-
@HalemoGPA - as discussed, here is the goals of this task. We want to run the full eval via a simple command:
Evaluation:
- `python evaluate.py --api-key="my-key"` and run the full Vidore evalua…
-
### System Info
Using infinity 0.0.67
### Information
- [ ] Docker
- [ ] The CLI directly via pip
### Tasks
- [ ] An officially supported command
- [ ] My own modifications
### Reproduction
Whe…
-
Hi, I noticed that 10 padding tokens are added as a suffix to queries:
https://github.com/illuin-tech/colpali/blob/dd8e88894b1224ebdd1af4f6480474195015fdaa/colpali_engine/collators/visual_retriever_c…
-
Here are the results I am getting when training BiPali model `colpali/scripts/configs/pali/train_bipali_model.yaml` (which is basically ColPali but with avg embedding of patches instead of contextuali…
-
Your work is impressive. When I **run the RAG.index() function in ColPali_+_Qwen2_VL.ipynb in Windows environment**, an error message "RuntimeError: Input type (torch.cuda.FloatTensor) and weight type…
-
When I run the example given in the README.md, I encounter the following error:
`RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (CUDABFloat16Type) should be the same
`
Code…
-
I hosted the LLM and embedding model locally and I tried to embed the txt file using `rag.insert()`. Here is my embedding and LLM configuration:
```
rag = LightRAG(
working_dir=WORKING_DIR,
…
-
Hi,
Thanks for the implementation, currently I see that you have primarily 2 models : PaliGemma and Qwen, giving rise to ColPali and Colqwen, the question I had was how can we try other VLMs apart fr…
-
Hi Team,
I tried for a PR and had a error. If the project setup is intended to be gated, KIndly fix the two suggestions from the image.
![image](https://github.com/user-attachments/assets/07d1…
-
When i paste the copied text on the cell, the pasted results is wrong in the lastest jupyter Lab version
For example ,
After copy the text "colqwen2-v0.1",
I want to use the pasted text to re…