Text Generation
Transformers
Safetensors
biomedical
genomics
variant-interpretation
lora
clinical-genomics
bioinformatics
research
conversational
Instructions to use Babajaan/KAU-BioMedLLM with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Babajaan/KAU-BioMedLLM with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="Babajaan/KAU-BioMedLLM") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("Babajaan/KAU-BioMedLLM", dtype="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use Babajaan/KAU-BioMedLLM with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Babajaan/KAU-BioMedLLM" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Babajaan/KAU-BioMedLLM", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/Babajaan/KAU-BioMedLLM
- SGLang
How to use Babajaan/KAU-BioMedLLM with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "Babajaan/KAU-BioMedLLM" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Babajaan/KAU-BioMedLLM", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "Babajaan/KAU-BioMedLLM" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Babajaan/KAU-BioMedLLM", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use Babajaan/KAU-BioMedLLM with Docker Model Runner:
docker model run hf.co/Babajaan/KAU-BioMedLLM
Resolve resource versions (UniProt 2026_02, AlphaFold v6, Ensembl 116) + Llama-3.1 license compliance ('Built with Llama')
Browse files- README.md +10 -2
- reports/DATA_PROVENANCE.md +6 -3
README.md
CHANGED
|
@@ -269,7 +269,7 @@ Full guarded report examples and summaries are included in the repository under
|
|
| 269 |
- Missense prediction remains below specialist predictors such as REVEL and AlphaMissense.
|
| 270 |
- Broad score model performance is partly driven by easier consequence classes.
|
| 271 |
- Structure evidence is only used when transcript/protein/residue mapping can be reconciled.
|
| 272 |
-
-
|
| 273 |
- The model has not been prospectively validated in clinical settings.
|
| 274 |
- Generated text can still contain errors and must be reviewed by domain experts.
|
| 275 |
- The uploaded artifact is an adapter, not a standalone full model.
|
|
@@ -384,4 +384,12 @@ Suggested acknowledgement:
|
|
| 384 |
|
| 385 |
## License and Base Model Notes
|
| 386 |
|
| 387 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 269 |
- Missense prediction remains below specialist predictors such as REVEL and AlphaMissense.
|
| 270 |
- Broad score model performance is partly driven by easier consequence classes.
|
| 271 |
- Structure evidence is only used when transcript/protein/residue mapping can be reconciled.
|
| 272 |
+
- External resource versions (resolved 2026-06-19): UniProt release 2026_02, AlphaFold DB v6, Ensembl release 116 (REST API 15.12), all accessed 2026-06-15; exact releases were reconstructed from access dates rather than captured at runtime.
|
| 273 |
- The model has not been prospectively validated in clinical settings.
|
| 274 |
- Generated text can still contain errors and must be reviewed by domain experts.
|
| 275 |
- The uploaded artifact is an adapter, not a standalone full model.
|
|
|
|
| 384 |
|
| 385 |
## License and Base Model Notes
|
| 386 |
|
| 387 |
+
**Built with Llama.**
|
| 388 |
+
|
| 389 |
+
This repository contains LoRA adapters and project documentation. It does not redistribute full Llama or Qwen base model weights.
|
| 390 |
+
|
| 391 |
+
The root adapter is a derivative of `meta-llama/Llama-3.1-8B-Instruct` and its use is governed by the [Meta Llama 3.1 Community License](https://www.llama.com/llama3_1/license/) and the [Acceptable Use Policy](https://www.llama.com/llama3_1/use-policy/). Per that license, derivative models carry the **"Built with Llama"** notice above; note also that the license asks derivative model names to begin with "Llama" — this adapter is published as **KAU-BioMedLLM** for research continuity, and adopters who redistribute should review that naming clause for their own use.
|
| 392 |
+
|
| 393 |
+
The `qwen_v0_1/` adapter is a derivative of `Qwen/Qwen2.5-1.5B-Instruct` and its use is subject to the [Qwen license terms](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct/blob/main/LICENSE).
|
| 394 |
+
|
| 395 |
+
Users are responsible for verifying license compatibility for their own use case. This research release makes no warranty of fitness for any purpose.
|
reports/DATA_PROVENANCE.md
CHANGED
|
@@ -51,7 +51,8 @@ This file records sources actually present or used in the HPC project. Unknown v
|
|
| 51 |
- Evidence file: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/knowledge/kaubiomed_gene_evidence.jsonl`
|
| 52 |
- Exists: `True`
|
| 53 |
- Size bytes: `253848`
|
| 54 |
-
- Access date
|
|
|
|
| 55 |
- Scope: 20-gene evidence panel, not genome-wide.
|
| 56 |
|
| 57 |
## AlphaFold DB
|
|
@@ -59,13 +60,15 @@ This file records sources actually present or used in the HPC project. Unknown v
|
|
| 59 |
- Manifest: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/knowledge/alphafold/alphafold_manifest.jsonl`
|
| 60 |
- Exists: `True`
|
| 61 |
- Size bytes: `12754`
|
| 62 |
-
- Access date
|
|
|
|
| 63 |
|
| 64 |
## Ensembl VEP REST
|
| 65 |
|
| 66 |
- Mapper script: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/pipeline/vep_residue_mapper.py`
|
| 67 |
- Exists: `True`
|
| 68 |
-
-
|
|
|
|
| 69 |
- Mapping policy: MANE-preferred, HGVS protein required, UniProt reconciliation required, abstain on mismatch.
|
| 70 |
|
| 71 |
## Base LLM Snapshots
|
|
|
|
| 51 |
- Evidence file: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/knowledge/kaubiomed_gene_evidence.jsonl`
|
| 52 |
- Exists: `True`
|
| 53 |
- Size bytes: `253848`
|
| 54 |
+
- Access date: 2026-06-15 (HPC file mtime; queried via rest.uniprot.org/uniprotkb).
|
| 55 |
+
- UniProt release: 2026_02 (released 10-Jun-2026), the release live on the access date. Resolved 2026-06-19 from UniProt relnotes; release was not captured at runtime.
|
| 56 |
- Scope: 20-gene evidence panel, not genome-wide.
|
| 57 |
|
| 58 |
## AlphaFold DB
|
|
|
|
| 60 |
- Manifest: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/knowledge/alphafold/alphafold_manifest.jsonl`
|
| 61 |
- Exists: `True`
|
| 62 |
- Size bytes: `12754`
|
| 63 |
+
- Access date: 2026-06-15 (HPC file mtime).
|
| 64 |
+
- AlphaFold DB model version: v6 (from structure URLs, e.g. AF-P32245-F1-model_v6.cif). Note: an earlier evidence-builder pass fetched model_v4; final structures were refreshed to v6, which the manifest reflects.
|
| 65 |
|
| 66 |
## Ensembl VEP REST
|
| 67 |
|
| 68 |
- Mapper script: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/pipeline/vep_residue_mapper.py`
|
| 69 |
- Exists: `True`
|
| 70 |
+
- Access date: 2026-06-15 (residue-mapping output mtime; rest.ensembl.org/vep/human/region, assembly GRCh38).
|
| 71 |
+
- Ensembl release: 116 (REST API 15.12), the current live release. Resolved 2026-06-19 from rest.ensembl.org/info; release was not captured at runtime.
|
| 72 |
- Mapping policy: MANE-preferred, HGVS protein required, UniProt reconciliation required, abstain on mismatch.
|
| 73 |
|
| 74 |
## Base LLM Snapshots
|