BrownianNotion
/

TinyLlama_v1.1_mix_wikitext_alpaca_2bit_BitDistiller_baseline

Eval Results (legacy)

Model card Files Files and versions

TODO: check the splits of each dataset

Loss curves:

Downloads last month: 8

Safetensors

Model size

1B params

Tensor type

BF16

·

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for BrownianNotion/TinyLlama_v1.1_mix_wikitext_alpaca_2bit_BitDistiller_baseline

Base model

TinyLlama/TinyLlama_v1.1

Finetuned

(54)

this model

Datasets used to train BrownianNotion/TinyLlama_v1.1_mix_wikitext_alpaca_2bit_BitDistiller_baseline

Evaluation results

Accuracy on ARC-Challenge
test set self-reported

0.215
Normalized Accuracy on ARC-Challenge
test set self-reported

0.246
Accuracy on HellaSwag
test set self-reported

0.324
Normalized Accuracy on HellaSwag
test set self-reported

0.373
Accuracy on PIQA
validation set self-reported

0.608
Normalized Accuracy on PIQA
validation set self-reported

0.607
Accuracy on Winogrande
test set self-reported

0.520
QA Average on QA-Avg
self-reported

0.417