mamba_pico_large_x_dt_out_proj

Model Card for Model ID

Adapter for mlsquare/pico_seshu_test using LoRA on "model.layers.3.dt_proj", "model.layers.3.x_proj", "model.layers.3.out_proj". Standard use of PEFT on Mamba-hf model

Model Details

Model Description

• Developed by: MLsquare
• Model type: Next Character Generation
• Language(s) (NLP): All languages in ai4bharat/samanantar dataset
• License: MIT

Model Details

Model Description

• Developed by: MLsquare
• Model type: Next Character Generation
• Language(s) (NLP): All languages in ai4bharat/samanantar dataset
• License: MIT

Model Sources [optional]

• Repository: https://github.com/LegallyCoder/mamba-hf
• Paper: https://arxiv.org/abs/2312.00752

Uses

Refer to the github repository for more information

Direct Use

Refer to the github repository for more information

How to Get Started with the Model

Refer to the github repository: https://github.com/mlsquare/fedem

Training Details

Training Data

Individual target and source sentences from the AI4Bharat Samanantar dataset. All 11 language sentences and their translations have been stacked and used for next character generation task.

Training Procedure

Trained on the next character generation task using cross-entropy loss.

Preprocessing [optional]

converted to raw UTF8 characters before training by using ByT5-large tokenizer

Training Hyperparameters

• Training regime: output_dir="mamba", per_device_train_batch_size=1, per_device_eval_batch_size=1, num_train_epochs=4, weight_decay=0.1, lr_scheduler_type="cosine", learning_rate=5e-4, fp16=False,

Evaluation

A simple cross-entropy loss has been used to test the pipeline and working of the model.

Model Card Contact

MLsquare