twiggly/astrobridge-model-v4_gemma

AstroBridge Captioner

n-modality (image + spectra) astronomy captioner. LoRA adapter + fusion stack trained on top of a frozen google/gemma-4-12B-it. The base model itself is NOT included here — load it fresh from google/gemma-4-12B-it and apply this adapter on top.

How to load

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
import torch

base = AutoModelForCausalLM.from_pretrained(
    "google/gemma-4-12B-it", dtype=torch.bfloat16, trust_remote_code=True
)
llm = PeftModel.from_pretrained(base, "twiggly/astrobridge-model-v4_gemma")
tokenizer = AutoTokenizer.from_pretrained("twiggly/astrobridge-model-v4_gemma")

# middle.pt (fusion stack: projectors/modality_identity/qformer/adapter) needs the captioner
# package's FusionStack class to reload — see captioner/model/captioner.py and
# captioner/train/stage1.py's run_stage1 for how it's constructed and wired to the LLM.

Training info

• config_hash: b1a6d42674db7337
• quantization: None
• git_sha: 84d698621d60cb62860bd386972d9282a0a62ab3
• tier_histogram: {"single": 4091, "joint": 1215}

Eval (groundedness gate)

Not yet run — see `make eval`.