lucidrains/RETRO-pytorch
Implementation of RETRO, Deepmind's Retrieval based Attention net, in Pytorch
PythonApache-2.0
Issues
- 1
how to deal with the problem ,
#42 opened by zhouchang123 - 0
Clarification about the code.
#45 opened by aneesh-shetye - 0
- 0
No embeddings found in folder .tmp/embeddings
#44 opened by Zahin112 - 1
I am revising the model to solve QA task..
#32 opened by 0525stone - 0
Convert embedded tokens to English
#41 opened by yerinNam - 6
How to give Prompt to trained RETRO Model?
#33 opened by shahmeer99 - 7
Retro-fitting a pretrained model
#26 opened by dean-sh - 2
- 5
- 3
- 8
TrainingWrapper does not support line breaks
#23 opened by 0x7o - 0
Question-Answer Dataset Format ?
#38 opened by szh-max - 0
Causal mask in Chunked Cross Attention
#35 opened by NathanaelBeau - 6
Scann vs faiss
#28 opened by afcruzs - 0
Error # could not open .tmp/.index/knn.index for reading: No such file or directory
#36 opened by siddgarg88 - 5
Why are there so many position embeddings?
#34 opened by jasperhyp - 4
- 4
- 0
Huggingface model
#31 opened by loretoparisi - 1
'NoneType' object is not callable
#29 opened by f-guoqiang - 0
Clarification on Architecture
#27 opened by bjascob - 1
- 2
Extra layer encoder_output_to_decoder_dim cause issue with distributed training
#20 opened by ncoop57 - 18
Error Reconstructing FAISS Index
#15 opened by ncoop57 - 0
Unable to do non-retrieval sampling
#19 opened by ncoop57 - 1
Could you provide a public dataset for RETRO?
#14 opened by liulhdarks - 3
rotary embedding question
#12 opened by jinmang2 - 2
- 13
Model training
#4 opened by malteos - 1
- 2
- 3
Autoregressivity
#3 opened by sdpmas - 5
`doc_ids_memmap` shape
#2 opened by josephcappadona