xtremedistil-l6-h256-zeroshot-v1.1-all-33

This model was fine-tuned using the same pipeline as described in the model card for MoritzLaurer/deberta-v3-large-zeroshot-v1.1-all-33 and in this paper.

The foundation model is microsoft/xtremedistil-l6-h256-uncased. The model only has 22 million backbone parameters and 30 million vocabulary parameters. The backbone parameters are the main parameters active during inference, providing a significant speedup over larger models. The model is 25 MB small.

This model was trained to provide a very small and highly efficient zeroshot option, especially for edge devices or in-browser use-cases with transformers.js.

Usage and other details

For usage instructions and other details refer to this model card MoritzLaurer/deberta-v3-large-zeroshot-v1.1-all-33 and this paper.

Metrics:

I didn't not do zeroshot evaluation for this model to save time and compute. The table below shows standard accuracy for all datasets the model was trained on (note that the NLI datasets are binary).

General takeaway: the model is much more efficient than its larger sisters, but it performs less well.

Datasets	mnli_m	mnli_mm	fevernli	anli_r1	anli_r2	anli_r3	wanli	lingnli	wellformedquery	rottentomatoes	amazonpolarity	imdb	yelpreviews	hatexplain	massive	banking77	emotiondair	emocontext	empathetic	agnews	yahootopics	biasframes_sex	biasframes_offensive	biasframes_intent	financialphrasebank	appreviews	hateoffensive	trueteacher	spam	wikitoxic_toxicaggregated	wikitoxic_obscene	wikitoxic_identityhate	wikitoxic_threat	wikitoxic_insult	manifesto	capsotu
Accuracy	0.894	0.895	0.854	0.629	0.582	0.618	0.772	0.826	0.684	0.794	0.91	0.879	0.935	0.676	0.651	0.521	0.654	0.707	0.369	0.858	0.649	0.876	0.836	0.839	0.849	0.892	0.894	0.525	0.976	0.88	0.901	0.874	0.903	0.886	0.433	0.619
Inference text/sec (A10G GPU, batch=128)	4117.0	4093.0	1935.0	2984.0	3094.0	2683.0	5788.0	4926.0	9701.0	6359.0	1843.0	692.0	756.0	5561.0	10172.0	9070.0	7511.0	7480.0	2256.0	3942.0	1020.0	4362.0	4034.0	4185.0	5449.0	2606.0	6343.0	931.0	5550.0	864.0	839.0	837.0	832.0	857.0	4418.0	4845.0