Repository Issue Activity (beta)

#303 请问你们用flash attention的话是用哪个版本的呢？

#80 Is this training result correct?

#305 Design bags

#304 Demo website is down

#160 The Size of ema.safetensors.

#273 What are the minimum training requirements?

#298 加载 MoT 权重时，copy_init_moe 应该设置为什么

#48 Why are torch and python old versions?

#230 Visual grounding capability in Bagel: support & evaluation?

#291 多batch推理

#295 FSDP training error: _dim_group_infos list index out of range

#245 NCCL超时

#269 Can we train Bagel with grpo？

#294 About text2img generation finetuning

#123 Unable to reproduce the results on GenEval

#293 Why need prefill for new generated text token?

#292 Model collapses after fine-tuning on the toy data

#114 Does `--resume_from` parameter matter?

#290 Time needed for fine-tuning Bagel

#288 SFT之后保存的模型大小翻倍

#140 训练edit时loss不正常

#289 请问为什么仅训练edit数据集的时候已经把vit_cond_dropout_prob，text_cond_dropout_prob和vae_cond_dropout_prob设置为0了还是有ce=None的情况呀？

#265 Mismatch between Bagel model's vocab size (152064) and Qwen2 tokenizer's vocab size (151644) causing initialization errors with custom tokens

#92 Guide to install and run Bagel on Windows 11 with RTX 5090 32GB Vram

#259 The generated picture is very blurry.