Views
No views yet
| Parameter | Details |
|---|---|
| Base Model | Meta LLaMa 3.1 8B |
| Fine-Tuning Framework | LoRA |
| Dataset Size | 67,299 high-quality Q&A pairs |
| Context Length | 4,096 tokens |
| Training Steps | 100,000 |
| Model Size | 8 billion parameters |
| Source | Description |
|---|---|
| PubMed | Extracted insights from open-access medical research. |
| Clinical Guidelines | Data sourced from WHO, CDC, and specialty-specific guidelines. |
| EHR-Simulated Data | Synthetic datasets modeled on real-world patient records for anamnesis workflows. |
pip install llama-cpp-python --prefer-binary --extra-index-url=https://jllllll.github.io/llama-cpp-python-cuBLAS-wheels/AVX2/cu1181from llama_cpp import Llama
2
3llm = Llama(
4 model_path="recurv_medical_llama.gguf",
5 n_ctx=2048, # Context window
6 n_threads=4 # Number of CPU threads to use
7)1prompt = "What is Paracetamol?"
2output = llm(
3 prompt,
4 max_tokens=256, # Maximum number of tokens to generate
5 temperature=0.5, # Controls randomness (0.0 = deterministic, 1.0 = creative)
6 top_p=0.95, # Nucleus sampling parameter
7 stop=["###"], # Optional stop words
8 echo=True # Include prompt in the output
9)
10
11# Print the generated text
12print(output['choices'][0]['text'])