ProtGPT2-Oxido / README.md
zypchn's picture
Update README.md
605e0c1 verified
---
license: apache-2.0
base_model:
- nferruz/ProtGPT2
tags:
- axolotl
- lora
- transformers
- gpt2
- protein_design
---
# Model
This is a fine-tuned version of [ProtGPT2](https://huggingface.co/nferruz/ProtGPT2) : a language model that speaks the protein language and can be used for de novo protein design and engineering.
<br/>
# Dataset
Protein data set retrieved from Research Collaboratory for Structural Bioinformatics (RCSB) Protein Data Bank (PDB). \
Only the OXIDOREDUCTASE enzymes were used. \
You can find the JSON formatted data @ [oxidos.json](https://github.com/zypchn/pLM/blob/main/data/oxidos.json)
<br/>
# How to Use?
```
>>> from transformers import pipeline
>>> pipe = pipeline("text-generation", model="zypchn/ProtGPT2-Oxido")
>>> sequences = pipe("", max_length=100, do_sample=True, top_k=950, repetition_penalty=1.2, num_return_sequences=5, eos_token_id=0)
# input field has left blank for diversity
```
```json
[
{"generated_text": "SNANQAPQPQTPTRATDAKKGSYGHPADRVGMEDNKYQVGVFYYDGPNPSYAEWNRDTQFWVETAKTAEKGKFDSIFFADTLGIYDSFKGSFEANLRHGAQFPVNDPLVAISAIAGATTKLGLVATASTTYSEPFHIARRFASLDHLSNGRAGWNIVTSYLDSAARNFGRTEQMEHDERYAIAEEYIDVVYKLWEGSWEDNAVIKDKETGLFTDPAKVHQINHEGEHFRVAGPLNIPRSPQGHPVIFQAGTSERGRDFAARHAEAVFTAQLDLEAGREFYEDIKSRAAKLGRDPDDVKILPGISVFVGKTREEAERKFRELQSLIDEEGALTRFSSYTGTDLSTYDPDGPLPELAGIDPTTPIAKLEGLLGKSKMTVREIALKQGGVSLREYQPFVGATAGSALVGGTPEQIADFMQDWFIEGTVDGFNIMPPYLPDGLEDFVDHVVPELQRRGLFRTEYEGTTLREHLGLAKPLEHHHHHH"},
{"generated_text": "MGSSHHHHHHSSGLVPRGSHMASMTGGQQMGRGSMGPCLICRSTSLKCVFCVRDPNGYKKCSKCDAFFCSRECQTEHWQRHHKFECPAAVAQPQIPPLPKPQQKQLTAAELGMFMEVRNQFALLKTNLERLDYEIFILERNVKLANTVTPPTNRTYFQSTMRYAPNPLRPNMTDAMRQQYLDKNKSSAALEHDLKELIKFKCYLLNDEYVEKEREENPFIWEYFLNKEWRKRNVWGNK\n"},
{"generated_text": "MGHHHHHHSSGLVPRGSHMTVEQAKKLRAEAEAQAQIQDKAKAIAQTHGKVEVMVDGKHRVVDLDATTRRQLTDGELQAIVVAAQEAAAKQLKAQRQALLEQHQDAELRKLALEGEIV\nAVITGAAQGIGRAIALRLAKDGFRVAVADIDLAAAEAVAAEIEAKGGKALVIEGDVSREEDVKRLVRKAIDQFGRLDYAVNNAGIQGPLAPTEELPLALWNKVIDVNLTGVFLCMKYEIAQMVKQGRGGAIVNTASVAGLSGQPGMVAYCASKHGVVGLTKTVAIEYAKHGIRINAVAPGFIDTPMVQKLPEEKRARIAAAIPMRRLGQPDEIAAVVAFLLSDDASFITGQCIAVDGGFTAGLLA"},
{"generated_text": "MAASKAADSLAEGAAKLEHHHHHH"},
{"generated_text": "GSKPQPGVQVEGAKCQVLQAVYDFTVQSASELSFKAGDVICVTGQYDPTLGWWLAEERRTGKSGLVPENYVELLSTGPAQHHHHHH"}
]
```