ThakrePranjal/pharma-tinyllama-merged

Pharma TinyLlama — Merged Model

This is the fully merged standalone model: TinyLlama/TinyLlama-1.1B-Chat-v1.0 with the ThakrePranjal/pharma-tinyllama-domain-lora LoRA adapter merged in. Use this if you want a single self-contained model without needing PEFT.

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model_repo = "ThakrePranjal/pharma-tinyllama-merged"

model = AutoModelForCausalLM.from_pretrained(model_repo)
tokenizer = AutoTokenizer.from_pretrained(model_repo)

Relationship to other repos

• LoRA-only adapter: ThakrePranjal/pharma-tinyllama-domain-lora
• Training dataset: ThakrePranjal/pharma-domain-corpus

Intended use & limitations

Same as the LoRA adapter repo — domain-adaptive continued pretraining only, not instruction-tuned, not validated for clinical/production use.