Skip to content

Model Selection of "nucleotide-transformer" #17

@Lvchangze

Description

@Lvchangze

Dear authors:
I tried to infer your model, Molly-4B, running "scripts/inference_nt_lora.sh".
But I found that the selection of the DNA-BERT model called "nucleotide-transformer" is unclear.
I downloaded model "InstaDeepAI/nucleotide-transformer-500m-human-ref" (https://huggingface.co/InstaDeepAI/nucleotide-transformer-500m-human-ref) on HuggingFace, but it seemed there was a dimension mismatch:

`
size mismatch for dna_rna_model.esm.encoder.layer.0.output.dense.weight: copying a param with shape torch.Size([1024, 4096]) from checkpoint, the shape in current model is torch.Size([1280, 5120]).

size mismatch for dna_rna_model.esm.encoder.layer.0.LayerNorm.weight: copying a param with shape torch.Size([1024]) from checkpoint, the shape in current model is torch.Size([1280]).

size mismatch for dna_rna_model.esm.encoder.layer.0.LayerNorm.bias: copying a param with shape torch.Size([1024]) from checkpoint, the shape in current model is torch.Size([1280]).
`

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions