Muzhi Zhu

## ❓ Questions and Help The link to explore the LVIS dataset on the official website (https://www.lvisdataset.org/explore) is currently unavailable.

``` File "/home/code/Key-Locked-Rank-One-Editing-for-Text-to-Image-Personalization/main.py", line 461, in model = instantiate_from_config(config.model) File "/home/code/Key-Locked-Rank-One-Editing-for-Text-to-Image-Personalization/main.py", line 137, in instantiate_from_config return get_obj_from_str(config["target"])(**config.get("params", dict())) File "/home/code/Key-Locked-Rank-One-Editing-for-Text-to-Image-Personalization/perfusion/perfusion.py", line 65, in __init__ self.embedding_manager = instantiate_from_config(personalization_config, embedder=self.cond_stage_model) File "/home/code/Key-Locked-Rank-One-Editing-for-Text-to-Image-Personalization/ldm/util.py",...

``` torchshow.save(images) Traceback (most recent call last): File "torchshow/utils.py", line 60, in tensor_to_array return x.detach().clone().cpu().numpy() TypeError: Got unsupported ScalarType BFloat16 ``` BFloat16 is currently not supported. It would be preferable...

Support for Zero3 or Zero3 Offload? Error when loading model state_dict

1

Hello, I encountered an error while trying to load a model using the following code in [`clip_encoder.py`] ```python self.vision_tower.load_state_dict(torch.load(self.clip_model), strict=False) ``` The error message is as follows: [rank0]: raise RuntimeError('Error(s)...

I notice during evaluation we use class ReferSegmDataset(torch.utils.data.Dataset): ``` if len(sents) >= self.num_classes_per_sample: sampled_inds = np.random.choice( list(range(len(sents))), size=self.num_classes_per_sample, replace=False ) else: sampled_inds = list(range(len(sents))) ``` it randomly samples 3 sentences...

SegNext require many clicks when I run demo on DAVIS test image

5

00013.jpg the checkpoint is vitb_sa2_cocolvis_hq44k_epoch_0.pth  I compare with simpleclick, which only require a few clicks  another case 00340.jpg  

[BUG] 模型进行全量微调后，loss正常，但推理时乱码

8

### 是否已有关于该错误的issue或讨论？ | Is there an existing issue / discussion for this? - [X] 我已经搜索过已有的issues和讨论 | I have searched the existing issues / discussions ### 该问题是否在FAQ中有解答？ | Is there an...

input_features values are all -1.5 in modeling_qwen2_5_omni.py when using lora sft Qwen2.5Omni

4

### Reminder - [x] I have read the above rules and searched the existing issues. ### System Info [2025-04-08 14:28:56,552] [INFO] [real_accelerator.py:219:get_accelerator] Setting ds_accelerator to cuda (auto detect) - `llamafactory`...