jploski/falcon40b-mini-shakespeare

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

falcon40b-mini-shakespeare

This model was trained from scratch on "tinyshakespeare" text file.

Model description

The configuration and code is adapted from tiiuae/falcon-40b, with configuration parameters changed to make it a very tiny model.

• —License: Apache 2.0.

Intended uses & limitations

Intended just to aid debugging efforts of a GGML port of Falcon.

Training and evaluation data

https://raw.githubusercontent.com/karpathy/char-rnn/master/data/tinyshakespeare/input.txt

Training procedure

Just used the single tinyshakespeare text file as both the training and validation set (split up into paragraphs). See:

https://colab.research.google.com/drive/1dl7Jko78CX1yEYZJWcRW9b7X-0Bst8?usp=sharing

Training hyperparameters

The following hyperparameters were used during training:

• —learning_rate: 0.0005
• —trainbatchsize: 32
• —evalbatchsize: 32
• —seed: 42
• —gradientaccumulationsteps: 8
• —totaltrainbatch_size: 256
• —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• —lrschedulertype: cosine
• —lrschedulerwarmup_steps: 10
• —num_epochs: 10

Framework versions

• —Transformers 4.28.0
• —Pytorch 2.0.1+cu118
• —Datasets 2.12.0
• —Tokenizers 0.13.3