YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
GeneLM – Gene Expression Language Model for Spatial Transcriptomics
Epoch 2 checkpoint trained on mouse brain scRNA-seq (C57BL/6J Allen Brain Atlas sections).
Architecture
- Gene encoder: CellFM (custom Vision/Transformer backbone,
enc_dims=1536) - LLM: meta-llama/Llama-3.1-8B (LoRA fine-tuned then merged)
- Projection:
Linear(1536→4096) → GELU → Linear(4096→4096) - Task: given cell gene-expression + metadata instruction, predict spatial (x, y) coordinates
Files
| File | Description |
|---|---|
llama_merged/ |
Merged LLaMA-3.1-8B + LoRA weights (safetensors) + tokenizer |
custom_modules/custom_modules.safetensors |
CellFM encoder, LayerNorm, projection layer |
custom_modules/genelm_config.json |
Architecture metadata |
Usage
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from safetensors.torch import load_file
# Load merged LLaMA
llama = AutoModelForCausalLM.from_pretrained("yue/genelm-spatial/llama_merged", torch_dtype=torch.bfloat16)
tokenizer = AutoTokenizer.from_pretrained("yue/genelm-spatial/llama_merged")
# Load custom modules
custom = load_file("yue/genelm-spatial/custom_modules/custom_modules.safetensors")
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support