You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
ubuntu:/data/.cache/huggingface/hub/datasets--mPLUG--DocDownstream-1.0/snapshots/18fb3b5cb8b840dcde61be40ccf6247187af760b$ cat train.jsonl |head
{"image": ["./imgs/DUE_Benchmark/DocVQA/pngs/xnbl0037_1.png"], "messages": [{"role": "user", "content": "<|image|>what is the date mentioned in this letter?"}, {"role": "assistant", "content": "1/8/93"}], "task_name": "qa_sft", "dataset_name": "DocVQA"}
{"image": ["./imgs/DUE_Benchmark/DocVQA/pngs/xnbl0037_1.png"], "messages": [{"role": "user", "content": "<|image|>what is the date mentioned in this letter?"}, {"role": "assistant", "content": "1/8/93"}], "task_name": "qa_sft", "dataset_name": "DocVQA"}
{"image": ["./imgs/DUE_Benchmark/DocVQA/pngs/xnbl0037_1.png"], "messages": [{"role": "user", "content": "<|image|>what is the date mentioned in this letter?"}, {"role": "assistant", "content": "1/8/93"}], "task_name": "qa_sft", "dataset_name": "DocVQA"}
{"image": ["./imgs/DUE_Benchmark/DocVQA/pngs/xnbl0037_1.png"], "messages": [{"role": "user", "content": "<|image|>what is the contact person name mentioned in letter?"}, {"role": "assistant", "content": "p. carter"}], "task_name": "qa_sft", "dataset_name": "DocVQA"}
{"image": ["./imgs/DUE_Benchmark/DocVQA/pngs/xnbl0037_1.png"], "messages": [{"role": "user", "content": "<|image|>what is the contact person name mentioned in letter?"}, {"role": "assistant", "content": "p. carter"}], "task_name": "qa_sft", "dataset_name": "DocVQA"}
{"image": ["./imgs/DUE_Benchmark/DocVQA/pngs/xnbl0037_1.png"], "messages": [{"role": "user", "content": "<|image|>what is the contact person name mentioned in letter?"}, {"role": "assistant", "content": "p. carter"}], "task_name": "qa_sft", "dataset_name": "DocVQA"}
I found there are these duplicated entires in the train.jsonl, is this normal?
The text was updated successfully, but these errors were encountered:
I found there are these duplicated entires in the train.jsonl, is this normal?
The text was updated successfully, but these errors were encountered: