---

base_model: ProTrekHub/Protein_Encoder_35M 

library_name: peft

---



# Model Card for Model-targetP-ProTrek-35M
## Base model

[ProTrekHub/Protein_Encoder_35M](https://huggingface.co/ProTrekHub/Protein_Encoder_35M)

## Task description

Detecting Sequence Signals in Targeting Peptides.It is a protein level classification task specificly sub cellular location prediction.

## Label meaning

noTP 0, SP 1, mTP 2, cTP 3, luTP 4

## Citation

Please consider citing the following work if you find the model useful:

https://github.com/JJAlmagro/TargetP-2.0



## Task type
Protein-level Classification

## Model input type
AA Sequence

## LoRA config

- **r:** 8
- **lora_dropout:** 0.0
- **lora_alpha:** 16
- **target_modules:** ['intermediate.dense', 'query', 'output.dense', 'key', 'value']
- **modules_to_save:** ['classifier']

## Training config

- **optimizer:**
  - **class:** AdamW
  - **betas:** (0.9, 0.98)
  - **weight_decay:** 0.01
- **learning rate:** 0.0005
- **epoch:** 3
- **batch size:** 32
- **precision:** 16-mixed