttention(config)
+        self.ln_2 = RMSNorm(
+            hidden_size,
+            eps=config.layer_norm_epsilon,
+        )
+
+        self.mlp = QWenMLP(config)
+
+    def forward(
+        self,
+        hidden_states: Optional[Tuple[torch.FloatTensor]],
+        rotary_pos_emb: Optional[List[torch.Tensor]] = None,
+        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,
+            rotary_pos_emb,
+            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:]
+
+        residual = hidden_states
+        layernorm_input = attn_output + residual
+
+        layernorm_output = self.ln_2(layernorm_input)
+
+        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.num_hidden_layers)
+                    ),
+                )
+
+    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.vocab_size
+        self.num_hidden_layers = config.num_hidden_layers
+        self.embed_dim = config.hidden_size
+
+        self.gradient_checkpointing = False
+        self.use_dynamic_ntk = config.use_dynamic_ntk
+        self.seq_length = config.seq_length
+
+        self.wte = nn.Embedding(self.vocab_size, self.embed_dim)
+
+        self.drop = nn.Dropout(config.emb_dropout_prob)
+
+
+        if config.rotary_pct == 1.0:
+            self.rotary_ndims = None
+        else:
+            assert config.rotary_pct < 1
+            self.rotary_ndims = int(
+                config.kv_channels * config.rotary_pct
+            )
+        dim = (
+            self.rotary_ndims
+            if self.rotary_ndims is not None
+            else config.kv_channels
+        )
+        self.rotary_emb = RotaryEmbedding(dim, base=config.rotary_emb_base)
+
+        self.h = nn.ModuleList(
+            [
+                QWenBlock(
+                    config,
+                )
+                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.num_hidden_layers)
+
+        if inputs_embeds is None:
+            inputs_embeds = self.wte(input_ids)
+        hidden_states = inputs_embeds
+
+        kv_seq_len = hidden_states.size()[1]
+        if past_key_values[0] is not None:
+            # past key values[0][0] shape: bs * seq_len * head_num * dim
+            kv_seq_len += past_key_values[0][0].shape[1]
+        if (
+            self.use_dynamic_ntk
+            and kv_seq_len == hidden_states.size()[1]
+            and not self.training
+        ):
+            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)
+        else:
+            ntk_alpha = self.rotary_emb._ntk_alpha_cached
+
+        rotary_pos_emb = self.rotary_emb(kv_seq_len, ntk_alpha=ntk_alpha)
+        for idx in range(len(rotary_pos_emb)):
+            rotary_pos_emb[idx] = rotary_pos_emb[idx].to(hidden_states.device)
+
+        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,
+                    rotary_pos_emb,
+                    None,
+                    attention_mask,
+                    head_mask[i],
+                    encoder_hidden_states,
+                    encoder_attention_mask,
+                )
+            else:
+                outputs = block(
+                    hidden_states,
+                    layer_past=layer_past,
+                    rotary_pos_emb=rotary_pos_emb,
+                    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[1],)
+
+            if output_attentions:
+                all_self_attentions = all_self_attentions + (outputs[2 if use_cache else 1],)
+
+        hidden_states = self.ln_f(hidden_states)
+        hidden_states = hidden_states.view(output_shape)
+        # Add last hidden state
+        if output_hidden_states:
+            all_hidden_states = all_hidden_states + (hidden_states,)
+
+        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)
+        assert (
+            config.bf16 + config.fp16 + config.fp32 <= 1
+        ), "Only one of \"bf16\", \"fp16\", \"fp32\" can be true"
+
+        autoset_precision = config.bf16 + config.fp16 + config.fp32 == 0
+
+        if autoset_precision:
+            if SUPPORT_BF16:
+                logger.warn(
+                    "The model is automatically converting to bf16 for faster inference. "
+                    "If you want to disable the automatic precision, please manually add bf16/fp16/fp32=True to \"AutoModelForCausalLM.from_pretrained\"."
+                )
+                config.bf16 = True
+            elif SUPPORT_FP16:
+                logger.warn(
+                    "The model is automatically converting to fp16 for faster inference. "
+                    "If you want to disable the automatic precision, please manually add bf16/fp16/fp32=True to \"AutoModelForCausalLM.from_pretrained\"."
+                )
+                config.fp16 = True
+            else:
+                config.fp32 = True
+
+        if config.bf16 and SUPPORT_CUDA and not SUPPORT_BF16:
+            logger.warn("Your device does NOT seem to support bf16, you can switch to fp16 or fp32 by by passing fp16/fp32=True in \"AutoModelForCausalLM.from_pretrained\".")
+        if config.fp16 and SUPPORT_CUDA and not SUPPORT_FP16:
+            logger.warn("Your device does NOT support faster inference with fp16, please switch to fp32 which is likely to be faster")
+        if config.fp32:
+            if SUPPORT_BF16:
+                logger.warn("Your device support faster inference by passing bf16=True in \"AutoModelForCausalLM.from_pretrained\".")
+            elif SUPPORT_FP16:
+                logger.warn("Your device support faster inference by passing fp16=True in \"AutoModelForCausalLM.from_pretrained\".")
+        
+        if config.use_flash_attn == "auto":
+            if config.bf16 or config.fp16:
+                logger.warn("Try importing flash-attention for faster inference...")
+                config.use_flash_attn = True
+            else:
+                config.use_flash_attn = False
+        if config.use_flash_attn and config.fp32:
+            logger.warn("Flash attention will be disabled because it does NOT support fp32.")
+
+        if config.use_flash_attn:
+            _import_flash_attn()
+
+        self.transformer = QWenModel(config)
+        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
+
+        if config.bf16:
+            self.transformer.bfloat16()
+            self.lm_head.bfloat16()
+        if config.fp16:
+            self.transformer.half()
+            self.lm_head.half()
+        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,
+        stream: Optional[bool] = _SENTINEL,
+        stop_words_ids: Optional[List[List[int]]] = None,
+        generation_config: Optional[GenerationConfig] = None,
+        **kwargs,
+    ) -> Tuple[str, HistoryType]:
+        generation_config = generation_config if generation_config is not None else self.generation_config
+
+        assert stream is _SENTINEL, _ERROR_STREAM_IN_CHAT
+        assert generation_config.chat_format == 'chatml', _ERROR_BAD_CHAT_FORMAT
+        if history is None:
+            history = []
+        if stop_words_ids is None:
+            stop_words_ids = []
+
+        max_window_size = kwargs.get('max_window_size', None)
+        if max_window_size is None:
+            max_window_size = generation_config.max_window_size
+        raw_text, context_tokens = make_context(
+            tokenizer,
+            query,
+            history=history,
+            system=system,
+            max_window_size=max_window_size,
+            chat_format=generation_config.chat_format,
+        )
+
+        stop_words_ids.extend(get_stop_words_ids(
+            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,
+                    generation_config=generation_config,
+                    **kwargs,
+                )
+
+        response = decode_tokens(
+            outputs[0],
+            tokenizer,
+            raw_text_len=len(raw_text),
+            context_length=len(context_tokens),
+            chat_format=generation_config.chat_format,
+            verbose=False,
+            errors='replace'
+        )
+
+        if append_history:
+            history.append((query, response))
+
+        return response, history
+
+    def chat_stream(
+            self,
+            tokenizer: PreTrainedTokenizer,
+            query: str,
+            history: Optional[HistoryType],
+            system: str = "You are a helpful assistant.",
+            stop_words_ids: Optional[List[List[int]]] = None,
+            logits_processor: Optional[LogitsProcessorList] = None,
+            generation_config: Optional[GenerationConfig] = None,
+            **kwargs,
+    ) -> Generator[str, Any, None]:
+        generation_config = generation_config if generation_config is not None else self.generation_config
+        assert generation_config.chat_format == 'chatml', _ERROR_BAD_CHAT_FORMAT
+        if history is None:
+            history = []
+        if stop_words_ids is None:
+            stop_words_ids = []
+
+        max_window_size = kwargs.get('max_window_size', None)
+        if max_window_size is None:
+            max_window_size = generation_config.max_window_size
+        raw_text, context_tokens = make_context(
+            tokenizer,
+            query,
+            history=history,
+            system=system,
+            max_window_size=max_window_size,
+            chat_format=generation_config.chat_format,
+        )
+
+        stop_words_ids.extend(get_stop_words_ids(
+            generation_config.chat_format, tokenizer
+        ))
+        if stop_words_ids is not None:
+            stop_words_logits_processor = StopWordsLogitsProcessor(
+                stop_words_ids=stop_words_ids,
+                eos_token_id=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)
+        input_ids = torch.tensor([context_tokens]).to(self.device)
+
+        from transformers_stream_generator.main import NewGenerationMixin, StreamGenerationConfig
+        self.__class__.generate_stream = NewGenerationMixin.generate
+        self.__class__.sample_stream = NewGenerationMixin.sample_stream
+        stream_config = StreamGenerationConfig(**generation_config.to_dict(), do_stream=True)
+
+        def stream_generator():
+            outputs = []
+            for token in self.generate_stream(
+                    input_ids,
+                    return_dict_in_generate=False,
+                    generation_config=stream_config,
+                    logits_processor=logits_processor,
+                    seed=-1,
+                    **kwargs):
+                outputs.append(token.item())
+                yield tokenizer.decode(outputs, skip_special_tokens=True, errors='ignore')
+
+        return stream_generator()
+
+    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,
+        assistant_model: Optional["PreTrainedModel"] = None,
+        streamer: Optional["BaseStreamer"] = None,
+        **kwargs,
+    ) -> Union[GenerateOutput, torch.LongTensor]:
+        generation_config = generation_config if generation_config is not None else self.generation_config
+
+        # 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(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=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=generation_config,
+            logits_processor=logits_processor,
+            stopping_criteria=stopping_criteria,
+            prefix_allowed_tokens_fn=prefix_allowed_tokens_fn,
+            synced_gpus=synced_gpus,
+            assistant_model=assistant_model,
+            streamer=streamer,
+            **kwargs,
+        )
+
+
+class RotaryEmbedding(torch.nn.Module):
+    def __init__(self, dim, base=10000):
+        super().__init__()
+        self.dim = dim
+        self.base = base
+        self.inv_freq = 1.0 / (base ** (torch.arange(0, dim, 2).float() / dim))
+        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 = max(2 * seqlen, 16)
+            self._ntk_alpha_cached = ntk_alpha
+            seq = torch.arange(self._seq_len_cached, 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
+
+            emb = rearrange(emb, "n d -> 1 n 1 d")
+
+            cos, sin = emb.cos(), emb.sin()
+            self._rotary_pos_emb_cache = [cos, sin]
+
+    def forward(self, max_seq_len, offset=0, ntk_alpha=1.0):
+        self.update_rotary_pos_emb_cache(max_seq_len, offset, ntk_alpha)
+        cos, sin = self._rotary_pos_emb_cache
+        return [cos[:, offset : offset + max_seq_len], sin[:, 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):
+    cos, sin = freqs
+    if apply_rotary_emb_func is not None and t.is_cuda:
+        t_ = t.float()
+        cos = cos.squeeze(0).squeeze(1)[:, : cos.shape[-1] // 2]
+        sin = sin.squeeze(0).squeeze(1)[:, : sin.shape[-1] // 2]
+        output = apply_rotary_emb_func(t_, cos, sin).type_as(t)
+        return output
+    else:
+        rot_dim = freqs[0].shape[-1]
+        cos, sin = freqs
+        t_, t_pass_ = t[..., :rot_dim], t[..., rot_dim:]
+        t_ = t_.float()
+        t_pass_ = t_pass_.float()
+        t_ = (t_ * cos) + (_rotate_half(t_) * 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 and x.is_cuda:
+            return rms_norm(x, self.weight, self.eps)
+        else:
+            output = self._norm(x.float()).type_as(x)
+            return output * self.weight

pytorch_model-00001-of-00008.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7745c0eaee03bce032fce81490cfa4e334fa59bdbf14840db184bddcd3b86a35
+size 1964070447

pytorch_model-00002-of-00008.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2524b0f63b9d29bd9d90104a1fdccbf6c2dcb273e04425c9b5b6f1cf0e992fd8
+size 1933792141

pytorch_model-00003-of-00008.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a9ea1b3a9ee6f37eed063e3e5c644353309cf94bc56982828c5a6d9928a3cd4
+size 1933792141

pytorch_model-00004-of-00008.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2cc1bb2883740aa748acd9cee2d855290a1f93b13461fa73b7b8e5531a32ab7f
+size 1990406779

pytorch_model-00006-of-00008.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f46c56789137dbcd6920b7887cd639648934cf0442bc587fc05a8f7afbb02ada
+size 1933783675

pytorch_model-00007-of-00008.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e995e6f9502910e1f1dbf109df6b6a7db3733cd9adccd2d408e2f3a0d5549900
+size 1933792205

pytorch_model-00008-of-00008.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be9fc9fd671dd29258478946724308e1f63c3fbf18c689294b46532ccf0515db
+size 1829818663

pytorch_model.bin.index.json

@@ -0,0 +1,266 @@
+{
+  "metadata": {
+    "total_size": 15442649088
+  },
+  "weight_map": {
+    "lm_head.weight": "pytorch_model-00008-of-00008.bin",
+    "transformer.h.0.attn.c_attn.bias": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.0.attn.c_attn.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.0.attn.c_proj.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.0.ln_1.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.0.ln_2.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.0.mlp.c_proj.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.0.mlp.w1.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.0.mlp.w2.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.1.attn.c_attn.bias": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.1.attn.c_attn.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.1.attn.c_proj.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.1.ln_1.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.1.ln_2.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.1.mlp.c_proj.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.1.mlp.w1.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.1.mlp.w2.weight": "pytorch_model-00001-of-00008.bin",
+    "transformer.h.10.attn.c_attn.bias": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.10.attn.c_attn.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.10.attn.c_proj.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.10.ln_1.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.10.ln_2.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.10.mlp.c_proj.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.10.mlp.w1.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.10.mlp.w2.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.11.attn.c_attn.bias": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.11.attn.c_attn.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.11.attn.c_proj.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.11.ln_1.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.11.ln_2.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.11.mlp.c_proj.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.11.mlp.w1.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.11.mlp.w2.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.12.attn.c_attn.bias": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.12.attn.c_attn.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.12.attn.c_proj.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.12.ln_1.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.12.ln_2.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.12.mlp.c_proj.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.12.mlp.w1.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.12.mlp.w2.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.13.attn.c_attn.bias": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.13.attn.c_attn.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.13.attn.c_proj.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.13.ln_1.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.13.ln_2.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.13.mlp.c_proj.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.13.mlp.w1.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.13.mlp.w2.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.14.attn.c_attn.bias": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.14.attn.c_attn.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.14.attn.c_proj.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.14.ln_1.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.14.ln_2.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.14.mlp.c_proj.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.14.mlp.w1.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.14.mlp.w2.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.15.attn.c_attn.bias": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.15.attn.c_attn.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.15.attn.c_proj.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.15.ln_1.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.15.ln_2.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.15.mlp.c_proj.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.15.mlp.w1.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.15.mlp.w2.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.16.attn.c_attn.bias": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.16.attn.c_attn.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.16.attn.c_proj.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.16.ln_1.weight": "pytorch_model-00004-of-00008.bin",
+    "transformer.h.16.ln_2.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.16.mlp.c_proj.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.16.mlp.w1.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.16.mlp.w2.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.17.attn.c_attn.bias": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.17.attn.c_attn.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.17.attn.c_proj.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.17.ln_1.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.17.ln_2.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.17.mlp.c_proj.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.17.mlp.w1.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.17.mlp.w2.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.18.attn.c_attn.bias": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.18.attn.c_attn.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.18.attn.c_proj.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.18.ln_1.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.18.ln_2.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.18.mlp.c_proj.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.18.mlp.w1.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.18.mlp.w2.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.19.attn.c_attn.bias": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.19.attn.c_attn.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.19.attn.c_proj.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.19.ln_1.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.19.ln_2.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.19.mlp.c_proj.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.19.mlp.w1.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.19.mlp.w2.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.2.attn.c_attn.bias": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.2.attn.c_attn.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.2.attn.c_proj.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.2.ln_1.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.2.ln_2.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.2.mlp.c_proj.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.2.mlp.w1.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.2.mlp.w2.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.20.attn.c_attn.bias": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.20.attn.c_attn.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.20.attn.c_proj.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.20.ln_1.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.20.ln_2.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.20.mlp.c_proj.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.20.mlp.w1.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.20.mlp.w2.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.21.attn.c_attn.bias": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.21.attn.c_attn.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.21.attn.c_proj.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.21.ln_1.weight": "pytorch_model-00005-of-00008.bin",
+    "transformer.h.21.ln_2.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.21.mlp.c_proj.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.21.mlp.w1.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.21.mlp.w2.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.22.attn.c_attn.bias": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.22.attn.c_attn.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.22.attn.c_proj.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.22.ln_1.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.22.ln_2.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.22.mlp.c_proj.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.22.mlp.w1.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.22.mlp.w2.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.23.attn.c_attn.bias": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.23.attn.c_attn.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.23.attn.c_proj.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.23.ln_1.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.23.ln_2.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.23.mlp.c_proj.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.23.mlp.w1.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.23.mlp.w2.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.24.attn.c_attn.bias": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.24.attn.c_attn.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.24.attn.c_proj.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.24.ln_1.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.24.ln_2.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.24.mlp.c_proj.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.24.mlp.w1.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.24.mlp.w2.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.25.attn.c_attn.bias": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.25.attn.c_attn.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.25.attn.c_proj.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.25.ln_1.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.25.ln_2.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.25.mlp.c_proj.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.25.mlp.w1.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.25.mlp.w2.weight": "pytorch_model-00006-of-00008.bin",
+    "transformer.h.26.attn.c_attn.bias": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.26.attn.c_attn.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.26.attn.c_proj.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.26.ln_1.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.26.ln_2.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.26.mlp.c_proj.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.26.mlp.w1.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.26.mlp.w2.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.27.attn.c_attn.bias": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.27.attn.c_attn.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.27.attn.c_proj.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.27.ln_1.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.27.ln_2.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.27.mlp.c_proj.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.27.mlp.w1.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.27.mlp.w2.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.28.attn.c_attn.bias": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.28.attn.c_attn.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.28.attn.c_proj.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.28.ln_1.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.28.ln_2.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.28.mlp.c_proj.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.28.mlp.w1.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.28.mlp.w2.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.29.attn.c_attn.bias": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.29.attn.c_attn.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.29.attn.c_proj.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.29.ln_1.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.29.ln_2.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.29.mlp.c_proj.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.29.mlp.w1.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.29.mlp.w2.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.3.attn.c_attn.bias": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.3.attn.c_attn.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.3.attn.c_proj.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.3.ln_1.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.3.ln_2.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.3.mlp.c_proj.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.3.mlp.w1.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.3.mlp.w2.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.30.attn.c_attn.bias": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.30.attn.c_attn.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.30.attn.c_proj.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.30.ln_1.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.30.ln_2.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.30.mlp.c_proj.weight": "pytorch_model-00008-of-00008.bin",
+    "transformer.h.30.mlp.w1.weight": "pytorch_model-00007-of-00008.bin",
+    "transformer.h.30.mlp.w2.weight": "pytorch_model-00008-of-00008.bin",
+    "transformer.h.31.attn.c_attn.bias": "pytorch_model-00008-of-00008.bin",
+    "transformer.h.31.attn.c_attn.weight": "pytorch_model-00008-of-00008.bin",
+    "transformer.h.31.attn.c_proj.weight": "pytorch_model-00008-of-00008.bin",
+    "transformer.h.31.ln_1.weight": "pytorch_model-00008-of-00008.bin",
+    "transformer.h.31.ln_2.weight": "pytorch_model-00008-of-00008.bin",
+    "transformer.h.31.mlp.c_proj.weight": "pytorch_model-00008-of-00008.bin",
+    "transformer.h.31.mlp.w1.weight": "pytorch_model-00008-of-00008.bin",
+    "transformer.h.31.mlp.w2.weight": "pytorch_model-00008-of-00008.bin",
+    "transformer.h.4.attn.c_attn.bias": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.4.attn.c_attn.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.4.attn.c_proj.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.4.ln_1.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.4.ln_2.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.4.mlp.c_proj.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.4.mlp.w1.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.4.mlp.w2.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.5.attn.c_attn.bias": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.5.attn.c_attn.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.5.attn.c_proj.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.5.ln_1.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.5.ln_2.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.5.mlp.c_proj.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.5.mlp.w1.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.5.mlp.w2.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.6.attn.c_attn.bias": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.6.attn.c_attn.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.6.attn.c_proj.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.6.ln_1.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.6.ln_2.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.6.mlp.c_proj.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.6.mlp.w1.weight": "pytorch_model-00002-of-00008.bin",
+    "transformer.h.6.mlp.w2.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.7.attn.c_attn.bias": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.7.attn.c_attn.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.7.attn.c_proj.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.7.ln_1.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.7.ln_2.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.7.mlp.c_proj.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.7.mlp.w1.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.7.mlp.w2.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.8.attn.c_attn.bias": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.8.attn.c_attn.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.8.attn.c_proj.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.8.ln_1.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.8.ln_2.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.8.mlp.c_proj.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.8.mlp.w1.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.8.mlp.w2.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.9.attn.c_attn.bias": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.9.attn.c_attn.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.9.attn.c_proj.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.9.ln_1.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.9.ln_2.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.9.mlp.c_proj.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.9.mlp.w1.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.h.9.mlp.w2.weight": "pytorch_model-00003-of-00008.bin",
+    "transformer.ln_f.weight": "pytorch_model-00008-of-00008.bin",
+    "transformer.wte.weight": "pytorch_model-00001-of-00008.bin"
+  }
+}

qwen_generation_utils.py

@@ -0,0 +1,416 @@
+# 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, allowed_special=set()
+            ) + nl_tokens + tokenizer.encode(content, allowed_special=set())
+
+        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,
+    errors: str='replace',
+):
+    trim_decode_tokens = tokenizer.decode(tokens, errors=errors)[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,
+    errors: str='replace'
+):
+    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], errors=errors)[raw_text_len:]
+    if verbose:
+        print("\nRaw Generate w/o EOD:", tokenizer.decode(tokens, errors=errors)[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,
+    errors: str="replace",
+) -> 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,
+            errors=errors,
+        )
+    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,
+            errors=errors,
+        )
+    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**15)
+        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 +1,228 @@
+# 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."""
+
+import base64
+import logging
+import os
+import unicodedata
+from typing import Collection, Dict, List, Set, Tuple, Union
+
+import tiktoken
+from transformers import PreTrainedTokenizer, AddedToken
+
+logger = logging.getLogger(__name__)
+
+
+VOCAB_FILES_NAMES = {"vocab_file": "qwen.tiktoken"}
+
+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+"""
+ENDOFTEXT = "<|endoftext|>"
+IMSTART = "<|im_start|>"
+IMEND = "<|im_end|>"
+# as the default behavior is changed to allow special tokens in
+# regular texts, the surface forms of special tokens need to be
+# as different as possible to minimize the impact
+EXTRAS = tuple((f"<|extra_{i}|>" for i in range(205)))
+SPECIAL_TOKENS = (
+    ENDOFTEXT,
+    IMSTART,
+    IMEND,
+) + EXTRAS
+
+
+def _load_tiktoken_bpe(tiktoken_bpe_file: str) -> Dict[bytes, int]:
+    with open(tiktoken_bpe_file, "rb") as f:
+        contents = f.read()
+    return {
+        base64.b64decode(token): int(rank)
+        for token, rank in (line.split() for line in contents.splitlines() if line)
+    }
+
+class QWenTokenizer(PreTrainedTokenizer):
+    """QWen tokenizer."""
+
+    vocab_files_names = VOCAB_FILES_NAMES
+
+    def __init__(
+        self,
+        vocab_file,
+        errors="replace",
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+
+        self.errors = errors  # how to handle errors in decoding
+
+        self.mergeable_ranks = _load_tiktoken_bpe(vocab_file)  # type: dict[bytes, int]
+        self.special_tokens = {
+            token: index
+            for index, token in enumerate(
+                SPECIAL_TOKENS, start=len(self.mergeable_ranks)
+            )
+        }
+
+        enc = tiktoken.Encoding(
+            "Qwen",
+            pat_str=PAT_STR,
+            mergeable_ranks=self.mergeable_ranks,
+            special_tokens=self.special_tokens,
+        )
+        assert (
+            len(self.mergeable_ranks) + len(self.special_tokens) == enc.n_vocab
+        ), f"{len(self.mergeable_ranks) + len(self.special_tokens)} != {enc.n_vocab} in encoding"
+
+        self.decoder = {
+            v: k for k, v in self.mergeable_ranks.items()
+        }  # type: dict[int, bytes|str]
+        self.decoder.update({v: k for k, v in self.special_tokens.items()})
+
+        self.tokenizer = enc  # type: tiktoken.Encoding
+
+        self.eod_id = self.tokenizer.eot_token
+        self.im_start_id = self.special_tokens[IMSTART]
+        self.im_end_id = self.special_tokens[IMEND]
+
+    def __len__(self) -> int:
+        return self.tokenizer.n_vocab
+
+    def get_vocab(self) -> Dict[bytes, int]:
+        return self.mergeable_ranks
+
+    def convert_tokens_to_ids(
+        self, tokens: Union[bytes, str, List[Union[bytes, str]]]
+    ) -> List[int]:
+        ids = []
+        if isinstance(tokens, (str, bytes)):
+            if tokens in self.special_tokens:
+                return self.special_tokens[tokens]
+            else:
+                return self.mergeable_ranks.get(tokens)
+        for token in tokens:
+            if token in self.special_tokens:
+                ids.append(self.special_tokens[token])
+            else:
+                ids.append(self.mergeable_ranks.get(token))
+        return ids
+
+    def _add_tokens(self, new_tokens: Union[List[str], List[AddedToken]], special_tokens: bool = False) -> int:
+        if not special_tokens and new_tokens:
+            raise ValueError('Adding regular tokens is not supported')
+        for token in new_tokens:
+            surface_form = token.content if isinstance(token, AddedToken) else token
+            if surface_form not in SPECIAL_TOKENS:
+                raise ValueError('Adding unknown special tokens is not supported')
+        return 0
+
+    def save_vocabulary(self, save_directory: str, **kwargs) -> Tuple[str]:
+        """
+        Save only the vocabulary of the tokenizer (vocabulary).
+
+        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,
+        allowed_special: Union[Set, str] = "all",
+        disallowed_special: Union[Collection, str] = (),
+        **kwargs,
+    ) -> List[Union[bytes, str]]:
+        """
+        Converts a string in a sequence of tokens.
+
+        Args:
+            text (`str`):
+                The sequence to be encoded.
+            allowed_special (`Literal["all"]` or `set`):
+                The surface forms of the tokens to be encoded as special tokens in regular texts.
+                Default to "all".
+            disallowed_special (`Literal["all"]` or `Collection`):
+                The surface forms of the tokens that should not be in regular texts and trigger errors.
+                Default to an empty tuple.
+
+            kwargs (additional keyword arguments, *optional*):
+                Will be passed to the underlying model specific encode method.
+
+        Returns:
+            `List[bytes|str]`: The list of tokens.
+        """
+        tokens = []
+        text = unicodedata.normalize("NFC", text)
+
+        # this implementation takes a detour: text -> token id -> token surface forms
+        for t in self.tokenizer.encode(
+            text, allowed_special=allowed_special, disallowed_special=disallowed_special
+        ):
+            tokens.append(self.decoder[t])
+        return tokens
+
+    def convert_tokens_to_string(self, tokens: List[Union[bytes, str]]) -> str:
+        """
+        Converts a sequence of tokens in a single string.
+        """
+        text = ""
+        temp = b""
+        for t in tokens:
+            if isinstance(t, str):
+                if temp:
+                    text += temp.decode("utf-8", errors=self.errors)
+                    temp = b""
+                text += t
+            elif isinstance(t, bytes):
+                temp += t
+            else:
+                raise TypeError("token should only be of type types or str")
+        if temp:
+            text += temp.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) -> Union[bytes, str]:
+        """Converts an id to a token, special tokens included"""
+        if index in self.decoder:
+            return self.decoder[index]
+        raise ValueError("unknown ids")
+
+    def _convert_token_to_id(self, token: Union[bytes, str]) -> int:
+        """Converts a token to an id using the vocab, special tokens included"""
+        if token in self.special_tokens:
+            return self.special_tokens[token]
+        if token in self.mergeable_ranks:
+            return self.mergeable_ranks[token]
+        raise ValueError("unknown token")
+
+    def _tokenize(self, text: str, **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,
+        errors: str = None,
+        **kwargs,
+    ) -> str:
+        if isinstance(token_ids, int):
+            token_ids = [token_ids]
+        if skip_special_tokens:
+            token_ids = [i for i in token_ids if i < self.eod_id]
+        return self.tokenizer.decode(token_ids, errors=errors or self.errors)

tokenizer_config.json

@@ -0,0 +1,10 @@
+{
+  "model_max_length": 8192,
+  "tokenizer_class": "QWenTokenizer",
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_qwen.QWenTokenizer",
+      null
+      ]
+  }
+}