metadata

tags:
  - generated_from_trainer
datasets:
  - pile-instruct/
metrics:
  - accuracy
model-index:
  - name: layer_4,5,6,7,8
    results:
      - task:
          type: text-generation
          name: Causal Language Modeling
        dataset:
          name: pile-instruct/
          type: pile-instruct/
          split: None
        metrics:
          - type: accuracy
            value: 0.20994595912408442
            name: Accuracy

layer_4,5,6,7,8

This model is a fine-tuned version of P1ayer-1/pythia-deduped-1b-chat-base on the pile-instruct/ dataset. It achieves the following results on the evaluation set:

Loss: 6.9437
Accuracy: 0.2099

Model description

More information needed

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: 0.0001
train_batch_size: 1
eval_batch_size: 8
seed: 42
distributed_type: multi-GPU
num_devices: 8
gradient_accumulation_steps: 8
total_train_batch_size: 64
total_eval_batch_size: 64
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
7.6017	0.02	200	7.5928	0.1605
7.1871	0.03	400	7.2690	0.1847
7.0356	0.05	600	7.0897	0.1980
6.93	0.07	800	6.9870	0.2064
6.9089	0.08	1000	6.9437	0.2099

Framework versions

Transformers 4.28.1
Pytorch 2.0.0+cu117
Datasets 2.11.0
Tokenizers 0.13.3

Wandb Report

https://wandb.ai/ontocord/pythia-1b-deduped-layer-test-min-pile-instruct/runs/6hvfd11h