yujun

### PR types Bug fixes ### PR changes Models ### Description 1、修改artist模型的激活函数，目前gpt使用的是approximate=True的glue激活函数，而原版使用的是approximate=False的激活函数。 2、修改dallebart的tokenizer，以防出现warning。

### PR types New features ### PR changes Examples ### Description - [x] add paddle version [diffusers](https://github.com/huggingface/diffusers) repo. (this pr is based on this [branch](https://github.com/JunnYu/diffusers/tree/dfnew1011 )) - [x] add some...

[PPDiffusers] Fix ppdiffusers bug and support ZH stablediffusion

2

### PR types Bug fixes ### PR changes Models ### Description - 发布0.6.1版本 - 支持 IDEA-CCNL/Taiyi-Stable-Diffusion-1B-Chinese-v0.1 和 IDEA-CCNL/Taiyi-Stable-Diffusion-1B-Chinese-EN-v0.1 中文权重 - 修复windows环境下载模型失败的问题。 - paddlenlp的attention_mask逻辑与HF不一致，因此人工设定全为1的attention_mask。 - 增添更多的图片后缀格式，ext = ['png', 'jpg', 'jpeg', 'bmp',...

### PR types New features ### PR changes examples ### Description add text_to_image_laion400m pretrain example

### PR types Others ### PR changes Readme ### Description 添加权重转换脚本及对应的Readme

HF hub [blip_text_model] num_attention_heads is 8? [blip_vision_model] eps is 1e-5?

1

- i find in this repo https://github.com/salesforce/LAVIS/blob/main/lavis/configs/models/med_config.json num_attention_heads is 12. but in https://huggingface.co/Salesforce/blip-image-captioning-large/blob/main/config.json [blip_text_model] num_attention_heads is 8. - and blip_vision_model's eps should be 1e-6. https://github.com/salesforce/LAVIS/blob/2b6c6caf223e1a9a5139842d3191cad4166466b8/lavis/models/vit.py#L209

Can't reproduce the result in fill50k

3

Hi, i can't reproduce the result in fill50k, https://github.com/lllyasviel/ControlNet/blob/main/docs/train.md. i use the same Hyperparameter in the https://github.com/lllyasviel/ControlNet/blob/main/tutorial_train.py Could you help me to reproduce the result? Thx, here is my generated...

[Trainer] ignore_save_lr_and_optim

2

### PR types New features ### PR changes APIs ### Description 新增忽略保存lr和optim的可选项，默认值为False，跟以前一样，只有想要使用的时候指定开启。

[NEW Model] add jamba

2

### PR types New features ### PR changes Models ### Description [Jamba: A Hybrid Transformer-Mamba Language Model](https://arxiv.org/abs/2403.19887) [huggingface仓库](https://github.com/huggingface/transformers/tree/main/src/transformers/models/jamba) [huggingface_hub权重](https://huggingface.co/ai21labs/Jamba-v0.1) run.py代码 ```python import paddle from paddlenlp.transformers import JambaForCausalLM, JambaConfig, JambaTokenizer from...

[OPS] add mamba ops

2

### PR types New features ### PR changes APIs ### Description 新增mamba的自定义算子，需要依赖triton，如果不想安装torch的话。可以参照 https://github.com/zhoutianzi666/UseTritonInPaddle 文档运行。