Pinned issues
Issues
- 0
- 0
[Bug] FlashAttention Error During InternVL Fine-tuning on Tesla T4 GPU
#918 opened by kachhadiyaraj15 - 0
Finetune是否有针对训练集的衡量指标?
#917 opened by WalkerWorldPeace - 0
[Docs] 路径问题
#916 opened by Sakura-not-sleep - 0
How can I determine which region of the image the model is focusing on when answering a specific question?
#915 opened by phongkhanh - 7
lora微调后的模型需要进行合并吗
#885 opened by HiXxxSss - 1
如何得到某些文本的输出概率?
#903 opened by wzczc - 0
- 0
微调internvl-chat loral,loss从2.6594下降到0.0001,但是模型输出结果为roproproproproproproproproproproproproproproproprop ...
#910 opened by limin-al - 0
- 0
预训练上下文问题
#911 opened by samaritan1998 - 0
预训练Loss不收敛问题
#884 opened by samaritan1998 - 6
[Bug] ImportError: cannot import name 'LLAMA_ATTENTION_CLASSES' from 'transformers.models.llama.modeling_llama'
#912 opened by kachhadiyaraj15 - 1
model.forward使用方法?
#907 opened by cyj95 - 0
关于internvl2.5-1B visual grounding
#909 opened by xvjiawen - 0
[Bug] concat_pad_data_collator的pad_id为0可能有问题
#908 opened by MrChen314 - 1
[Bug] After lora ft, the lora_A and load_B tensors in ckpt remain unchanged.
#905 opened by heibaidaolx123 - 1
LORA微调8B的2.5VL所需的显存
#902 opened by liziming5353 - 1
[Docs] https://internvl.readthedocs.io/en/latest/internvl1.0/clip_benchmark.html#
#904 opened by WanderingMars - 1
关于AI2D评测中w / wo M的问题
#899 opened by Super-Shen - 0
- 2
Reusing image encodings to save inference time
#882 opened by karanl1 - 2
- 1
关于全参数多图微调的问题
#901 opened by EmeryBAI - 2
FineTune 78B model on multi-node slrum system
#898 opened by spcrobocar - 0
About the license of internVL2.5-4B model
#900 opened by NaokiAsatani - 2
关于InternVL定位多类别时的问题
#897 opened by daihuidai - 1
[Bug] Is there any official elegant implementations about the error:TypeError: 'NoneType' object is not callable
#894 opened by MengHao666 - 1
internvl2.5 78b在2张A100情况下推理DEMO卡住,无返回结果
#895 opened by Sariel2 - 2
关于InternVL的不同使用方法
#896 opened by ChenJian7578 - 1
关于InternVL2.5-MPO的Grounding 数据集制作
#890 opened by daihuidai - 4
[Bug]orjson.JSONDecodeError: unexpected end of data: line 2 column 1 (char 2)
#893 opened by MengHao666 - 3
InternVL2_5-8B模型没有输出
#891 opened by Hello-Worldd - 2
[Docs] Deploy InternVL2_5-78B-MPO
#892 opened by itsyaboyksi - 5
[Docs] import错误
#886 opened by Newbie-pz - 3
- 0
[Bug] internvl的MPO的loss一开始就为0
#889 opened by amoreZgx1n - 1
- 3
[Bug] output Lots of single "r" when inference
#870 opened by dongwhfdyer - 1
[Bug] OpenGVLab/InternVL2_5-78B-MPO Lmdeploy report bug "AttributeError: 'NoneType' object has no attribute 'split'"
#880 opened by CedrusLNZ - 1
pretrain 训练数据会开源,想从0开始训练一个模型
#878 opened by cqray1990 - 7
[Bug] 不仅仅是未能完成卷积正常初始化,同样的,新增加的MLP也无法正常初始化。
#874 opened by corkiyao - 2
- 0
[Feature] Any plans to fine tune the OpenGVLab/Mini-InternVL2-4B-DA-Medical vlm with UCSC-VLAA/MedTrinity-25M dataset ?
#881 opened by satheeshkola-532 - 0
[Bug]设置use_packed_ds=True的时候报错
#879 opened by cqray1990 - 1
Mixed Precision float16 training for InternVL2.5
#877 opened by saeedkhaki92 - 0
InternVL2.5-78b model's performance significantly degrades after AWQ quantization
#876 opened by vladimiralbrekhtccr - 1
Distillation code of InternViT
#875 opened by jiangyitong - 2
[Errors on MPO data] Some image and question are mismatched in `geomverse_extracted_pairs_vqa_correctness_rules.jsonl `
#871 opened by linhaojia13 - 0
Context extrapolation
#873 opened by jyC23333