princeton-nlp / AutoCompressors

[EMNLP 2023] Adapting Language Models to Compress Long Contexts

https://arxiv.org/abs/2305.14788

princeton-nlp/AutoCompressors Issues

Inquire on data of Table 1
Updated 2 months ago1
Reduce the number of summary vectors
Closed 2 months ago2
Question about the data preprocessing
Closed 2 months ago1
Install instructions are not clear
Closed 2 months ago2
Dimension of last_hidden_state size
Closed 2 months ago2
question about `position_ids`
Closed 2 months ago2
Held-out perplexity question
Closed 3 months ago3
Finetuning an autocompressor model
Closed 3 months ago4
AttributeError: 'SubstepTrainer' object has no attribute 'do_grad_scaling'
Closed 3 months ago3
RuntimeError: FlashAttention only support fp16 and bf16 data type
Closed 3 months ago3
Summary Vector Failures and Incomplete Answers with Numerical Contexts
Closed 5 months ago4
BUG REPORT
Closed 8 months ago1
CUDA out of memory.
Closed 10 months ago3
Inquiry for the release date of the pre-trained model
Closed 10 months ago1
Question on the preprocessed data
Closed a year ago3
Follow-up on Code Release Timeline
Closed a year ago4
Timeline for Release of Code and Pre-Trained Models
Closed a year ago2