-
Hi, I'm interested in your great work.
The `./scripts/v1_5/eval/eval_all.sh` is not avalilable now. Could you release the evaluation tools? **Especially the few-shot VQA/Caption.**
And the mmc4 …
-
Hi [@JosephPai](https://github.com/JosephPai),
Thanks for your awesome repository, which serves as a great resource in this community.
Could you please consider adding our recent paper on ACL 20…
-
Hello, I would like to ask for assistance in solving a problem I've encountered.
I am currently training a MLLM with DeepSpeed, and I've introduced an additional modality to the existing ones. Howe…
-
您好,
看论文中的效果很好,在InternVL2‑26B中进行尝试,不知道需要怎么修改?
感谢!
zhly0 updated
3 weeks ago
-
Hi,
I start the training with run_stage1.sh. One error message comes up when saving the training model at step 3000.
Any suggestions?
Thanks,
--Ruida
tensor(0., device='cuda:1', gra…
ruida updated
6 months ago
-
```shell
[rank2]: Traceback (most recent call last):
[rank2]: File "/mnt/Lumina-mGPT/lumina_mgpt/finetune_solver.py", line 113, in
[rank2]: solver = Solver(args)
[rank2]: File "/mnt/Lumin…
-
Fail To Load Models! Please Check if models exists at /sdcard/Download/model and restart
but i have copied model in that location. still getting same error message what would be the reason.
-
Thanks for your impressive work! The idea of integrating a text encoder before the large language model caught my attention. Could you share the specific benefits this approach provides? Also, were an…
-
in load_pretrained_model
model = CambrianLlamaForCausalLM.from_pretrained(
File "/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py", line 3531, in from_pretrained
) =…
-
File "/root/.local/lib/python3.12/site-packages/vlmeval/dataset/image_mcq.py", line 181, in evaluate
answer_map = {i: c for i, c in zip(meta['index'], meta['answer'])}
…