# Few-shot fine-tuning

This model is a fine-tuned image segmentation model based on mukesh3444/window_detection_model on the nagarajuthirupathi/indoor_window_detection_swf dataset, focusing on indoor window detection tasks.

This is a text-to-image generation model based on the Flax framework, specifically designed for generating plush toy-style images.

An image segmentation model fine-tuned based on nvidia/mit-b0, performing excellently on the NICOPOI-9/morphpad_coord_hgo_512_4class dataset

Light-R1-32B-DS is a near-SOTA level 32B mathematical model, fine-tuned based on DeepSeek-R1-Distill-Qwen-32B, achieving high performance with only 3K SFT data.

XTTS V2 text-to-speech model fine-tuned on a custom Tunisian dataset

TinyTimeMixers (TTMs) are compact pretrained models for multivariate time-series forecasting open-sourced by IBM Research, starting from 1 million parameters, introducing the concept of 'miniature' pretrained models in time-series forecasting for the first time.

Urdu TTS model fine-tuned from microsoft/speecht5_tts with small training dataset (4,200 sentences). Commercial use requires retraining.

This is a monolingual Czech SpeechT5 base model, pre-trained on 120,000 hours of Czech audio and a 17.5 billion-word text corpus, designed as a starting point for Czech TTS fine-tuning.

This is a version of Microsoft's Florence-2 model fine-tuned for 1 day using the Docmatix dataset (5% of the data) with a learning rate of 1e-6

This is a multimodal model fine-tuned based on Microsoft's Kosmos-2 model, specifically designed for recognizing Pokemon names on Pokemon cards.

A model fine-tuned on a small dataset of 16,000 English translations of Korean patents based on Meta-Llama-3-8B-Instruct, for testing purposes only.

Chinese language model fine-tuned on the generator dataset based on google/gemma-1.1-7b-it

A video understanding model fine-tuned on an unknown dataset based on the MCG-NJU/videomae-base model, with an accuracy of 67.13%

This is a Spanish TTS model based on MMS, fine-tuned using the VITS architecture, requiring only 80-150 samples and 20 minutes of training time to generate Spanish speech with a Colombian accent.

This is a fine-tuned model based on the MMS Spanish version, built using the VITS architecture, trained with only 80 to 150 samples in approximately 20 minutes.

An audio classification model fine-tuned on the DonateACry dataset based on bookbot/distil-ast-audioset, designed for identifying infant cries

This is an ABAP programming language model fine-tuned based on Llama-2-7b-chat-hf, specifically designed for generating ABAP code

Image segmentation model fine-tuned on the Isaacks/ihc_slide_tissue dataset based on nvidia/mit-b0 model

A voice processing model fine-tuned based on microsoft/unispeech-sat-base, specializing in digit masking tasks, with outstanding performance on evaluation sets.

A vision model fine-tuned on an image folder dataset based on microsoft/swin-tiny-patch4-window7-224

A Vision Transformer model fine-tuned from google/vit-base-patch16-224-in21k, achieving 99.26% accuracy on the evaluation set

A model based on the Donut architecture, specifically fine-tuned for Latvian receipt data

This model is a fine-tuned version of donut-base on a Latvian receipt dataset, primarily used for receipt image processing tasks

A lightweight image classification model based on the DeiT-tiny architecture, achieving 94.8% accuracy after fine-tuning on a custom image dataset

This is a speech recognition model fine-tuned on the Catalan Common Voice 11.0 dataset based on OpenAI Whisper Medium.

A vision Transformer model based on the BEiT architecture, fine-tuned on FER2013CKPlus and SFEW datasets for facial expression recognition tasks.

A fine-tuned model based on Google Vision Transformer (ViT) architecture, suitable for image classification tasks

This is a tiny version of the image classification model based on the Swin Transformer architecture, fine-tuned on the EuroSAT dataset, suitable for remote sensing image classification tasks.

An image classification model fine-tuned for binary classification tasks based on facebook/convnext-small-224, achieving an F1 score of 0.9620 on the evaluation set

This is a computer vision model based on the ConvNeXt-Tiny architecture, fine-tuned on unsupervised data using pseudo-labels generated by Snorkel

A question-answering model based on the BART-base architecture, fine-tuned on the SQuAD dataset, suitable for reading comprehension tasks.

This model is a fine-tuned version based on facebook/convnext-tiny-224, primarily used for image classification tasks, demonstrating excellent performance on the evaluation set.

A vision classification model fine-tuned based on Google's Vision Transformer (ViT) foundation model, suitable for image classification tasks

A fine-tuned image classification model based on the Swin Transformer architecture, achieving 98.3% accuracy on the image folder dataset.

A fine-tuned image classification model based on Swin Transformer architecture, achieving 93.94% accuracy on the EuroSAT dataset

Fine-tuned image classification model based on Swin Transformer architecture, achieving 98% accuracy on the EuroSAT dataset

A fine-tuned image classification model based on Swin Transformer Tiny architecture, achieving 97.26% accuracy on the EuroSAT dataset

This model is a fine-tuned BEiT base version on the CIFAR-10 dataset, focusing on image classification tasks, achieving 99.18% accuracy on the evaluation set.

This is a fine-tuned model based on the Swin Transformer Tiny architecture, specifically designed for image classification tasks, achieving an accuracy of 97.59% on the evaluation set.

A fine-tuned image classification model based on the Swin Transformer architecture, achieving 97.44% accuracy on the image folder dataset