84 lines
1.9 KiB
Python
84 lines
1.9 KiB
Python
|
import torch
|
||
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
||
|
|
|
||
|
|
# INSTRUCTIONS: Replace the JSON below with your material's properties
|
||
|
|
# Common data sources: materialsproject.org, DFT calculations, experimental databases
|
||
|
|
|
||
|
|
JSON_INPUT = """
|
||
|
|
{
|
||
|
|
"material_id": "mp-8062",
|
||
|
|
"formula": "SiC",
|
||
|
|
"elements": [
|
||
|
|
"Si",
|
||
|
|
"C"
|
||
|
|
],
|
||
|
|
"spacegroup": "P63mc",
|
||
|
|
"band_gap": 3.26,
|
||
|
|
"formation_energy_per_atom": -0.73,
|
||
|
|
"density": 3.21,
|
||
|
|
"volume": 41.2,
|
||
|
|
"nsites": 8,
|
||
|
|
"is_stable": true,
|
||
|
|
"elastic_modulus": 448,
|
||
|
|
"bulk_modulus": 220,
|
||
|
|
"thermal_expansion": 4.2e-06,
|
||
|
|
"electron_affinity": 4.0,
|
||
|
|
"ionization_energy": 6.7,
|
||
|
|
"crystal_system": "Hexagonal",
|
||
|
|
"magnetic_property": "Non-magnetic",
|
||
|
|
"thermal_conductivity": 490,
|
||
|
|
"specific_heat": 0.69,
|
||
|
|
"is_superconductor": false,
|
||
|
|
"band_gap_type": "Indirect"
|
||
|
|
}
|
||
|
|
"""
|
||
|
|
|
||
|
|
def load_model(model_path):
|
||
|
|
model = AutoModelForCausalLM.from_pretrained(
|
||
|
|
model_path,
|
||
|
|
torch_dtype=torch.float16,
|
||
|
|
device_map="auto",
|
||
|
|
trust_remote_code=True
|
||
|
|
)
|
||
|
|
|
||
|
|
tokenizer = AutoTokenizer.from_pretrained(
|
||
|
|
model_path,
|
||
|
|
trust_remote_code=True
|
||
|
|
)
|
||
|
|
|
||
|
|
return model, tokenizer
|
||
|
|
|
||
|
|
def generate_response(model, tokenizer, topic):
|
||
|
|
topic = topic.strip()
|
||
|
|
|
||
|
|
prompt = f"USER: {topic}\nASSISTANT:"
|
||
|
|
|
||
|
|
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
|
||
|
|
|
||
|
|
outputs = model.generate(
|
||
|
|
**inputs,
|
||
|
|
max_new_tokens=3000,
|
||
|
|
temperature=0.7,
|
||
|
|
top_p=0.9,
|
||
|
|
repetition_penalty=1.1,
|
||
|
|
do_sample=True
|
||
|
|
)
|
||
|
|
|
||
|
|
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
|
||
|
|
|
||
|
|
return response.split("ASSISTANT:")[-1].strip()
|
||
|
|
|
||
|
|
def run():
|
||
|
|
model_path = "./" # Path to the directory containing your model weight files
|
||
|
|
|
||
|
|
model, tokenizer = load_model(model_path)
|
||
|
|
|
||
|
|
result = generate_response(model, tokenizer, JSON_INPUT)
|
||
|
|
|
||
|
|
print(result)
|
||
|
|
|
||
|
|
if __name__ == "__main__":
|
||
|
|
run()
|
||
|
|
|
||
|
|
|