Hugging Face's logo
Activity Feed

AI & ML interests

None defined yet.
pranavupadhyaya52 
posted an update 17 days ago
view post
Post
536
Wiki-SmartBotLM Benchmarked. Here are the results:

Benchmark WikiSB | Gemma3_270M
BoolQ 37.8 | 61.4
ARC Challenge 24.2 | 28.2
HellaSwag 26.3 | 37.7
PIQA 53.5 | 66.2
Winogrande 50.2 | 52.3

Summary:
Closest benchmark: Winogrande (2.1-point gap)
Strong performance: ARC Challenge (4.0-point gap)
Moderate gap: HellaSwag and PIQA
Largest gap: BoolQ (Reading Comprehension)

It shows despite the training difference between Gemma3 270M and WikiSmartBotLM, It does come close to Gemma 3 in some benchmarks.

pranavupadhyaya52/Wiki-SmartBotLM-Instruct
  • 3 replies
·
pranavupadhyaya52 
posted an update 22 days ago
view post
Post
3623
After several weeks of experimenting, debugging, and iterating, I am excited to share WikiSmartBotLM.

WikiSmartBotLM is a compact decoder only language model built from the ground up as an educational and practical project. The goal was not simply to train another language model, but to create one that is easy to understand, modify, and experiment with while following many of the architectural ideas used in modern LLMs.

The model is built on a custom Transformer architecture featuring Rotary Positional Embeddings, RMSNorm, SwiGLU feed forward layers, grouped query attention, and an efficient autoregressive decoder optimized for local inference.

The repository includes the complete model implementation, configuration files, tokenizer integration, training pipeline, inference scripts, checkpoint conversion utilities, and examples that demonstrate how each component works together. Whether you want to understand the forward pass, train your own model, or build applications on top of WikiSmartBotLM, everything is designed to be approachable.
You can directly run the model via the models Huggingfaces space, which I've included in the post.

Model Repository: pranavupadhyaya52/Wiki-SmartBotLM-Instruct

I hope WikiSmartBotLM becomes a useful resource for anyone who enjoys learning by building. Feedback, issues, feature requests, and contributions are always welcome.
https://huggingface.co/spaces/pranavupadhyaya52/WikiSmartBot
  • 1 reply
·
pranavupadhyaya52 
updated a collection 3 months ago
pranavupadhyaya52 
updated a model 3 months ago
pranavupadhyaya52 
posted an update about 1 year ago
view post
Post
271
Here are my complete project notebooks hosted to my project titled "Multimodal Custom Finetuned LLM for Virtual Healthcare Consultations" at OSF - https://osf.io/ne2zh/
pranavupadhyaya52 
posted an update about 1 year ago
view post
Post
2286
Hello everyone. I've built a medical AI assistant application.

pranavupadhyaya52/MediWiki_Medical_Assistant

It is a multimodal chatbot and can accept text, radiology images, prescription and lab reports (currently it only accepts one image per chat.) and audio files (wav and MP3 extension files).

It is built on top of a finetuned Llama 3.2 11B vision instruct. It also uses a 41000 medically related question answer pair stored in the form of chromadb embedding for Retrieval Augmented Generation (RAG).

Please let me know your thoughts on my project and how I can improve it further. Thank you.
  • 5 replies
·