kuleshov-group/caduceus

Issues

• ImportError with newer mamba_ssm

  #51 opened 2 months ago by davidaknowles
  1

• Pretraining with HF trainer

  #50 opened 2 months ago by loieeyet
  3

• loss nan in Genomic Benchmark

  #34 opened 6 months ago by multydoffer
  10

• Error when reproducing the pre-training

  #16 opened 2 months ago by richardsun-voyager
  12

• MCC calculations

  #48 opened 2 months ago by leannmlindsey
  2

• Inference?

  #49 opened 2 months ago by leannmlindsey
  1

• Confusion with batches shown in progress bar

  #47 opened 2 months ago by cbirchsy
  2

• Conda environment installation issues

  #44 opened 4 months ago by mathinic
  2

• About pretrain GPUS

  #42 opened 5 months ago by wyhsleep
  3

• Conda environment doesn't work out of the box

  #43 opened 5 months ago by edenlum
  0

• Difficulty Reproducing Downstream Task Finetune Results with HF Trainer

  #37 opened 6 months ago by yangzhao1230
  10

• error when using hf model in genomic_benchmark pretrain

  #40 opened 6 months ago by lyksj200
  4

• High Storage Requirements for Running vep_embeddings.py in EQTL VEP Analysis

  #28 opened 6 months ago by yangzhao1230
  11

• The TF binding or enhancer activation is orientation independent

  #41 opened 6 months ago by freedom-cognit
  3

• Does Caduceus Support Variable-Length Inputs in Batches?

  #39 opened 6 months ago by yangzhao1230
  1

• Strange results using mcc on genomic benchmark?

  #38 opened 6 months ago by leannmlindsey
  2

• Customize dimension and layers

  #35 opened 6 months ago by zyj1729
  5

• Error using seq classification with d_model 256

  #27 opened 6 months ago by leannmlindsey
  5

• "omegaconf.errors.InterpolationKeyError: Interpolation key 'metric' not found" in NucelotideTransformer Benchmark

  #36 opened 6 months ago by yangzhao1230
  1

• reproducing the pre-training

  #26 opened 6 months ago by GengGengJiuXi
  8

• Package requirements issue

  #30 opened 6 months ago by amelie-iska
  3

• runtime terminal behaves

  #33 opened 6 months ago by yongrenr
  3

• loss nan in Genomic Benchmark

  #32 opened 6 months ago by multydoffer
  2

• num_workers=1 warning

  #29 opened 6 months ago by leannmlindsey
  2

• "Unexpected segmentation fault encountered in worker" in multi-gpu training

  #31 opened 6 months ago by yangzhao1230
  0

• Fine Tuning on custom datasets

  #25 opened 7 months ago by leannmlindsey
  1

• import quesition

  #23 opened 7 months ago by GengGengJiuXi
  0

• question about training lengths

  #24 opened 7 months ago by leannmlindsey
  1

• val/loss=nan.0 when training with ddp (distributed data parallel)

  #22 opened 7 months ago by leannmlindsey
  7

• How to load pre-trained model for fine-tuning?

  #14 opened 7 months ago by zhan8855
  6

• Unexpected data loader error during pre-training

  #20 opened 7 months ago by zhan8855
  3

• Unexpected low test accuracy in reproducing dummy_mouse_enhancers_ensembl experiments

  #19 opened 7 months ago by zhan8855
  4

• pretraining parameters for caduceus-ph and caduceus w/o equivariance

  #21 opened 7 months ago by leannmlindsey
  2

• Inconsistency between pre-training slurm scripts and the training details in the paper

  #15 opened 7 months ago by zhan8855
  2

• Slurm scripts for downstream tasks

  #18 opened 7 months ago by zhan8855
  2

• LLVM ERROR: Cannot select: intrinsic %llvm.nvvm.shfl.sync.bfly.i32

  #2 opened 8 months ago by Jellymoon
  9

• The size of the vocab is inconsistent with the output dimension of the lm_head

  #17 opened 7 months ago by jaminzzz
  1

• CPU

  #13 opened 8 months ago by yongrenr
  3

• Question about pre-training

  #11 opened 8 months ago by leannmlindsey
  5

• Small typo in wrapper scripts

  #12 opened 8 months ago by leannmlindsey
  1

• Issues with genomic_benchmark Experiment code in Caduceus

  #9 opened 8 months ago by xwx1999
  3

• Something about specifying the GPU

  #7 opened 8 months ago by yongrenr
  5

• 1.9million parameters, 470K parameters Caduceus model query

  #8 opened 8 months ago by MarcAmil30
  1

• What does the length extension mean for pre-training dataset?

  #6 opened 8 months ago by KatarinaYuan
  1

• Unable to run the model on input sequence with length of 131072

  #4 opened 8 months ago by qiaoqiaoLF
  4

• There seems a bug of related to d_model in class "CaduceusForSequenceClassification" for the RCPS case

  #5 opened 8 months ago by qiaoqiaoLF
  4

• Questions about experimental code

  #3 opened 8 months ago by yongrenr
  5