flytech/togetherchat-dev-7b

togetherchat-dev-7b

This model is a fine-tuned version of togethercomputer/LLaMA-2-7B-32K using 5000 examples and 3 datasets:

platypusdataset = loaddataset("garage-bAInd/Open-Platypus") codealpacadataset = loaddataset("sahil2801/CodeAlpaca-20k") evolcodealpacadataset = load_dataset("theblackcat102/evol-codealpaca-v1")

Model description

Step Training Loss --------------------- 60 1.293000 120 0.673600 180 0.633200 240 0.611600 300 0.633000 360 0.589500 480 0.587600 540 0.569000 600 0.548700 660 0.553100 720 0.531500 780 0.506400 840 0.512500

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• —learning_rate: 0.0002
• —trainbatchsize: 8
• —evalbatchsize: 8
• —seed: 42
• —gradientaccumulationsteps: 2
• —totaltrainbatch_size: 16
• —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• —lrschedulertype: constant
• —lrschedulerwarmup_ratio: 0.1
• —num_epochs: 3

Training results

Framework versions

• —Transformers 4.33.1
• —Pytorch 2.0.1+cu118
• —Datasets 2.14.5
• —Tokenizers 0.13.3