Hugging Face's logo
Anurag's picture
Building on HF
11 28 100

Anurag

edwixx
ModelsLab
Patil's profile picture 21world's profile picture theshresthshukla's profile picture
Ā·
https://anuragkanade.com/

AI & ML interests

Machine Learning, and Speech

Recent Activity

new activity about 15 hours ago
huggingface/InferenceSupport:edwixx/whisper-large-hebrew-finetune
reacted to sagar007's post with šŸ¤ about 16 hours ago
šŸš€ I built a Multimodal Vision-Language Model from using Gemma-270M + CLIP! Just finished training my multimodal model on the full LLaVA-Instruct-150K dataset (157K samples) and wanted to share the results! šŸ”§ What I Built: A vision-language model that can understand images and answer questions about them, combining: - Google Gemma-3-270M (language) - OpenAI CLIP ViT-Large/14 (vision) - LoRA fine-tuning for efficiency šŸ“Š Training Stats: - 157,712 training samples (full LLaVA dataset) - 3 epochs on A100 40GB - ~9 hours training time - Final loss: 1.333 training / 1.430 validation - Only 18.6M trainable params (3.4% of 539M total) šŸ“ˆ https://huggingface.co/sagar007/multigemma Benchmark Results: - VQA Accuracy: 53.8% - Works great for: animal detection, room identification, scene understanding šŸ”— **Try it yourself:** - šŸ¤— Model: https://huggingface.co/sagar007/multigemma - šŸŽ® Demo: https://huggingface.co/spaces/sagar007/Multimodal-Gemma - šŸ’» GitHub: https://github.com/sagar431/multimodal-gemma-270m Built with PyTorch Lightning + MLflow for experiment tracking. Full MLOps pipeline with CI/CD! Would love to hear your feedback! šŸ™ #multimodal #gemma #clip #llava #vision-language #pytorch
reacted to sagar007's post with šŸ”„ about 16 hours ago
šŸš€ I built a Multimodal Vision-Language Model from using Gemma-270M + CLIP! Just finished training my multimodal model on the full LLaVA-Instruct-150K dataset (157K samples) and wanted to share the results! šŸ”§ What I Built: A vision-language model that can understand images and answer questions about them, combining: - Google Gemma-3-270M (language) - OpenAI CLIP ViT-Large/14 (vision) - LoRA fine-tuning for efficiency šŸ“Š Training Stats: - 157,712 training samples (full LLaVA dataset) - 3 epochs on A100 40GB - ~9 hours training time - Final loss: 1.333 training / 1.430 validation - Only 18.6M trainable params (3.4% of 539M total) šŸ“ˆ https://huggingface.co/sagar007/multigemma Benchmark Results: - VQA Accuracy: 53.8% - Works great for: animal detection, room identification, scene understanding šŸ”— **Try it yourself:** - šŸ¤— Model: https://huggingface.co/sagar007/multigemma - šŸŽ® Demo: https://huggingface.co/spaces/sagar007/Multimodal-Gemma - šŸ’» GitHub: https://github.com/sagar431/multimodal-gemma-270m Built with PyTorch Lightning + MLflow for experiment tracking. Full MLOps pipeline with CI/CD! Would love to hear your feedback! šŸ™ #multimodal #gemma #clip #llava #vision-language #pytorch
View all activity

Organizations

Stable Diffusion API's profile picture Tune a video concepts library's profile picture ModelsLab's profile picture StyleTTS 2 Community's profile picture Chinese LLMs on Hugging Face's profile picture