feat(transformers): add hunyuan v1 dense/moe (v4.57.1)

[Fzilan]

add

Add the HunyuanV1Dense, HuanyuanV1MoE models, aligned with transformers v4.57.1

usage

• HunyuanV1Dense

  from transformers import AutoModelForCausalLM, AutoTokenizer
  import mindspore as ms
  
  model_name_or_path = "tencent/Hunyuan-MT-7B"
  tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
  model = AutoModelForCausalLM.from_pretrained(model_name_or_path)
  messages = [
      {"role": "user", "content": "Translate the following segment into Chinese, without additional explanation.\n\nIt’s on the house."},
  ]
  tokenized_chat = tokenizer.apply_chat_template(
      messages,
      tokenize=True,
      add_generation_prompt=False,
      return_tensors="np"
  )
  
  outputs = model.generate(ms.tensor(tokenized_chat), max_new_tokens=2048)
  output_text = tokenizer.decode(outputs[0])

  results:

  <|startoftext|>Translate the following segment into Chinese, without additional explanation.
  It’s on the house.<|extra_0|>这顿饭由我们公司来买单。<|eos|>
  

• HuanyuanV1MoE

  To be released with the official model launch.

  Currently the unit tests of the modeling script pass.

performance

Experiments are tested on Ascend Atlas 800T A2 machines with mindspore 2.7.0 pynative mode,

model	precision	weight loading time	speed
tencent/Hunyuan-MT-7B	bf16	311s	1.91 token/s

before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline?
• Did you make sure to update the documentation with your changes? E.g. record bug fixes or new features in What's New. Here are the
  documentation guidelines
• Did you build and run the code without any errors?
• Did you report the running environment (NPU type/MS version) and performance in the doc? (better record it for data loading, model inference, or training tasks)
• Did you write any new necessary tests?

[gemini-code-assist]

Summary of Changes

Hello @Fzilan, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request significantly expands the mindone/transformers library by adding comprehensive support for the HunYuan V1 family of models. It enables users to leverage both the dense and Mixture-of-Experts (MoE) versions of HunYuan V1 for various natural language processing tasks, enhancing the model diversity and capabilities of the framework.

Highlights

• New Model Integration: Introduced support for HunYuan V1 Dense and HunYuan V1 Mixture-of-Experts (MoE) models within the mindone/transformers library.
• Model Architectures: Added specific model classes for both HunYuanDenseV1 and HunYuanMoEV1, including ForCausalLM and ForSequenceClassification variants, along with their respective configurations and core components like RMSNorm, MLP, Attention, and Rotary Embeddings.
• Auto-Class Registration: Integrated the new HunYuan V1 models and their configurations into the auto classes (configuration_auto.py and modeling_auto.py) to allow for dynamic loading based on model names.
• Unit Testing: Included dedicated test files for both HunYuanDenseV1 and HunYuanMoEV1 models to ensure their correct functionality and compatibility within the framework.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request adds support for Hunyuan v1 dense and MoE models. The changes include adding the model implementations, tests, and registering them in the auto classes. The implementation is a port from PyTorch and contains some porting errors, such as using .contiguous() which is not available in MindSpore and will cause runtime errors. I've also found a bug in the MoE gate logic, a potential simplification for better readability, a typo in auto-configuration, and an incorrect model ID in a docstring example. I've provided suggestions to fix these issues.