RetroMAE

RetroMAE copied to clipboard

Some questions about the data part

1

Great work, I have some questions below: class RetroMAECollator(DataCollatorForWholeWordMask): max_seq_length: int = 512 encoder_mlm_probability: float = 0.15 decoder_mlm_probability: float = 0.15 def __call__(self, examples): input_ids_batch = [] attention_mask_batch = []...

chenchongthu

How to finetune with ANCE?

4

Hellow, Is that the code publiced in [code](https://github.com/staoxiao/RetroMAE/tree/master/examples/retriever/msmarco) is for finetuning the model by dpr? Where can i find the code for finetuning with ANCE? And there's another question that...

Victoriaheiheihei

How did you evaluate on NQ?

1

Hi, I see that you have also reported results on NQ dataset but there is no details for nq evaluation in the paper. Could you provide more instructions (codebase, data...

jordane95

Could you please provide the teacher score file for distillation fine-tuning?

2

Hi, Thanks for the great work! I was wondering if it is possible to provide the teacher score file you used to fine-tune the model on the MSMarco dataset? Thanks!

Gouzi3618

one GPU with zero GPU-util

1

Gavinthisisit

Any code for finetune by RetroMAE-v2?

1

I find that the finetune code is using RetroMAE V1, where i can't find something about ot-embedding in finetune code. Is it released? Dose the code released now use the...

kkkjyu